[ccp4bb] Quest for resolution-sensitive substructure determination cases

[Randy Read]

Dear members of the BBs,

When determining the anomalous substructure as part of SAD phasing, one of the 
most important parameters can be the choice of resolution limit for the data 
provided to the substructure determination algorithm.  Extending to too high 
resolution can hamper the search by introducing too much noise. The resolution 
limit is varied internally in phenix.hyss, but is provided as a parameter to 
SHELXD.  There are various rules of thumb that people use to make a first 
choice of resolution limit, such as half-dataset anomalous correlation or the 
average precision of the anomalous differences, or even just adding 0.5 to dmin.

We’re currently exploring some alternative measures and we would like to test 
them on a significant number of relevant test cases.  It turns out that a large 
proportion of SAD data deposited in the PDB have such good signal-to-noise that 
substructure determination succeeds with a wide variety of parameters, so we’ve 
only collected a few cases so far.  

It would be great if you could let us know of cases that you’re aware of, where 
substructure determination succeeds with the right choice of resolution limit 
but fails with the full resolution range of data.  Of course, these have to be 
cases where the diffraction data have been deposited at the PDB or otherwise 
been made available.  We prefer data for which the intensities, and not just 
the amplitudes, are available, but we won’t be too picky if that would limit 
the number of examples too much!

If you email me directly, I’ll post a summary to the BBs.

Thanks!

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Postdoc position: posted on behalf of Julia Shifman and Joel Sussman

[Randy Read]

=
Shifman lab in the Hebrew University of Jerusalem is looking for a postdoctoral 
scholar to study evolution of protein-protein interactions. Our lab has 
pioneered computational and experimental methodology for mapping changes in 
binding affinity due to thousands of mutations in protein complexes.  The goal 
of the project is to apply the new methodology to various protein-protein 
interactions. We are looking for a highly motivated person with a background in 
Biophysics/Biochemistry/Structural Biology to take over this exciting project. 
The candidate should have experience with molecular biology and protein 
expression/purification. Experience in yeast surface display is a plus. 
contact Julia Shifman: jshif...@mail.huji.ac.il

Julia
Julia Shifman
Prof. of Biophysics and Biochemistry
Chair, Department of Biological Chemistry
Hebrew University of Jerusalem
Phone:   972-2-658-4078

WEB:  https://openscholar.huji.ac.il/shifmanlab/people/julia-shifman
=


To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Interesting article on the synchrotron data deluge...

[Randy Read]

…with lots of familiar names!

https://physicstoday.scitation.org/do/10.1063/PT.6.2.20200925a/full/

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] [EXTERNAL] Re: [ccp4bb] Best protocols to advance a low resolution twin

[Randy Read]

Hi Huw,

Thanks for pointing that out.  I guess I hadn’t presumed to brand myself as an 
expert!  I’ll have to find out now what else I’ve been missing...

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 14 Sep 2020, at 12:36, Huw Jenkins 
> <288da93ae744-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi Randy,
> 
>> On 14 Sep 2020, at 12:09, Randy Read  wrote:
>> 
>>   Unfortunately, it looks like this is one of the few options that isn’t yet 
>> available from the ccp4i2 interface! 
> 
> Isn't it this one? 
> 
> 
> 
> 
> Note this only appears when 'expert' options are shown:
> 
> 
> 
> 
> Best regards,
> 
> 
> Huw
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] [EXTERNAL] Re: [ccp4bb] Best protocols to advance a low resolution twin

[Randy Read]

Just to follow up on this — if you’ve already tried searching for the 
additional domain with the already placed ones fixed, and that didn’t work, 
it’s possible that the missing domain is less well-ordered than the ones you’ve 
already placed.  So one thing you could try is increasing the relative B-factor 
of the new domain you’re searching for, with the SEARCH BFACTOR command in a 
keyword script, or activating the “Search with bfactor” option in the Expert 
parameters section of ccp4i.  Unfortunately, it looks like this is one of the 
few options that isn’t yet available from the ccp4i2 interface!  I’d probably 
try a number like 20 A^2 as a guess.

If this doesn’t work, then it’s probably worth cutting out the density and 
treating it as a real-space MR problem, using the phased translation target at 
the translation step of the search.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 14 Sep 2020, at 11:27, Schreuder, Herman /DE  
> wrote:
> 
> Dear Andy,
>  
> I few thoughts from my side, but no solution I am afraid:
>   • Your twinning operator -h, -k, l is the standard alternative indexing 
> for P3x space group, which makes a lot of sense.
>   • P32 is a low symmetry space group, which makes MR easier, but this is 
> offset by the NCS.
>   • In my hands, MR is surprisingly insensitive to twinning, so I would 
> search for the molecules using the twinned data. Also, for MR one does not 
> need extremely high resolution data. 
>   • However, twinning and low resolution data seriously hamper “de novo” 
> model building. So if you have good MR models for your d1 and d2 domains, you 
> may have a good chance of solving the structure. If you would have to build 
> them “de novo” in a MR electron density map, you may be doomed.
>   • What I would do, is to look at the phaser map using a very large map 
> radius: say 35-40 Å or more and look in the solvent region if there are 
> places with higher density that may suggest the presence of the missing 
> domains. If something shows up, you can focus on those regions in your MR, or 
> even try to manually fit the Ca chain of your MR model.
>   • You certainly have already done it, but Phaser has to option to 
> search for additional domains, given the domains you already found.
>  
> Best,
> Herman
>  
>  
> Von: CCP4 bulletin board  Im Auftrag von Andrew 
> Lovering
> Gesendet: Montag, 14. September 2020 11:17
> An: CCP4BB@JISCMAIL.AC.UK
> Betreff: [EXTERNAL] Re: [ccp4bb] Best protocols to advance a low resolution 
> twin
>  
> EXTERNAL : Real sender is owner-ccp...@jiscmail.ac.uk
> 
>  
> 
> To follow on from this thread:
>  
> To answer Jon, I did try to see if P6 subgroups were a possibility but can 
> rule this out for a few reasons (MR doesn’t give solutions, merging stats not 
> suggestive of P6, the other dataset with the twin fraction that is 
> significantly further from 50:50); and the d3:d3 NCS is not parallel to any 
> crystallographic axis
>  
> The spacegroup is indeed P3 sub 2, not P321, and the solution again only 
> possible in P3 sub 2 not P3 sub 1, so spacegroup confidence is high
>  
> I did get one reply from Petrus Zwart that twin refinement / map improvement 
> is a subject being worked on
>  
> What I might try is a Phaser MR where the “missing domains” are searched for 
> using cut out density of the one placed domain, rather than model (which 
> could possibly be a better choice at this low resolution? Thoughts 
> appreciated)
>  
> Best wishes & thanks everyone
> Andy
>  
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] taking information from a deposited structure

[Randy Read]

It’s perfectly normal and accepted to use a released entry from the PDB, even 
if there’s no underlying publication.  That is true of many of the structures 
from the structural genomics initiatives, for instance.  As you say, you should 
cite the PDB entry ID and it would be nice also to name the authors of the 
entry.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 9 Sep 2020, at 06:41, Firdous Tarique  wrote:
> 
> Dear CCP4 community members.
> 
> I have solved  a crystal structure of a protein and am now trying to get a 
> structure with a ligand bound to it, but so far unsuccessful. A homologous 
> structure with the same ligand is present in the RCSB PDB (un published). Is 
> it permissible to fetch structural information from that unpublished data and 
> use it for docking and simulation studies with my protein? Will it be 
> copyright infringement or it is just a normal thing.  I will be mentioning in 
> my manuscript that I have used this unpublished structure (deposited by this 
> author) for these studies. 
> 
> Best
> 
> Kahkashan
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] Homology modeling

[Randy Read]

Dear Amit,

I don’t think that PROCHECK is the right tool for assessing the quality of a 
comparative model.  It was revolutionary in its time, but the criteria it looks 
at can mostly be satisfied by energy minimisation even of an incorrect model, 
especially if you don’t have to simultaneously satisfy experimental data.  For 
evaluating refined experimental models, I would now be inclined to use 
Molprobity (available from a website, CCP4 and Phenix) or tools built into 
graphics programs like ISOLDE and coot.

Model quality assessment is actually a thriving subfield in the protein 
modeling community and it has its own category in the CASP modeling challenges. 
 You want to be looking at the methods that have been judged best by seeing how 
well they perform in blind predictions.  Comparative models differ from 
experimental models in that they haven’t been constrained by the requirement to 
fit experimental data, so quality assessment has to look at more subtle 
features like residue environment.

The most recent CASP was CASP13, so a good start would be the main model 
quality assessment evaluation in the special issue from that event: 
https://onlinelibrary.wiley.com/doi/10.1002/prot.25767. You’ll find a variety 
of good tools described there.  The one we’ve been playing with, in terms of 
local assessment of model quality for MR, is ProQ3D, which has an online 
server: http://proq3.bioinfo.se.

Best wishes,
Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 15 Aug 2020, at 00:39, amit gaur  wrote:
> 
> Hi All,
>I am trying to generate a model using comparative modeling. Can anybody 
> suggest the quality of the model based on procheck summary.
> 
>  +--<<<  P  R  O  C  H  E  C  K S  U  M  M  A  R  Y  
> >>>--+
>  |
> |
>  | /var/www/PROCHECK/Jobs/7358861/7358861.pdb   1.5 1253 residues 
> |
>  |
> |
> *| Ramachandran plot:   87.9% core   10.2% allow1.4% gener0.4% disall 
> |
>  |
> |
> *| All Ramachandrans:   67 labelled residues (out of1235) 
> |
> +| Chi1-chi2 plots:  3 labelled residues (out of 771) 
> |
>  | Side-chain params:5 better 0 inside  0 worse   
> |
>  |
> |
> *| Residue properties: Max.deviation: 4.9  Bad contacts:0 
> |
> *| Bond len/angle:9.9Morris et al class:  1  1  2 
> |
>  |
> |
>  | G-factors   Dihedrals:  -0.13  Covalent:  -0.07Overall:  -0.10 
> |
>  |
> |
> *| Planar groups:90.0% within limits  10.0% highlighted  13 off graph 
> |
>  |
> |
>  
> ++
>+ May be worth investigating further.  * Worth investigating further.
> 
> 
> 
> Thanks,
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

Re: [ccp4bb] Phaser Warning Message

[Randy Read]

Thanks to Damian for the clear explanation!

Just a bit more background.  Airlie put this feature in because, with small 
fragments such as helices (which our collaborator Isabel Usón uses in 
ARCIMBOLDO), the initial translation function scores are often higher for 
trials where symmetry-related copies are superimposed, and better solutions 
were otherwise being excluded because they had much lower scores than these 
deceptive ones.  The advisory is printed out so that the user can be aware this 
is happening, in case what you need to do (as Damian says) is to go back and 
edit out a part of your model.

This case turns out to be interesting.  Before the structure has been solved, 
it’s not clear whether the space group is P6522 or its mirror image, P6122.  
Although P6122 is the answer, P6522 is tested first.  In the translation 
search, there will be peaks where part of the structure is right (the molecules 
related by the symmetry operators shared by the two space groups) so the 
likelihood score is much better than random, but the incorrect symmetry will 
lead to clashes.  The advisories come from the translation searches in P6522, 
and then much better scores without serious clashes are found in P6122.

Thanks for sharing the log file so we could see what is going on!

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 16 Jun 2020, at 22:40, Damian Ekiert  wrote:
> 
> Hi Silvia,
> 
> Looking at the end of the log file, it looks like you have a very clear 
> solution and are ready to go!
> 
> This message means that a better solution from the Fast Translation Function 
> was excluded because it failed the packing test (too many clashes with 
> neighboring molecules). Sometimes this will prevent you from getting a usable 
> solution altogether, and this message would clue you into that. In such a 
> situation, you could prune flexible/non-conserved loops from your protein and 
> try MR again with the truncated model.
> 
> But in this case, your still managed to get a very clear solution, so I think 
> you are ready to look at your maps and begin rebuilding and refinement.
> 
> Good luck!
> 
> Best,
> 
> Damian
> 
>> On Jun 16, 2020, at 4:47 PM, Napolitano Silvia 
>>  wrote:
>> 
>> 
>> Dear ccp44 Bulletin Board,
>> 
>> I am trying to solve a structure by molecular replacement.
>> I am doing that by using PhaserMR (simple one-component interface).
>> 
>> I tried out setting up different parameters (e.g.changing # of components, 
>> #search, different space group search settings...), no matter what I try I 
>> always get this warning message:
>> 
>> --
>> Advisory: The top solution from a FTF did not pack
>> --
>> 
>> --
>> Advisory: The top solution from a TF rescoring did not pack
>> 
>> (please, see log file attached).
>> 
>> What does this Warning mean? I tried to look it up online but, 
>> unfortunately, I could not find the answer I was looking for.
>> 
>> Please, let me know if you need more detailed information.
>> 
>> Many thanks in advance for your help and your time!
>> 
>> Best Regards
>> 
>> Silvia
>> 
>> Silvia Napolitano
>> ETH Zurich
>> Institute of Molecular Biology and Biophysics
>> Otto-Stern-Weg 5, HPK E14
>> CH–8093 Zurich
>> Phone: +41 44 6332482
>> 
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>> 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 
> This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing 
> list hosted by www.jiscmail.ac.uk, terms & conditions are available at 
> https://www.jiscmail.ac.uk/policyandsecurity/



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1

This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list 
hosted by www.jiscmail.ac.uk, terms & conditions are available at 
https://www.jiscmail.ac.uk/policyandsecurity/

[ccp4bb] Update on COVID-19 Open Structures initiative

[Randy Read]

Dear all,

Last week I posted the message below, suggesting an open science approach to 
ensure that there will be no unnecessary delays in the determination of crystal 
structures related to COVID-19, whether they’re from the SARS-CoV-2 virus, 
interacting host proteins, or complexes.  Working together, and combining 
expertise in data analysis, structure determination and computational 
modelling, we believe we can greatly accelerate progress on difficult 
structures.

Since last week we’ve chosen a platform, Microsoft Teams, to coordinate the 
effort, created a new Team and built the bare bones of a site ready for new 
targets.  (For those who prefer other collaboration platforms such as Slack or 
Zulip, they might well have advantages but I had easy access to the Teams 
infrastructure and we’re already using it for working from home.)

If you believe that you can bring expertise to this initiative and you’re 
willing to share your insights and any promising results immediately with 
others, please get in touch and I’ll add you to the team.

If you have diffraction data for a target structure for this initiative, please 
send an email to me, to Massimo Sammito (md...@cam.ac.uk) and to the CASP 
organisers (c...@predictioncenter.org). Either Massimo or I will add you to the 
team, upload the target data, links and other information and send an alert to 
everyone who has registered.

Best wishes,

Randy Read

===

The Covid-19 crisis is bringing out the best in the communities we belong to, 
with many people giving deep thought to how we can use our skills to help.  On 
the crystallography bulletin boards we've seen offers to help in solving 
Covid-19-related structures that prove difficult, offers to help with improving 
protein stability, suggestions that the deposition of raw diffraction images 
would allow the community to help get the best possible version of any relevant 
structure, and requests to share bioinformatics analyses and predictions of 
what are the most interesting targets.

We're writing to suggest an additional way that the community can help to 
accelerate progress in the structural understanding of Covid-19.  The CASP 
(Critical Assessment of Structure Prediction) organisers have recently launched 
an initiative to mobilise the structure prediction community to predict and 
refine 3D structures of SARS-2-Covid proteins and relevant complexes that 
either have unknown structure or are non-trivial modelling targets: 
http://predictioncenter.org/caspcommons/index.cgi.
Note that the models will be refined much more extensively than typically done 
in a normal CASP round, which should make them even better than the impressive 
results seen in recent years. 

We would like to build on this initiative, and the enthusiasm this has revealed 
in the prediction community, to help to accelerate the determination of 
structures needed for a molecular-level understanding of Covid-19.  Structure 
prediction has reached a level of maturity where predicted ab initio models and 
distant homology models can be accurate enough to solve new structures by 
molecular replacement.  The best way to bring the prediction and experimental 
communities together to exploit these developments and accelerate progress is 
to embrace an open science approach.  To that end, we propose the following:

* If you have diffraction data involving a SARS-2-Covid protein, a host protein 
relevant to pathogenesis or a complex, but you are not immediately able to 
solve the structure, contact the CASP organisers (c...@predictioncenter.org) 
with the sequence(s) of the construct(s) that went into the crystallisation 
drop.  If relevant predictions have already been made, any unreleased models 
will be released at this point (along with predictions of local accuracy).  If 
proteins in your crystals are not already modelling targets, the CASP 
organisers will consider them as potential new targets for the modelling 
community.

* We all want the fastest possible progress on scientific understanding of 
Covid-19, and this can best be achieved by completely open science.  On a 
number of occasions at crystallographic computing schools and workshops, we 
have seen extremely difficult structures yield to the combined expertise of a 
number of developers and "power users" of the software, none of whom knew how 
to solve every problem that arose.  Even before CASP models are available, some 
other crystallographer may find a way to solve the structure!  So it would be 
ideal if the sequence information you provide to the CASP organisers was 
accompanied by a DOI or URL pointing at the diffraction data, preferably in the 
form of raw images as well as integrated data.  Data from different crystal 
forms or poorly isomorphous crystals can also be incredibly valuable.  Openness 
should go both ways, so people who wish to access these data will be asked to 
agree to immediately release any positive results, even if these 

[ccp4bb] Covid-19: open science to harness synergies in structure prediction and structure solution

[Randy Read]

The Covid-19 crisis is bringing out the best in the communities we belong to, 
with many people giving deep thought to how we can use our skills to help.  On 
the crystallography bulletin boards we've seen offers to help in solving 
Covid-19-related structures that prove difficult, offers to help with improving 
protein stability, suggestions that the deposition of raw diffraction images 
would allow the community to help get the best possible version of any relevant 
structure, and requests to share bioinformatics analyses and predictions of 
what are the most interesting targets.

We're writing to suggest an additional way that the community can help to 
accelerate progress in the structural understanding of Covid-19.  The CASP 
(Critical Assessment of Structure Prediction) organisers have recently launched 
an initiative to mobilise the structure prediction community to predict and 
refine 3D structures of SARS-2-Covid proteins and relevant complexes that 
either have unknown structure or are non-trivial modelling targets: 
http://predictioncenter.org/caspcommons/index.cgi.
Note that the models will be refined much more extensively than typically done 
in a normal CASP round, which should make them even better than the impressive 
results seen in recent years. 

We would like to build on this initiative, and the enthusiasm this has revealed 
in the prediction community, to help to accelerate the determination of 
structures needed for a molecular-level understanding of Covid-19.  Structure 
prediction has reached a level of maturity where predicted ab initio models and 
distant homology models can be accurate enough to solve new structures by 
molecular replacement.  The best way to bring the prediction and experimental 
communities together to exploit these developments and accelerate progress is 
to embrace an open science approach.  To that end, we propose the following:

* If you have diffraction data involving a SARS-2-Covid protein, a host protein 
relevant to pathogenesis or a complex, but you are not immediately able to 
solve the structure, contact the CASP organisers (c...@predictioncenter.org) 
with the sequence(s) of the construct(s) that went into the crystallisation 
drop.  If relevant predictions have already been made, any unreleased models 
will be released at this point (along with predictions of local accuracy).  If 
proteins in your crystals are not already modelling targets, the CASP 
organisers will consider them as potential new targets for the modelling 
community.

* We all want the fastest possible progress on scientific understanding of 
Covid-19, and this can best be achieved by completely open science.  On a 
number of occasions at crystallographic computing schools and workshops, we 
have seen extremely difficult structures yield to the combined expertise of a 
number of developers and "power users" of the software, none of whom knew how 
to solve every problem that arose.  Even before CASP models are available, some 
other crystallographer may find a way to solve the structure!  So it would be 
ideal if the sequence information you provide to the CASP organisers was 
accompanied by a DOI or URL pointing at the diffraction data, preferably in the 
form of raw images as well as integrated data.  Data from different crystal 
forms or poorly isomorphous crystals can also be incredibly valuable.  Openness 
should go both ways, so people who wish to access these data will be asked to 
agree to immediately release any positive results, even if these fall short of 
a full structure solution, so that others can build on them.

Best wishes,

Randy Read (in cooperation with the CASP organisers)

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Se SAD phasing map becomes worse after refinement

[Randy Read]

Dear Zhu,

Questions specific to Phenix should really go to the Phenix-BB, so I am 
cross-posting my reply there.  Here I’ll focus more on generic issues.  There 
are also CCP4 tools that you could consider and presumably other people will 
offer advice on those.

One point you raise comes up so much that we have an entry in our FAQ 
(https://www.phaser.cimr.cam.ac.uk/index.php/FAQ) about it: “Can I use Phaser 
to check the correctness of a model I have already built and refined?”  The 
answer is no, because once you’ve refined the model it becomes better than 
random at predicting the data and therefore achieves a high LLG score, 
regardless of whether or not it is correct.

In this case, you might be able to use Phaser to help complete the model if one 
of the two copies of the complex is more completely modelled than the other.  
If, say, you had a model for one copy you could fix that and search for a 
second copy: this should work because the refinement didn’t know anything about 
the second copy.

Phenix.autosol attempts to determine the NCS operators, so you need to check 
whether that has succeeded.  If not, you might need to try some more manual 
approaches to defining the NCS, which would help a great deal in map 
improvement.  Tom Terwilliger might answer this in more detail (perhaps on the 
Phenix-BB), but I don’t think you can build protein and nucleic acid in the 
same job, so you should look at the documentation to see how to do that.

Good luck!

Randy Read
-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 24 Mar 2020, at 04:31, Zhu Qiao  wrote:
> 
> Dear All
> 
> I am sorry for the long context. 
> 
> I have one protein (252 AAs, 2 Met) bound to double-stranded DNA (24 bp) 
> crystalized.  I collected the Se-Met data of the crystal in C222 up to 2.8 
> angstrom. the space group is confirmed by running the pointless. 
> 
> I used the Phenix.Autosol to find the heavy atoms and get a quite nice map 
> after the density modification.  It seems there are two proteins and two DNA 
> duplex are in one ASU. Phenix.Autobuild can only build less than half of the 
> protein sequence into the map and fill in the potential DNA map with amino 
> acids. The Rwork/Rfree is 0.40 and 0.46, with the map CC=0.60. If I do the MR 
> with the initial model built by Autobuild, the result TFZ=40, LLG=200+, which 
> suggests the partial correction of the initial model. 
> 
> Here is the problem. From the map, I can see one of my protein domain and a 
> clear feature of DNA double helix. But whatever I go further for manual build 
> using coot, like building the DNA double-strand into the map and building the 
> resolved domain, the refinement statistics go bad with R free ~0.50. 
> 
> I am wondering what's going wrong and how come the refinement can't improve 
> the R factor. 
> 
> I have attached the relevant photos. 
> https://drive.google.com/drive/folders/1dJ4kn7CEHkCL3sMcCJtBqa5OsVFQngBx?usp=sharing
> 
> 
> Sincerely
> Zhu 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Error running Phaser-MR

[Randy Read]

Dear Gihan,

The most likely thing I can think of is that the sequence contains a character 
not recognised as a legitimate one-letter code.  This could happen if you use a 
special letter coding for an ambiguous residue, or if you say it’s a nucleic 
acid sequence instead of a protein sequence.

Something else we have seen occasionally is that people save a sequence file 
from a program that writes rich text format or something other than pure text.  
You could try looking at the file with a pure text editor, or using the “file” 
command from the terminal on a Unix (Linux, Mac) system.

If neither of these explain the problem, please send the offending sequence 
file to me (off-list) and I’ll see if I can find out what’s wrong.

Best wishes,

Randy Read
-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building   Fax: +44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.  
www-structmed.cimr.cam.ac.uk

> On 22 Mar 2020, at 03:26, Gihan Ketawala  wrote:
> 
> Hi. 
> 
> I get this "INPUT ERROR: Erro interpreting Composition sequence" message when 
> I try to MR in ccp4 Pherser MR. I've been using regular *.fasta file for the 
> sequence. First, I thought the error coming from the changes I made to the 
> *.fasta fill, but even when I'm using the original file straight from the PDB 
> I still get the error. Any help you guys can give to point me in the right 
> direction is highly appreciated.
> 
> Gihan
> 
> i.e. I'm getting the same error with Phenix-Phaser MR as well
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] [3dem] Which resolution?

[Randy Read]

Hi Alexis,

A brief summary of the relevant points in the paper that Pavel mentioned 
(https://journals.iucr.org/d/issues/2020/03/00/ba5308 
):

The paper is about how to estimate the amount of information gained by making a 
diffraction measurement, not really about defining resolution limits.  Based on 
the results presented, I would argue that it's not a good idea to define a 
traditional resolution cutoff based on average numbers in a resolution shell, 
because there will almost certainly be useful data beyond that point (as long 
as you account properly for the effect of the measurement error, which is 
something that needs work for a lot of programs!).  In our program Phaser, we 
use all of the data provided to scale the data set and refine parameters that 
define the systematic variation in diffraction intensity (and hence signal).  
In this step, knowing which reflections are weak helps to define the parameters 
characterising the systematic variation due to effects like anisotropy and 
translational non-crystallographic symmetry (tNCS).  However, after this point 
the information gained by measuring any one of these reflections tells you how 
much power that observation will have in subsequent likelihood-based 
calculations.  As the information gain drops, the usefulness of the observation 
in determining refineable parameters with the likelihood also drops.  In the 
context of Phaser, we've found that there's a small amount of benefit from 
including reflections down to an information gain of 0.01 bit, but below that 
the observations can safely be ignored (after the scaling, anisotropy and tNCS 
steps).

However, it's possible that average information content is a useful way to 
think about *nominal* resolution.  We should probably do this systematically, 
but our impression from looking at a variety of deposited diffraction data is 
that the average information gain in the highest resolution shell is typically 
around 0.5 to 1 bit per reflection.  So it looks like the half-bit level agrees 
reasonably well with how people currently choose their resolution limits.

For the future, what I would like to see is, first, that everyone adopts 
something like our LLGI target that accounts very well for the effect of 
intensity measurement error:  the current Rice likelihood target using French & 
Wilson posterior amplitudes breaks down for very weak data with very low 
information gain.  Second, I would like to see people depositing at least their 
unpruned intensity data:  not just derived amplitudes, because the conversion 
from intensities to amplitudes cannot be reversed effectively, and not 
intensity data prescaled to remove anisotropy.  Third, I would like to see 
people distinguishing between nominal resolution (which is a useful number to 
make a first guess about which structures are likely to be most accurate) and 
the actual resolution of the data deposited.  There are diminishing returns to 
including weaker and weaker data, but the resolution cutoff shouldn't exclude a 
substantial number of observations conveying more than, say, 0.1 bit of 
information.

Best wishes,

Randy

> On 9 Mar 2020, at 04:06, Alexis Rohou  wrote:
> 
> Hi Colin,
> 
> It sounds to me like you are mostly asking about whether 1/2-bit is the 
> "correct" target to aim for, the "correct" criterion for a resolution claim. 
> I have no view on that. I have yet to read Randy's work on the topic - it 
> sounds very informative. 
> 
> What I do have a view on is, once one has decided one likes 1/2 bit 
> information content (equiv to SNR 0.207) or C_ref = 0.5, aka FSC=0.143 (equiv 
> to SNR 0.167) as a criterion, how one should turn that into an FSC threshold.
> 
> You say you were not convinced by Marin's derivation in 2005. Are you 
> convinced now and, if not, why?
> 
> No. I was unable to follow Marin's derivation then, and last I tried (a 
> couple of years back), I was still unable to follow it. This is despite being 
> convinced that Marin is correct that fixed FSC thresholds are not desirable. 
> To be clear, my objections have nothing to do with whether 1/2-bit is an 
> appropriate criterion, they're entirely about how you turn a target SNR into 
> an FSC threshold.
> 
> A few years ago, an equivalent thread on 3DEM/CCPEM (I think CCP4BB was 
> spared) led me to re-examine the foundations of the use of the FSC in 
> general. You can read more details in the manuscript I posted to bioRxiv a 
> few days ago (https://www.biorxiv.org/content/10.1101/2020.03.01.972067v1 
> ), but 
> essentially I conclude that:
> 
> (1) fixed-threshold criteria are not always appropriate, because they rely on 
> a biased estimator of the SNR, and in cases where n (the number of 
> independent samples in a Fourier shell) is low, this bias is significant
> (2) thresholds in use today do not involve a significance test; they 

Re: [ccp4bb] [3dem] Which resolution?

[Randy Read]

Dear Colin,

Over the last few years we've been implementing measures of information gain to 
evaluate X-ray diffraction data in our program Phaser. Some results in a paper 
that has been accepted for publication in the 2019 CCP4 Study Weekend special 
issue are relevant to this discussion.

First, looking at data deposited in the PDB, we see that the information gain 
in the highest resolution shell is typically about 0.5-1 bit per reflection 
(though we haven't done a comprehensive analysis yet).  A very rough 
calculation suggests that a half-bit resolution threshold is equivalent to 
something like an I/SIGI threshold of one.  So that would fit with the idea 
that a possible resolution limit measure would be the resolution where the 
average information per reflection drops to half a bit.

Second, even if the half-bit threshold is where the data are starting to 
contribute less to the image and to likelihood targets for tasks like molecular 
replacement and refinement, weaker data still contribute some useful signal 
down to limits as low as 0.01 bit per reflection.  So any number attached to 
the nominal resolution of a data set should not necessarily be applied as a 
resolution cutoff, at least as long as the refinement target (such as our 
log-likelihood-gain on intensity or LLGI score) accounts properly for large 
measurement errors.

Best wishes,

Randy

> On 20 Feb 2020, at 10:15, Nave, Colin (DLSLtd,RAL,LSCI) 
>  wrote:
> 
> Dear all,
> I have received a request to clarify what I mean by threshold in my 
> contribution of 17 Feb  below and then post the clarification on CCP4BB. 
> Being a loyal (but very sporadic) CCP4BBer I am now doing this. My musings in 
> this thread are as much auto-didactic as didactic. In other words I am trying 
> to understand it all myself.
>  
> Accepting that the FSC is a suitable metric (I believe it is) I think the 
> most useful way of explaining the concept of the threshold is to refer to 
> section 4.2 and fig. 4 of Heel and Schatz (2005), Journal of Structural 
> Biology, 151, 250-262. Figure 4C show an FSC together with a half bit 
> information curve and figure 4D shows the FSC with a 3sigma curve.
>  
> The point I was trying to make in rather an obtuse fashion is that the choice 
> of threshold will depend on what one is trying to see in the image. I will 
> try and give an example related to protein structures rather than uranium 
> hydride or axons in the brain. In general protein structures consist of atoms 
> with similar scattering power (C, N, O with the hydrogens for the moment 
> invisible) and high occupancy. When we can for example distinguish side 
> chains along the backbone we have a good basis for starting to interpret the 
> map as a particular structure. An FSC with a half bit threshold at the 
> appropriate resolution appears to be a good guide to whether one can do this. 
> However, if a particular sidechain is disordered with 2 conformations, or a 
> substrate is only 50% occupied, the contribution in the electron density map 
> is reduced and might be difficult to distinguish from the noise. A  higher 
> threshold might be necessary to see these atoms but this would occur at a 
> lower resolution than given by the half bit threshold. One could instead 
> increase the exposure to improve the resolution but of course radiation 
> damage lurks. For reporting structures, the obvious thing to do is to show 
> the complete FSC curves together with a few threshold curves (e.g. half bit, 
> one bit, 2 bits). This would enable people to judge whether the data is 
> likely to meet their requirements. This of course departs significantly from 
> the desire to have one number. A compromise might be to report FSC 
> resolutions at several thresholds.
>  
> I understand that fixed value thresholds (e.g. 0.143) were originally adopted 
> for EM to conform to standards prevalent for crystallography at the time. 
> This would have enabled comparison between the two techniques. For many cases 
> (as stated in Heel and Schatz) there will be little difference between the 
> resolution given by a half bit and that given by 0.143. However, if the 
> former is mathematically correct and easy to implement then why not use it 
> for all techniques? The link to Shannon is a personal reason I have for 
> preferring a threshold based on information content. If I had scientific 
> “heroes” he would be one of them.
>  
> I have recently had a paper on x-ray imaging of biological cells accepted for 
> publication. This includes
> 
> “In order to compare theory or simulations with experiment, standard methods 
> of reporting results covering parameters such as the feature examined (e.g. 
> which cellular organelle), resolution, contrast, depth of material (for 2D), 
> estimate of noise and dose should be encouraged. Much effort has gone in to 
> doing this for fields such as macromolecular crystallography but it has to be 
> admitted that this is still an ongoing 

Re: [ccp4bb] Shipping samples for neutron diffraction

[Randy Read]

Hi Gloria,

While we're confessing to minor economies with the truth, I never travelled to 
the Photon Factory in Japan with crystals of pertussis toxin: they were 
crystals of Bordetella pertussis ADP-ribosyltransferase.

Randy

> On 20 Feb 2020, at 02:47, Gloria Borgstahl  wrote:
> 
> Actually we don't declare anything to TSA, Jahaun just carries them in his 
> carryon and it is fine.  Easy peasy?
> 
> 
> On Wed, Feb 19, 2020 at 3:48 PM 
> <0c2488af9525-dmarc-requ...@jiscmail.ac.uk 
> > wrote:
> 
> ... and don't say you're travelling with heavy water!
> 
> Jon Cooper
> 
> On 19 Feb 2020 17:12, "Azadmanesh, Jahaun"  > wrote:
> Hello,
> 
> I have traveled to a neutron beamline ~6 times over the past several years.
> 
> I have had awful luck shipping crystals, so I decided to hand-carry and I 
> found this best.
> 
> I'm not sure what conditions you plan to shoot your crystals.  I mount my 
> crystals in a sturdy quartz capillary (from vitrocom), seal the ends with a 
> layer of wax, then two coats of nail polish (choose your favorite prettiest 
> color!) 
> 
> I would then wrap the capillary in cotton from a cottonball, then "stuff" the 
> capillary-cotton combination in a 15 mL falcon tube for a snug fit. For extra 
> assurance, the falcon tubes were placed inside a generous amount of 
> bubble-wrap that was then taped inside a small cardboard box. 
> 
> I have sought anaerobic conditions for several of my samples to acquire a 
> desired redox state of my metallo-protein. This was tough, so instead I used 
> a high-ish concentration of my redox agent dissolved in my reservoir solution 
> "slugs" in the capillary. These would be replaced in intervals with fresh 
> slugs prior to the beamtime, and for sure right before the beamtime. The 
> redox state held for my ~10 days of beamtime and several months after.
> 
> For cryo-conditions, I just placed a bunch of my crystals in a capillary 
> filled with reservoir solution and sealed as above. I would do my chemical 
> manipulations and freezing at the beamline.
> 
> Much more details are found in our publication: 
> https://www.ncbi.nlm.nih.gov/pubmed/30279321 
>  
> 
> Hope this helps!
> From: CCP4 bulletin board  > on behalf of stephen.c...@rc-harwell.ac.uk 
>   >
> Sent: Wednesday, February 19, 2020 4:22 AM
> To: CCP4BB@JISCMAIL.AC.UK  
> mailto:CCP4BB@JISCMAIL.AC.UK>>
> Subject: [ccp4bb] Shipping samples for neutron diffraction
>  
> Non-UNMC email 
> 
> Dear CCP4 community,
> 
> 
> I have an impending trip to a neutron source and was wondering how people 
> tend to ship their samples prior to beam time?  Is sending something frozen 
> best or is sealed in a capillary more sensible or is there another better 
> way?  One caveat is that ideally my sample should remain anaerobic which has 
> me leaning towards frozen, but I guess in pcr type tubes sealed in gas tight 
> vessels is also a viable option?
> 
> 
> Any thoughts or suggestions are very much appreciated.
> 
> 
> Best wishes,
> 
> 
> Steve
> 
> 
> Dr Stephen Carr
> Research Complex at Harwell (RCaH)
> Rutherford Appleton Laboratory
> Harwell Oxford
> Didcot
> Oxon OX11 0FA
> United Kingdom
> Email stephen.c...@rc-harwell.ac.uk 
> tel 01235 567717
> 
> This email and any attachments may contain confidential, copyright and or 
> privileged material, and are for the use of the intended addressee only. If 
> you are not the intended addressee or an authorized recipient of the 
> addressee, please notify us of receipt by returning the e-mail and do not 
> use, copy, retain, distribute or disclose the information in or attached to 
> this email.
> 
> Any views or opinions presented are solely those of the author and do not 
> necessarily represent those of the Research Complex at Harwell.
> 
> There is no guarantee that this email or any attachments are free from 
> viruses and we cannot accept liability for any damage which you may sustain 
> as a result of software viruses which may be transmitted in or with the 
> message.
> 
> We use an electronic filing system. Please send electronic versions of 
> documents, unless paper is specifically requested.
> 
> This email may have a protective marking, for an explanation, please see:
> https://urldefense.com/v3/__http://www.mrc.ac.uk/About/informationandstandards/documentmarking/index.htm__;!!JkUDQA!b5-sOFziDpUSOnWUVY3BWn_6eloBp1th-M3qGa2sBfF1-Ak7AgXiHT91jdMOlwo3yU5MtpQ$
>  
> 
>  .
> 
> 
> 
> 

Re: [ccp4bb] Dead link

[Randy Read]

It looks like someone at CCP4 will have to fix the link to their local copy of 
the wikidump, but the Phaserwiki link gives you a direct link to all of our 
documentation.

Best wishes,

Randy Read


> On 18 Dec 2019, at 20:29, David J. Schuller  wrote:
> 
> On this CCP4 documentation page:
> 
> http://www.ccp4.ac.uk/html/phaser.html
> 
> The link for "Full documentation" goes nowhere.
> 
> 
> -- 
> ===
> All Things Serve the Beam
> ===
>David J. Schuller
>modern man in a post-modern world
>MacCHESS, Cornell University
>schul...@cornell.edu
> 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: + 44 1223 336500
The Keith Peters Building   Fax: + 44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] [EXTERNAL] Re: [ccp4bb] 1 out of 2 proteins in asymmetric unit does not fit density

[Randy Read]

Dear Jessica,

Good to hear that you've solved it!

We're currently putting a lot of work into making the tNCS treatment in Phaser 
more robust.  Would it be possible to share more information about your case so 
we can understand why it was necessary to manually turn off tNCS in that case?  
Having the data (which we would only use for this purpose) would be best, if 
you could send that off-line.  If you don't feel able to share the data, then 
sample log files from an unsuccessful run and the eventual successful run might 
give us some hints.

Best wishes,

Randy Read

> On 16 Dec 2019, at 21:29, Jessica Besaw  wrote:
> 
> THE PROBLEM IS SOLVED! 
> 
> Thank you all for you suggestions. I applied all the suggestion across all 
> datasets collected (not just the one shown above).
> 
> The error & solution: Incorrect space group assignment (new space group P 21 
> 2 21) and ignoring tNCS. 
> 
> This is the BEST Christmas gift a crystallographer could get.  Thank you 
> everyone!
> 
> Cheers!
> 
> Jessica 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> On Mon, 16 Dec 2019 at 16:16, Wim Burmeister  <mailto:wim.burmeis...@ibs.fr>> wrote:
> Hello,
> I would guess that the badly fitting molecule may be upside down (related my 
> an 2-fold axis).
> I would use the first, partially refined structure for another round of 
> molecular replacement in P212121 with molrep, using the model as well as a 
> partial solution as as asearch model.
> The translational self peak in the native Patterson may be misleading. I came 
> recently across a similar problem.
> Regards
> Wim
> 
> De: "Jessica Besaw" mailto:jbesaw1...@gmail.com>>
> À: "CCP4BB" mailto:CCP4BB@JISCMAIL.AC.UK>>
> Envoyé: Lundi 16 Décembre 2019 20:29:38
> Objet: Re: [ccp4bb] [EXTERNAL] Re: [ccp4bb] 1 out of 2 proteins in asymmetric 
> unit does not fit density
> 
> There have been two potential space groups:
> P212121 - Rfree = 36%
> P21212 - Rfree = 45% 
> 
> Xtriage reports that twinning is unlikely. 
> 
> Cheers!
> 
> Jessica 
> 
> 
> 
> 
> 
> On Mon, 16 Dec 2019 at 13:56, Jürgen Bosch  <mailto:jxb...@case.edu>> wrote:
> What’s your spacegroup ? RWork / RFree?
> Twinning by any chance?
> 
> Jürgen 
> 
> __
> Jürgen Bosch, Ph.D.
> Division of Pediatric Pulmonology and Allergy/Immunology
> Case Western Reserve University
> 2109 Adelbert Rd, BRB 835
> Cleveland, OH 44106
> Phone: 216.368.7565
> Fax: 216.368.4223
> https://www.linkedin.com/in/jubosch/ <https://www.linkedin.com/in/jubosch/>
> 
> CEO & Co-Founder at InterRayBio, LLC
> 
> Johns Hopkins University
> Bloomberg School of Public Health
> Department of Biochemistry & Molecular Biology
> 
> On Dec 16, 2019, at 1:50 PM, Jessica Besaw  <mailto:jbesaw1...@gmail.com>> wrote:
> 
> I am crystallizing this membrane protein in a medium (bicelles) that forms 
> lamella like sheets that stack on top of each other. 
> The layer packing is shown below. Is this structure unreasonable?
> 
> 
> 
> On Mon, 16 Dec 2019 at 13:38, Reza Khayat  <mailto:rkha...@ccny.cuny.edu>> wrote:
> Hi Jessica,
> 
> 
> 
> The gap between the two proteins is a bit troubling. Perhaps it's the image, 
> but why would a crystal form if there is no crystal contact between the two 
> proteins?
> 
> 
> 
> Reza
> 
> 
> 
> Reza Khayat, PhD
> Assistant Professor 
> City College of New York
> Department of Chemistry
> New York, NY 10031
> From: CCP4 bulletin board  <mailto:CCP4BB@JISCMAIL.AC.UK>> on behalf of Ashish Kumar 
> mailto:mail2ashish...@gmail.com>>
> Sent: Monday, December 16, 2019 1:24 PM
> To: CCP4BB@JISCMAIL.AC.UK <mailto:CCP4BB@JISCMAIL.AC.UK>
> Subject: [EXTERNAL] Re: [ccp4bb] 1 out of 2 proteins in asymmetric unit does 
> not fit density
>  
> Hi Jessica,
> 
> It may be possible because of wrong MR solution as well. How were your stats 
> after MR. 
> Also it is correct that it could be possible because of wrong space group.
> Try changing the Space group and repeat MR.
> 
> Best Regards
> Ashish
> 
> On 16 Dec 2019 22:56, "Jessica Besaw"  <mailto:jbesaw1...@gmail.com>> wrote:
> Dear community, 
> 
> I am having a lot of trouble solving a protein structure. I think my problem 
> may caused by incorrectly placed proteins in molecular replacement. I have 
> two proteins in my asymmetric unit. It appears that one protein fits 
> perfectly, and the other one has many errors. (See snapshots below). I have 
> tried deleting the parts of the protein (and even the wh

Re: [ccp4bb] Figure of merit in refinement

[Randy Read]

Dear Eleanor,

Yes, difference maps are weighted by FOM, with the coefficient being m*Fo-D*Fc, 
phased by the model. If Fc is small, then m will be small because, even if Fo 
is large, you have no idea what phase to assign to the difference.  If Fc is 
large because you haven't treated bulk solvent, it turns out that D will 
effectively apply a Babinet scaling because D includes a term with the square 
root of the mean observed intensity divided by the mean calculated intensity.  
If Fc is large even with a bulk solvent correction, then you want a big 
negative term in the difference coefficient so that's fine too!

Best wishes,

Randy

> On 18 Oct 2019, at 07:31, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> This is hunch speak - not proper analysis, but it is possible to get huge 
> Fcalc, and hence large difference map terms,  at low resolution by assuming 
> the solvent volume is a vacuum, not full of partially ordered water 
> molecules. 
> The Babinet scaling can do something to correct this but it is a very blunt 
> tool.  And once a structure is more or less complete the Solvent masked 
> contribution to Fcalc helps, but there is an intermediate stage where 
> spurious differences can distort maps.
> 
> As Randy says - if either Eobs or Ecalc is small the FOM is also small. The 
> worst offenders are when Eobs is large but Ecalc is crazy. 
> I like to look at the plot of  v v resolution,  output by REFMAC 
> along with Rfactor plots. If thee are large discrepancies
> maybe it is time to worry about scaling options..
> 
> Eleanor
> 
> PS - But are difference map terms weighted by FOM? 
> 
> 
> On Thu, 17 Oct 2019 at 08:55, Jan Dohnalek  <mailto:dohnalek...@gmail.com>> wrote:
> Dear all,
> regarding the "remaining strong differences" between measured data and 
> calculated SFs from a a finished (high res structure) I once investigated a 
> bit into this going back to images and looking up some extreme outliers.
> I found the same - those were clear strong diffraction spots, not ice, not 
> small molecule, genuine protein diffraction. So I had no explanation for 
> those. Some were even "forbidden" intensities, because of screw axes which 
> were correct. structure refined perfectly, no problems at all.
> I then found some literature about the possibilities of multiple reflections 
> - I guess this is possible but I wonder if you could get easily say a 25 
> sigma I in this way.
> 
> And as we often end our beer-discussions - may be all protein space groups 
> are actually true P1, just close enough to satisfy the high symmetry rules .. 
> but this is getting a bit philosophical I know ..
> 
> Jan Dohnalek
> 
> 
> 
> 
> On Wed, Oct 16, 2019 at 6:24 PM Randy Read  <mailto:rj...@cam.ac.uk>> wrote:
> James,
> 
> Where we diverge is with your interpretation that big differences lead to 
> small FOMs.  The size of the FOM depends on the product of Fo and Fc, not 
> their difference.  The FOM for a reflection where Fo=1000 and Fc=10 is very 
> different from the FOM for a reflection with Fo=5000 and Fc=4010, even though 
> the difference is the same.
> 
> Expanding on this: 
> 
> 1. The FOM actually depends more on the E values, i.e. reflections smaller 
> than average get lower FOM values than ones bigger than average.  In the 
> resolution bin from 5.12 to 5.64Å of 2vb1, the mean observed intensity is 
> 20687 and the mean calculated intensity is 20022, which means that 
> Eobs=Sqrt(145.83/20687)=0.084 and Ecalc=Sqrt(7264/20022)=0.602.  This 
> reflection gets a low FOM because the product (0.050) is such a small number, 
> not because the difference is big.
> 
> 2. You have to consider the role of the model error in the difference, 
> because for precisely-measured data most of the difference comes from model 
> error.  In this resolution shell, the correlation coefficient between Iobs 
> and Fcalc^2 is about 0.88, which means that sigmaA is about Sqrt(0.88) = 
> 0.94.  The variance of both the real and imaginary components of Ec (as an 
> estimate of the phased true E) will be (1-0.94^2)/2 = 0.058, so the standard 
> deviations of the real and imaginary components of Ec will be about 0.24.  In 
> that context, the difference between Eobs and Ecalc is nothing like a 
> 2000-sigma outlier.
> 
> Looking at this another way, the reason why the FOM is low for this 
> reflection is that the conditional probability distribution of Eo given Ec 
> has significant values on the other side of the origin of the complex plane. 
> That means that the *phase* of the complex Eo is very uncertain.  The figures 
> in this web page 
> (https://www-structmed.cimr.cam.ac.uk/Course/Statistics/statistics.html 
>

Re: [ccp4bb] Figure of merit in refinement

[Randy Read]

James,

Where we diverge is with your interpretation that big differences lead to small 
FOMs.  The size of the FOM depends on the product of Fo and Fc, not their 
difference.  The FOM for a reflection where Fo=1000 and Fc=10 is very different 
from the FOM for a reflection with Fo=5000 and Fc=4010, even though the 
difference is the same.

Expanding on this: 

1. The FOM actually depends more on the E values, i.e. reflections smaller than 
average get lower FOM values than ones bigger than average.  In the resolution 
bin from 5.12 to 5.64Å of 2vb1, the mean observed intensity is 20687 and the 
mean calculated intensity is 20022, which means that 
Eobs=Sqrt(145.83/20687)=0.084 and Ecalc=Sqrt(7264/20022)=0.602.  This 
reflection gets a low FOM because the product (0.050) is such a small number, 
not because the difference is big.

2. You have to consider the role of the model error in the difference, because 
for precisely-measured data most of the difference comes from model error.  In 
this resolution shell, the correlation coefficient between Iobs and Fcalc^2 is 
about 0.88, which means that sigmaA is about Sqrt(0.88) = 0.94.  The variance 
of both the real and imaginary components of Ec (as an estimate of the phased 
true E) will be (1-0.94^2)/2 = 0.058, so the standard deviations of the real 
and imaginary components of Ec will be about 0.24.  In that context, the 
difference between Eobs and Ecalc is nothing like a 2000-sigma outlier.

Looking at this another way, the reason why the FOM is low for this reflection 
is that the conditional probability distribution of Eo given Ec has significant 
values on the other side of the origin of the complex plane. That means that 
the *phase* of the complex Eo is very uncertain.  The figures in this web page 
(https://www-structmed.cimr.cam.ac.uk/Course/Statistics/statistics.html) should 
help to explain that idea.

Best wishes,

Randy

> On 16 Oct 2019, at 16:02, James Holton  wrote:
> 
> 
> All very true Randy,
> 
> But nevertheless every hkl has an FOM assigned to it, and that is used to 
> calculate the map.  Statistical distribution or not, the trend is that hkls 
> with big amplitude differences get smaller FOMs, so that means large 
> model-to-data discrepancies are down-weighted.  I wonder sometimes at what 
> point this becomes a self-fulfilling prophecy?  If you look in detail and the 
> Fo-Fc differences in pretty much any refined structure in the PDB you will 
> find huge outliers.  Some are hundreds of sigmas, and they can go in either 
> direction.
> 
> Take for example reflection -5,2,2 in the highest-resolution lysozyme 
> structure in the PDB: 2vb1.  Iobs(-5,2,2) was recorded as 145.83 ± 3.62 (at 
> 5.4 Ang) with Fcalc^2(-5,2,2) = 7264.  A 2000-sigma outlier!  What are the 
> odds?   On the other hand, Iobs(4,-6,2) = 1611.21 ± 30.67 vs Fcalc^2(4,-6,2) 
> = 73, which is in the opposite direction.  One can always suppose 
> "experimental errors", but ZD sent me these images and I have looked at all 
> the spots involved in these hkls.  I don't see anything wrong with any of 
> them.  The average multiplicity of this data set was 7.1 and involved 3 
> different kappa angles, so I don't think these are "zingers" or other weird 
> measurement problems.
> 
> I'm not just picking on 2vb1 here.  EVERY PDB entry has this problem.  Not 
> sure where it comes from, but the FOM assigned to these huge differences is 
> always small, so whatever is causing them won't show up in an FOM-weighted 
> map.
> 
> Is there a way to "change up" the statistical distribution that assigns FOMs 
> to hkls?  Or are we stuck with this systematic error?
> 
> -James Holton
> MAD Scientist
> 
> On 10/4/2019 9:31 AM, Randy Read wrote:
>> Hi James,
>> 
>> I'm sure you realise this, but it's important for other readers to remember 
>> that the FOM is a statistical quantity: we have a probability distribution 
>> for the true phase, we pick one phase (the "centroid" phase that should 
>> minimise the RMS error in the density map), and then the FOM is the expected 
>> value of the phase error, obtained by taking the cosines of all possible 
>> phase differences and weighting by the probability of that phase difference. 
>>  Because it's a statistical quantity from a random distribution, you really 
>> can't expect this to agree reflection by reflection!  It's a good start to 
>> see that the overall values are good, but if you want to look more closely 
>> you have to look a groups of reflections, e.g. bins of resolution, bins of 
>> observed amplitude, bins of calculated amplitude.  However, each bin has to 
>> have enough members that the average will generally be close to the expected 
>> value.
>> 
>> Best wishes,
>

Re: [ccp4bb] Figure of merit in refinement

[Randy Read]

Hi James,

I'm sure you realise this, but it's important for other readers to remember 
that the FOM is a statistical quantity: we have a probability distribution for 
the true phase, we pick one phase (the "centroid" phase that should minimise 
the RMS error in the density map), and then the FOM is the expected value of 
the phase error, obtained by taking the cosines of all possible phase 
differences and weighting by the probability of that phase difference.  Because 
it's a statistical quantity from a random distribution, you really can't expect 
this to agree reflection by reflection!  It's a good start to see that the 
overall values are good, but if you want to look more closely you have to look 
a groups of reflections, e.g. bins of resolution, bins of observed amplitude, 
bins of calculated amplitude.  However, each bin has to have enough members 
that the average will generally be close to the expected value.

Best wishes,

Randy Read

> On 4 Oct 2019, at 16:38, James Holton  wrote:
> 
> I've done a few little experiments over the years using simulated data where 
> I know the "correct" phase, trying to see just how accurate FOMs are.  What I 
> have found in general is that overall FOM values are fairly well correlated 
> to overall phase error, but if you go reflection-by-reflection they are 
> terrible.  I suppose this is because FOM estimates are rooted in amplitudes.  
> Good agreement in amplitude gives you more confidence in the model (and 
> therefore the phases), but if your R factor is 55% then your phases probably 
> aren't very good either.  However, if you look at any given h,k,l those 
> assumptions become less and less applicable.  Still, it's the only thing 
> we've got.
> 
> 2qwAt the end of the day, the phase you get out of a refinement program is 
> the phase of the model.  All those fancy "FWT" coefficients with "m" and "D" 
> or "FOM" weights are modifications to the amplitudes, not the phases.  The 
> phases in your 2mFo-DFc map are identical to those of just an Fc map.  
> Seriously, have a look!  Sometimes you will get a 180 flip to keep the sign 
> of the amplitude positive, but that's it.  Nevertheless, the electron density 
> of a 2mFo-DFc map is closer to the "correct" electron density than any other 
> map.  This is quite remarkable considering that the "phase error" is the same.
> 
> This realization is what led my colleagues and I to forget about "phase 
> error" and start looking at the error in the electron density itself 
> (10.1073/pnas.1302823110).  We did this rather pedagogically.  Basically, 
> pretend you did the whole experiment again, but "change up" the source of 
> error of interest.  For example if you want to see the effect of sigma(F) 
> then you add random noise with the same magnitude as sigma(F) to the Fs, and 
> then re-refine the structure.  This gives you your new phases, and a new map. 
> Do this 50 or so times and you get a pretty good idea of how any  source of 
> error of interest propagates into your map.  There is even a little feature 
> in coot for animating these maps, which gives a much more intuitive view of 
> the "noise".  You can also look at variation of model parameters like the 
> refined occupancy of a ligand, which is a good way to put an "error bar" on 
> it.  The trick is finding the right source of error to propagate.
> 
> -James Holton
> MAD Scientist
> 
> 
> On 10/2/2019 2:47 PM, Andre LB Ambrosio wrote:
>> Dear all,
>> 
>> How is the phase error estimated for any given reflection, specifically in 
>> the context of model refinement? In terms of math I mean.
>> 
>> How useful is FOM in assessing the phase quality, when not for initial 
>> experimental phases?
>> 
>> Many thank in advance,
>> 
>> Andre.
>> 
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
>> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: + 44 1223 336500
The Keith Peters Building   Fax: + 44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] One protein, two data sets

[Randy Read]

Hi,

If you reindexed crystal 1, the cell dimensions could be a=61.0, b=133.23, 
c=100.34, which would be correspond roughly to doubling the b-cell edge of 
crystal 2.  Is there a sign of this in the data, i.e. in the current indexing 
of crystal form 1 is there a big peak around 0,0,1/2 in the native Patterson 
map?  If that's true, the two crystal forms could be closely related.

Best wishes,

Randy Read

> On 8 Sep 2019, at 12:15, Prerana G.  wrote:
> 
> Dear all,
> 
> We have two data sets of a protein with the following parameters:
> 1. Space group P212121 a=61.0, b=100.34, c=133.23 No. of molecules in ASU - 2
> 2. Space group P212121 a=58.33, b=62.86, c=109.45 No. of molecules in ASU - 1
> 
> Can we use them as two different structures?
> 
> Regards,
> Prerana
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: + 44 1223 336500
The Keith Peters Building   Fax: + 44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Another difficult MR case

[Randy Read]

Dear Napoleão,

Yes, I agree with Phil that it looks like a case where there *is* translational 
NCS but it's not being used by Phaser, either because it wasn't automatically 
detected (native Patterson peak just under the default limit? more than one 
off-origin Patterson peak?) or because Phaser was told to ignore tNCS.  You 
should look in detail at the relevant section of the logfile, and manually 
override Phaser's automated decision if you think there's good evidence for 
tNCS.  I would say that, as Phil noted, the fact that two molecules are in 
basically the same orientation separated by a translation is pretty strong 
evidence.  In particular, the fact that placing the second molecule in the same 
orientation gave such a large increase in both LLG and TFZ is strong evidence: 
this tells us that having two molecules in the same orientation (even if 
they're the wrong molecule or in the wrong orientation) explains some feature 
of the data, i.e. the modulation caused by tNCS.

I'd be happy to look at the log file for you if you find it hard to interpret.

Best wishes,

Randy Read

> On 29 Aug 2019, at 17:24, Phil Jeffrey  wrote:
> 
> Are you *sure* there's no translational NCS ?
> 
> For example your first molecular replacement solution out of Phenix shows
> 
> EULER  293.6   27.7  288.7
> FRAC -0.02  0.02  0.02
> (that's "first molecule at origin in P1")
> 
> and
> 
> EULER  294.0   27.9  288.8
> FRAC -0.37  0.02  0.02
> 
> which is essentially the same orientation, and a translation down one 
> crystallographic axis (a*)
> 
> And this suggests to me that either Xtriage or Phaser is missing something 
> here.  Does Phaser find translational NCS in its initial data analysis ?  
> Unmodeled translational NCS could cause significant problems with the 
> molecular replacement search.
> 
> Phil Jeffrey
> Princeton
> 
> 
> 
> 
> On 8/29/19 11:28 AM, Napoleão wrote:
>> Deal all,
>> Sorry for the long post.
>> I have a data set obtained from a crystal produced after incubating a 
>> protease with a protein which is mostly composed by an antiparallel beta 
>> sheet. I have tried numerous approaches to solve it, and failed. Molecular 
>> replacement using Phaser, and the protease or the protein as a template 
>> yields no solution. However, molecular replacement using only part of the 
>> beta sheet yields LLG=320 TFZ==28.0 (see below).
>> The apparently good data extends to 1.9 A, as processed by XDS, and the 
>> space group is P1 (pointless agree). XDS info below:
>> SPACE_GROUP_NUMBER=1
>> UNIT_CELL_CONSTANTS=44.4372.2977.30  97.802  89.939 101.576
>>  ab  ISa
>>  9.647E-01  3.176E-03   18.07
>>  RESOLUTION NUMBER OF REFLECTIONSCOMPLETENESS R-FACTOR  R-FACTOR 
>> COMPARED I/SIGMA   R-meas  CC(1/2)  Anomal  SigAno   Nano
>>LIMIT OBSERVED  UNIQUE  POSSIBLE OF DATA   observed  expected 
>>  Corr
>>  1.90   24890   19149 23814   80.4%  58.1% 63.7%
>> 114820.77 82.2%63.8* 30.694 492
>> total  163756  125884146938   85.7%  10.6% 10.8%
>> 757443.78 15.0%99.0*-30.7615834
>> Xtriage in Phenix 1.16-3549 gives me all green lights (print below), 
>> suggesting the data presents no twinning, no translational NCS, no ice rings 
>> and is not anisotropic.
>> http://fullonline.org/science/phenix_xtriage_green.png
>> Molecular replacement in Phaser yields single solutions like:
>>Solution annotation (history):
>>SOLU SET  RFZ=3.0 TFZ=* PAK=0 LLG=29 RFZ=2.8 TFZ=8.8 PAK=1 LLG=310 
>> TFZ==27.6
>> LLG=320 TFZ==28.0
>>SOLU SPAC P 1
>>SOLU 6DIM ENSE ensemble1 EULER  293.6   27.7  288.7 FRAC -0.02  0.02  
>> 0.02 BFAC
>> -6.03
>>SOLU 6DIM ENSE ensemble1 EULER  294.0   27.9  288.8 FRAC -0.37  0.02  
>> 0.02 BFAC
>> -6.52
>>SOLU ENSEMBLE ensemble1 VRMS DELTA -0.1983 RMSD  0.49 #VRMS  0.21
>> or partial solutions like:
>>Partial Solution #1 annotation (history):
>>SOLU SET  RFZ=3.7 TFZ=* PAK=0 LLG=32 RFZ=2.8 TFZ=13.0 PAK=0 LLG=317 
>> TFZ==30.2
>> LLG=331 TFZ==30.5 RFZ=2.4 TFZ=7.2 PAK=0 LLG=464 TFZ==18.5 RFZ=2.7 
>> TFZ=5.7 PAK=1
>> LLG=501 TFZ==6.8 LLG=509 TFZ==6.6
>>SOLU SPAC P 1
>>SOLU 6DIM ENSE ensemble1 EULER   85.4  153.0  138.5 FRAC -0.01 -0.00 
>> -0.00 BFAC
>> -12.30
>>SOLU 6DIM ENSE ensemble1 EULER   86.2  153.2  139.5 FRAC -0.36 -0.01 
>> -0.01 BFAC
>> -9.16
>>SOLU 6DIM ENSE ensemble1 EULER   83.8  152.3  135.9 FRAC -0.00  0.00 
&g

Re: [ccp4bb] higher older tNCS

[Randy Read]

Dear Jessica,

There's a dedicated Phenix bulletin board, where it would be better to post 
Phenix-specific questions.

Nonetheless, you can do this in both the CCP4 and Phenix interfaces to Phaser, 
so I'll explain how to do this in all the interfaces.

First, I'm guessing that the order of the tNCS wasn't unambiguous, that Phaser 
chose something other than NMOL 3 in the automated run, but now you want to 
test the alternative of NMOL 3.

ccp4i: Under the "User parameters" tab, change "Number of assemblies related by 
TNCS vector" to 3.  The toggle should automatically be checked when you do that.

ccp4i2: In the Expert Mode Phaser MR GUI, under the "Keywords" tab, change 
"Number of molecules or complexes related by translation vector" to 3.

Phenix: in the Phaser-MR (full-featured) GUI, "Input and general options" tab, 
click on "Other settings".  In the "Translational NCS" section of the popup, 
set "Number of molecules or complexes related by translation vector" to 3.

And for completeness, if you're writing a command-line script, the command is 
"TNCS NMOL 3".

Best wishes,

Randy Read

> On 23 Jul 2019, at 16:06, Jessica Besaw  wrote:
> 
> Hello everyone, 
> 
> Does anyone know of a good phenix tutorial on how to deal with higher order 
> translational non-crystallographic symmetry (tNCS) in a structure? 
> 
> Ultimately, I need to know how to set TNCS NMOL to 3 in phaser of Phenix. I 
> am using the phenix GUI, and I am uncertain how to alter this parameter.  
> 
> Any help would be greatly appreciated
> 
> Cheers!
> 
> Jessica 
> 
>  
> 
>  
> <http://www.avg.com/email-signature?utm_medium=email_source=link_campaign=sig-email_content=webmail_term=oa-4885-a>
>   Virus-free. www.avg.com 
> <http://www.avg.com/email-signature?utm_medium=email_source=link_campaign=sig-email_content=webmail_term=oa-4885-a>
>  
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: + 44 1223 336500
The Keith Peters Building   Fax: + 44 1223 336827
Hills Road   E-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Does ncs bias R-free? And if so, can it be avoided by special selection of the free set?

[Randy Read]

Dear Ian,

I think the missing ingredient in your argument is an assumption that may be 
implicit in what others have written: if you have NCS in your crystal, you 
should be restraining that NCS in your model.  If you do that, then the 
NCS-related Fcalcs will be similar (especially in the particularly problematic 
case where the NCS is nearly crystallographic), and if the working reflections 
are over-fit to match the Fobs values, then the free reflections that are 
related by the same NCS will also be overfit.  So the measurement errors don't 
have to be correlated, just the modelling errors.

Best wishes,

Randy

> On 5 Jun 2019, at 13:58, Ian Tickle  wrote:
> 
> 
> Hi Jon
> 
> Sorry I didn't intend for my response to be interpreted as saying that anyone 
> has suggested directly that the measurement errors of NCS-related reflection 
> amplitudes are correlated.  In fact the opposite is almost certainly true 
> since the only obvious way in practice that errors in Fobs could be 
> correlated is via errors in the batch scale factors which would introduce 
> correlations between errors in Fobs for reflections in the same or adjacent 
> images, but that has nothing to do with NCS.  That's the 'elephant in the 
> room': no-one has suggested that reflections on the same or adjacent images 
> should not be split between the working and test sets, yet that's easily the 
> biggest contributor to CV bias with or without NCS!  I think taking that 
> effect into account would be much more productive than worrying about NCS, 
> but performing the test-set sampling in shells can't possibly address that, 
> since the images obviously cut across all shells.
> 
> The point I was making was that correlation of errors in NCS-related Fobs 
> would appear to be the inevitable _implication_ of what certainly has been 
> claimed, namely that NCS can introduce bias into CV statistics if the 
> test-set sampling is not done correctly, i.e. by splitting NCS-related Fobs 
> between the working and test sets.  Unless there's something I've missed 
> that's the only possible explanation for that claim.  This is because 
> overfitting results from fitting the model to the errors in Fobs, and the CV 
> bias arises from correlation of those errors if the NCS-related Fobs are 
> split up, thus causing the degree of overfitting to be underestimated and 
> giving a too-rosy picture of the structure quality.  Indeed you seem to be 
> saying that because the NCS-related Fobs are correlated (a patently true 
> statement), then it follows that the errors in those Fobs are also 
> correlated, or at least no more correlated than for non-NCS-related Fobs, but 
> I just don't see how that can be true.
> 
> Rfree is not unbiased: as a measure of the agreement it is biased upwards by 
> overfitting (otherwise how could it be used to detect overfitting?), by 
> failing to fit with the uncorrelated errors in the test-set Fobs, just as 
> Rwork is biased downwards by fitting to the errors in the working-set Fobs.  
> Overfitting becomes immediately apparent whenever you perform any refinement, 
> so the only point at which there is no overfitting is for the initial model 
> when Rwork and Rfree are equal, apart from a small difference arising from 
> random sampling of the test-set (that sampling error could be reduced by 
> performing refinements with all 20 working/test sets combinations and 
> averaging the R values).  From there on the 'gap' between Rwork and Rfree is 
> a measure of the degree of overfitting, so we should really be taking some 
> average of Rwork and Rfree as the true measure of agreement (though the 
> biases are not exactly equal and opposite so it's not a simple arithmetic 
> mean).  The goal of choosing the appropriate refinement parameters, 
> restraints and weights is to _minimise_ overfitting, not eliminate it.  It is 
> not possible to eliminate it completely: if it were then Rwork and Rfree 
> would become equal (apart from that small effect from random sampling).
> 
> I don't follow your argument about correlation of Fobs from NCS.  
> Overfitting, and therefore CV bias, arises from the _errors_ in the Fobs not 
> from the Fobs themselves, and there's no reason to believe that the Fobs 
> should be correlated with their errors.  You say "any correlation between the 
> test-set and the working-set F's due to NCS would be expected to reduce 
> R-free".  If the working and test sets are correlated by NCS that would mean 
> that Rwork is correlated with Rfree so they would be reduced equally!  There 
> are two components of the Fobs - Fcalc difference: Fcalc - Ftrue (the model 
> error) and Fobs - Ftrue (the data error).  The former is completely 
> correlated between the working and test sets (obviously since it's the same 
> model) so what you do to one you must do to the other.  The latter can only 
> be correlated by NCS if NCS has an effect on errors in the Fobs, which it 
> doesn't, or by some other effect such as 

Re: [ccp4bb] High Rfree in last Shell

[Randy Read]

Hi,

No-one else seems to have said this, but just from the data in that table, if I 
were a referee I would almost certainly be asking why you seem to have 
discarded useful data at higher resolution than 3.1A!

Best wishes,

Randy Read

> On 16 Apr 2019, at 20:59, Jan van Agthoven  wrote:
> 
> Dear all,
> 
> The last resolution shell is 3.1-3.2 Å. I corrected high resolution Rmerge 
> and Rmeas which seemed truncated, and added cc1/2. Rsym is Rpim which was a 
> typo.
> 
> Resolution range (Å)   49.2-3.1   
>49.3-3.1
> Rfactor (%)24.0 (32.4)
>   23.4 (32.0)
> Rfree (%)  26.6 (29.2)
>   26.3 (31.6)
> 
> Data collection
> Completeness 100 (100)
> 100 (100)
> 
> Redundancy6.9 (7.0)   
>6.2 (6.3)
> 
> Molecules in asymmetric unit  1   
>1
> 
> Average I/σ 14.1 (1.7)
> 15.3 (2.0)
> 
> Rmerge (%)  14.9 (169.4)  
>12.7 (131.8)
> 
> Rmeas (%)16.2 (183.2) 
>  13.9 (143.9)
> 
> Rpim (%)   6.2 (68.6) 
>5.5 (57.1)
> Wilson B-factor 65.6  
>   62.7
> 
> cc1/20.997 (0.561)
>   0.998 (0.616)
> 
> 
>> On Apr 16, 2019, at 2:21 PM, Diana Tomchick 
>> > <mailto:diana.tomch...@utsouthwestern.edu>> wrote:
>> 
>> Why are the the values for R(free) in parentheses higher than R(factor) in 
>> parentheses? I would not expect that to happen, if the same resolution 
>> limits for the last shell are used for both values. And I echo the previous 
>> commentator, please provide the resolution limits for the values in 
>> parentheses.
>> 
>> What exactly is R(factor) in this case? The R(work), or the [R(work) + 
>> R(free)]? Please define.
>> 
>> Diana
>> 
>> **
>> Diana R. Tomchick
>> Professor
>> Departments of Biophysics and Biochemistry
>> UT Southwestern Medical Center
>> 5323 Harry Hines Blvd.
>> Rm. ND10.214A
>> Dallas, TX 75390-8816
>> diana.tomch...@utsouthwestern.edu <mailto:diana.tomch...@utsouthwestern.edu>
>> (214) 645-6383 (phone)
>> (214) 645-6353 (fax)
>> 
>> On Apr 16, 2019, at 11:57 AM, Jan van Agthoven > <mailto:janc...@gmail.com>> wrote:
>> 
>> Hi everyone,
>> I’m trying to publish two structures at 3.1Å resolution with the following 
>> refinement statistics:
>> 
>> Resolution range (Å)   49.2-3.1  
>> 49.3-3.1
>> Rfactor (%)24.0 (32.4)   
>>23.4 (32.0)
>> Rfree (%)  26.6 (29.2)   
>>26.3 (31.6)
>> 
>> Data collection
>> Completeness  100 (100)  
>>   100 (100)
>> 
>> Redundancy6.9 (7.0)  
>> 6.2 (6.3)
>> 
>> Molecules in asymmetric unit  1  
>> 1
>> 
>> Average I/σ 14.1 (1.7)   
>>  15.3 (2.0)
>> 
>> Rmerge (%)  14.9 (100)   
>> 12.7 (100)
>> 
>> Rmeas (%)16.2 (100)  
>>  13.9 (100)
>> 
>> Rsym (%)   6.2 (68.6)
>> 5.5 (57.1)
>> Wilson B-factor 65.6 
>>62.7
>> 
>> I’ve been told that the Rfree factor in the last shell are too high. Does 
>> anyone know how I can improve these Rfree factors other then cutting the 
>> resolution, which already is rather low?
>&

Re: [ccp4bb] Deposition requirement of anomalous pairs?

[Randy Read]

I’d be with you, Graeme, as a good compromise between merged data and raw 
diffraction images.  Though it does seem to be getting more practical to make 
image data available…

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 20 Mar 2019, at 12:31, graeme.win...@diamond.ac.uk 
>  wrote:
> 
> Hi Eleanor
> 
> I’d raise you scaled but unmerged intensities :-)
> 
> Best wishes Graeme
> 
> 
> On 20 Mar 2019, at 12:21, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk>
>  wrote:
> 
> 
> Wouldn't it be a good idea if the pdb required that data deposition provide h 
> k l I+ SIGI+  I- SIGI-  ??
> 
> Even if the depositor has not exploited the anomalous signal, it may well be 
> there sufficiently strongly to validate CYS or phosphates.
> 
> I believe aata processing programs provide this information ?
> 
> Eleanor Dodson
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1
> 
> -- 
> This e-mail and any attachments may contain confidential, copyright and or 
> privileged material, and are for the use of the intended addressee only. If 
> you are not the intended addressee or an authorised recipient of the 
> addressee please notify us of receipt by returning the e-mail and do not use, 
> copy, retain, distribute or disclose the information in or attached to the 
> e-mail.
> Any opinions expressed within this e-mail are those of the individual and not 
> necessarily of Diamond Light Source Ltd. 
> Diamond Light Source Ltd. cannot guarantee that this e-mail or any 
> attachments are free from viruses and we cannot accept liability for any 
> damage which you may sustain as a result of software viruses which may be 
> transmitted in or with the message.
> Diamond Light Source Limited (company no. 4375679). Registered in England and 
> Wales with its registered office at Diamond House, Harwell Science and 
> Innovation Campus, Didcot, Oxfordshire, OX11 0DE, United Kingdom
> 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] help with TNCS + MR

[Randy Read]

Dear Almu,

We have a discussion on the Phaser Wiki about some of the possibilities for 
tNCS, what you should look for, and what Phaser can handle: 
http://www.phaser.cimr.cam.ac.uk/index.php/Molecular_Replacement#Translational_Non-crystallographic_Symmetry.
  As it explains, if you have one major peak in the native Patterson then 
everything will be pretty automatic, and it can even be straightforward if you 
have several peaks corresponding to multiples of the same underlying 
translation.  If you have some specific questions about your problem after 
you’ve read that and run Phaser on your problem, then let us know!

Best wishes,

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 14 Mar 2019, at 05:07, Almudena Ponce Salvatierra  
> wrote:
> 
> Dear all, 
> 
> I have a dataset with strong TNCS and I would like to know if there are a 
> "series of steps" that I could follow in order to find out whether I can 
> solve my structure or not. 
> 
> I see on the Phaser wiki that structure determination in the presence of TNCS 
> is not fully automated. which steps are those in which my intervention should 
> be needed and what do I look for?
> 
> It is the first time that I deal with this problem and I'd be immensely 
> thankful if I could get some hints from those of you who have encountered it 
> in the past or understand where in particular I should keep and eye on :)
> 
> All the best, 
> 
> cheers, 
> 
> Almu
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Change dimer assembly in ASU

[Randy Read]

Dear Ezequiel,

There is nothing special about which particular symmetry copies a molecular 
replacement program chooses, so there is no good reason to stick to those 
symmetry copies.  On the other hand, there are very good reasons to present a 
molecule that is as close as possible to what is biologically relevant, so you 
should definitely change the choices of symmetry copies to make a proper 
heterodimer in your PDB entry.  The easiest way I know to do that is with the 
option in coot: Extensions->Modelling->Symm Shift Reference Chain Here (after 
centering on an atom in a symmetry copy that you want to make the master copy).

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 4 Mar 2019, at 17:27, Eze Chivi  wrote:
> 
> Dear CCP4bb community:
> My crystal is a heterodimeric complex. I solved the structure using MR with a 
> related structure (containig the dimer), using a highly automated pipeline. 
> However, the MR solution is not the dimer of biological relevance. The 
> experimentally validated dimer is formed between a protomer in the ASU and 
> one from the adjacent symmetry-related pair of molecules. Is it correct to 
> use a modelling program to assemble the "biologically correct" dimer and then 
> proceed to refinement? Or... is it need to keep the MR solution and inform in 
> the PDB header how the relevant dimer is formed? Many Thanks
>  
> Ezequiel
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] visualising anis B factors

[Randy Read]

Hi Eleanor,

It’s easy in coot.  Load in the PDB with ANISOU records, then select Draw -> 
Anisotropic atoms… -> tick box labelled “Show Anisotropic Atoms”!

Best wishes,

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 23 Jan 2019, at 15:17, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Is there any easy way to do this?
> 
> Coot? ccp4mg?
> 
> Eleanor Dodson
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] translational NCS & twinning

[Randy Read]

Dear Lan,

Yes, that’s a serious problem that has led some people astray, including a few 
papers where people got apparently good R-factors by invoking non-existent 
twinning.

You can find a brief discussion of this point on the CCP4 wiki 
(https://strucbio.biologie.uni-konstanz.de/ccp4wiki/index.php/R-factors), which 
has links to a couple of publications where Garib Murshudov has worked out what 
happens to R-factors when you treat the data as twinned even when it isn’t.

Best wishes,

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 11 Jan 2019, at 23:05, Guan, Lan  wrote:
> 
> Hi Randy,
> 
> 
>> As Jacob and others have mentioned, you will always get lower R-factors once 
>> you treat the data as being twinned, and the more twin operators the bigger 
>> the reduction in R-factors.  
> 
> Do normal data with no twinning, but refined with twin operator(s), show 
> similar phenomenon?  If it decreases R-factors for normal data, how much it 
> can achieve (5-10% lower)?
> 
> Thanks,
> 
> 
> Lan
> 
> 
> 
>> So you need very strong evidence, independent of R-factors, to invoke 
>> twinning.  In this case, the L-test should be reasonably trustworthy even in 
>> the presence of tNCS, and your L-test values are close to what one would 
>> expect for an untwinned crystal.  At most you have partial twinning, or 
>> perhaps twinning of a pseudo-symmetric crystal (a possibility Phil 
>> mentioned), where the effects on intensity statistics are reduced.
>> 
>> One way to address a problem like this is to solve the structure in a lower 
>> symmetry space group (as you have done), but then to check whether the MR 
>> solution actually obeys the higher symmetry.  You can do this by looking at 
>> the crystal packing or at merging statistics for Fcalcs after a refinement 
>> with highly restrained NCS.  A similar sort of analysis is automated in the 
>> Zanuda tool in CCP4.
>> 
>> Dealing with potential complications from combinations of twinning and 
>> pseudosymmetry is one of the more challenging aspects of crystallography, 
>> but it's a good learning experience.  Good luck!
>> 
>> Randy Read
>> 
>>> On 10 Jan 2019, at 22:38, Donghyuk Shin  wrote:
>>> 
>>> Dear all,
>>> 
>>> Thank you very much for all of your suggestions and sharing experiences.
>>> As many of you commented, the current small unit cell C2 refinement seems 
>>> to be incorrect or correct, and I should put some efforts to crack this 
>>> question.
>>> 
>>> - To Phill Jeffrey,
>>> The idea, trying to find high symmetry SG with small unit cell C2 data is 
>>> good idea, and I will try this.
>>> For your last comments, identifiable electron density differences between 
>>> each chain,
>>> I guess there should not be other densities between chains if my current SG 
>>> and model is correct. Am I right?
>>> 
>>> - To Ethan,
>>> Turning off the automatic_tNCS_option seems to be good option.
>>> I think, my current data seems to be twinned then tNCS which I am not sure 
>>> at this moment. But I will keep your advice in my mind.
>>> 
>>> - To Phoebe A. Rice,
>>> It is quite interesting that you also could get structure solution by 
>>> indexing strong spots and having smaller unit cell.
>>> Actually, I was wondering how it was possible that having half-sized unit 
>>> cell could have solution, while full-sized unit cell could not.
>>> It will be great if you can share your experience a bit more (e.g the size 
>>> of smaller unit cell used in initial search for both 1szp and 3pkz)
>>> 
>>> Again, thank you very much for all of your suggestion.
>>> 
>>> Best wishes,
>>> Donghyuk
>>> 
>>> 
>>> 
>>> To unsubscribe from the CCP4BB list, click the following link:
>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__www.jiscmail.ac.uk_cgi-2Dbin_webadmin-3FSUBED1-3DCCP4BB-26A-3D1=DwIFAg=dIAUvHjI5lMnDD45iB3vgA=SZfWxbou0cinb5MH936uQKMweCFFe1qb2Aj8yA2JJ_E=1tRkoB-BHhBWDCtbPikRHJfYIDIqpaGbVoF5xFjTLD4=007v9D7bqs0h2uVFwau-m4cmT4PVlWyYZ2UVxJvWkrs=
>> 
>> --
>> Randy J. Read
>> Departme

Re: [ccp4bb] translational NCS & twinning

[Randy Read]

Hi,

I'm just catching up on the BB after the CCP4 Study Weekend!

Going back to the beginning of this thread, it's true that tNCS and twinning 
have opposite effects on the intensity moments, which can mask each other.  
However, for simple cases (as this one appears to be) with a single NCS 
translation, the tNCS analysis in Phaser seems to be pretty successful at 
deconvoluting those effects.  It would be interesting to see the overall second 
moments and the second moments as a function of resolution from a run in 
Phaser, either in NCS mode or in MR_AUTO mode.

In such a case, once the tNCS is properly characterised, Phaser places 
molecules in pairs (or higher multiples in more complex cases of higher-order 
tNCS).  That should deal automatically with cases like the ones Phoebe 
mentions, where you could temporarily reduce the cell by a factor of two, but 
it works more generally when the tNCS doesn't correspond to a cell doubling.  
We'd appreciate hearing about any cases where reindexing in a smaller cell 
works and Phaser's automated treatment didn't!  Or, indeed, about cases like 
the one Ethan mentioned where turning off the automatic tNCS correction helped. 
 (I've already been in touch with Ethan off-line.)

As Jacob and others have mentioned, you will always get lower R-factors once 
you treat the data as being twinned, and the more twin operators the bigger the 
reduction in R-factors.  So you need very strong evidence, independent of 
R-factors, to invoke twinning.  In this case, the L-test should be reasonably 
trustworthy even in the presence of tNCS, and your L-test values are close to 
what one would expect for an untwinned crystal.  At most you have partial 
twinning, or perhaps twinning of a pseudo-symmetric crystal (a possibility Phil 
mentioned), where the effects on intensity statistics are reduced.

One way to address a problem like this is to solve the structure in a lower 
symmetry space group (as you have done), but then to check whether the MR 
solution actually obeys the higher symmetry.  You can do this by looking at the 
crystal packing or at merging statistics for Fcalcs after a refinement with 
highly restrained NCS.  A similar sort of analysis is automated in the Zanuda 
tool in CCP4.

Dealing with potential complications from combinations of twinning and 
pseudosymmetry is one of the more challenging aspects of crystallography, but 
it's a good learning experience.  Good luck!

Randy Read

> On 10 Jan 2019, at 22:38, Donghyuk Shin  wrote:
> 
> Dear all,
> 
> Thank you very much for all of your suggestions and sharing experiences.
> As many of you commented, the current small unit cell C2 refinement seems to 
> be incorrect or correct, and I should put some efforts to crack this question.
> 
> - To Phill Jeffrey, 
> The idea, trying to find high symmetry SG with small unit cell C2 data is 
> good idea, and I will try this.
> For your last comments, identifiable electron density differences between 
> each chain, 
> I guess there should not be other densities between chains if my current SG 
> and model is correct. Am I right?
> 
> - To Ethan,
> Turning off the automatic_tNCS_option seems to be good option.
> I think, my current data seems to be twinned then tNCS which I am not sure at 
> this moment. But I will keep your advice in my mind.
> 
> - To Phoebe A. Rice,
> It is quite interesting that you also could get structure solution by 
> indexing strong spots and having smaller unit cell.
> Actually, I was wondering how it was possible that having half-sized unit 
> cell could have solution, while full-sized unit cell could not.
> It will be great if you can share your experience a bit more (e.g the size of 
> smaller unit cell used in initial search for both 1szp and 3pkz)
> 
> Again, thank you very much for all of your suggestion.
> 
> Best wishes,
> Donghyuk
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

[ccp4bb] CCP4 Study Weekend: late addition and last chance to register

[Randy Read]

The CCP4 Study Weekend on “Integrated, rational molecular replacement” will be 
held at the East Midlands Conferenced Centre in Nottingham, from 8th to 10th 
January 2019.  There are still some places left, but registration will close 
very soon, on 10th December.

We’re pleased to announce a late addition to the programme, which couldn’t be 
publicised until the results of the CASP13 protein folding challenge had been 
announced.  As you may have heard (if you read tech websites or the Guardian 
newspaper) the people from DeepMind have applied their deep learning algorithms 
to the protein folding problem and have done extremely well in folding the 
protein targets in CASP13.  Andrew Senior, the lead of the AlphaFold team that 
achieved this result, will be attending the Study Weekend to speak about their 
algorithms and their potential applications.

To see the rest of our exciting programme and to find registration information, 
please go to the meeting website: http://www.cvent.com/d/sgq8q6/6X.

Best wishes,

Isabel Usón, Ronan Keegan and Randy Read
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] preparing ellipsoidally truncated data for PDB deposition

[Randy Read]

Dear Kevin,

I’ve just been working through the issue myself of how to deposit both the 
pruned data needed to reproduce the refinement and the unmassaged data.  The 
advice I got from the people at PDBe was to prepare an mmCIF file with two 
reflection loops.  The validation pipeline expects the first one to contain the 
data actually used in refinement, so I prepared one in which the second loop 
contained the full set of intensities.  It should also be possible to have a 
third loop containing unmerged original-index (but probably scaled) intensity 
data, and this would be a good intermediate ground between the worst option of 
just depositing French & Wilson (truncate) amplitudes and the ideal-world case 
of depositing the raw diffraction images.

Like you, I used the sf-tool at wwPDB, but I’ve just double-checked and the 
mmCIF file I got has the same free flags for both reflection loops.  In case 
this helps to narrow down why you had a problem and I didn’t, the data I 
submitted were in a single MTZ file, and the R-free flags had been prepared 
with the CCP4 convention, with values from 0 to 19, and not a binary convention 
with just values of 0 and 1.

I’ve also been wondering about when you get to label the different intensity 
sets!  The mmCIF standard doesn’t have data types that distinguish different 
kinds of observed intensities.  We haven’t yet deposited the structure with the 
data, and I’ve been hoping that you’re asked to provide labels during the 
deposition process.  Perhaps someone from one of the wwPDB sites, or someone 
who has gotten as far as depositing such data, can comment!

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 26 Sep 2018, at 19:34, Kevin Jude  wrote:
> 
> I am preparing to deposit several structures that I refined against 
> anisotropic data that I truncated with STARANISO. I will of course be 
> depositing the original data with spherical resolution limits, but it seems 
> that I should also deposit the ellipsoidally truncated data that I actually 
> refined against. To be clear, these are the same dataset but in the second 
> case the unmerged reflections have been rescaled and I/SIGIs that fall 
> outside the ellipsoid are set to empty values. I have a technical problem 
> with preparing the mmcif and a broader question about presenting the data.
> 
> I've tried to create an mmcif file from my mtz using 
> http://sf-tool.wwpdb.org, treating the two sets of I/sigI as two datasets. 
> For some reason, in the output file the test set flags are reversed for the 
> second "dataset" (whichever I choose to be second); ie, o becomes f and f 
> becomes o. This is a technical problem that I can correct with a text editor, 
> but still irritating.
> 
> More importantly, is there a way to distinguish in the file between the 
> spherically complete dataset and the truncated dataset that was used in 
> refinement in a way that is useful to future users? I have not worked with 
> mmcif before and am not sure what column names are permissible, nor what 
> would be recognizable to other users or software. I'm interested to hear the 
> thoughts and experiences of the community on this.
> 
> Best wishes
> Kevin
> 
> --
> Kevin Jude, PhD
> Structural Biology Research Specialist, Garcia Lab
> Howard Hughes Medical Institute
> Stanford University School of Medicine
> Beckman B177, 279 Campus Drive, Stanford CA 94305
> Phone: (650) 723-6431
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1
> 



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

[ccp4bb] CASP13 conference info

[Randy Read]

Sent on behalf of the CASP13 organisers!

=

CASP13 Dec. 1-4, 2018 Riviera Maya, Mexico

Come and join us at CASP13, the Critical Assessment of protein Structure 
Prediction
conference!

Established in 1994, CASP is the first of now several challenges in 
computational
and structural biology where new predictive methods are tested without an a 
priori
knowledge of the answers.

In the 2018 CASP13 experiment, over 57,000 predictions were collected on 90 
protein
targets. The conference will feature prediction assessment talks, reports from 
the
best predictors, and keynote presentations. Detailed program of the meeting 
will be
available in mid-November, 2018.

Assessment of predictions:

* High accuracy models - Randy Read (Cambridge University, UK)
* Model topology - Matteo Dal Peraro (EPFL, Lausanne, Switzerland)
* Contact prediction - Andras Fiser (Albert Einstein College of Medicine,
 New York, NY, USA)
* Model refinement - Randy Read (Cambridge University, UK)
* Assembly (quaternary structure and complexes) - Jose Duarte (RCSB, San Diego, 
CA,
 USA)
* Model accuracy estimation - Chaok Seok (Seoul National University, South 
Korea)
* Data assisted models - Gaetano Montelione (Rutgers, USA), Susan Tsutakawa 
/Greg
 Hura (LBL, Berkeley, CA, USA), and Andras Fiser (Albert Einstein College
 of Medicine, New York, USA)
* Biological relevance of models – Rosalba Lepore, University of Basel, 
Switzerland

The keynote speakers represent two current areas of special interest in CASP, 
machine
learning /deep learning in structural biology and cryo-EM driven modeling:

* David Koes (U Pittsburgh), is a pioneer in the application of convolutional 
neural
 networks to molecular docking
* Wah Chiu (Stanford), is a leader in cryo-electron microscopy.

CASP13 will be held December 1-4, 2018, at the Iberostar Paraiso Maya 5* resort
(destination airport is Cancun - CUN).

Registration:http://predictioncenter.org/casp13/meeting.cgi
Please share this information with your colleagues.
Please contact Krzysztof Fidelis,kfide...@ucdavis.edu, with any questions.


-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk


To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] tCNS and space group determination

[Randy Read]

Hi,

Looking at the pointless logfile for P222, there’s excellent evidence for a 
2(1) screw along the shortest and longest cell edges as Eleanor says, which 
would make the space group P 21 2 21 for the data set used in the Phaser run 
with the data merged in P222.  The merging statistics are equally good for all 
3 2-folds and for the identity, which implies, as Eleanor says, that the data 
most likely should be merged as orthorhombic.  If the data were twinned that 
would explain merging in higher symmetry even if the true space group were some 
version of P2 or P21, but the twinning statistics look pretty close to what one 
would expect for an untwinned crystal.

Maybe you’ve run Phaser in all 8 possible orthorhombic space groups in some 
other job, but in the job that the file name implies was run in P212121, it was 
actually only run in P222, i.e. the point group in which the data were merged.  
So this job has missed all the more likely space groups.

What Phaser flags as an outlier rejection changed a couple of years ago, though 
we still haven’t published this.  For a long time, Phaser has rejected 
improbably large structure factors, i.e. ones that would be expected to occur 
less than one time in a million according to the Wilson distribution.  For 
about 2 years now, it has also been ignoring reflections flagged as containing 
very little information about the true intensity, i.e. ones in which the 
standard deviation of the intensity (SIGI) is large compared to the Wilson 
expected intensity.  When a crystal has very high anisotropy (as in this case, 
where the anisotropic delta B, or the difference between the weakest and 
strongest directions, is nearly 60 A^2, many of the reflections in the weak 
directions in reciprocal space will contain so little information that they can 
be ignored in the calculation (because they would contribute almost nothing 
apart from raising the CPU time!).  The presence of tNCS introduces more 
reflections that are systematically very weak, as well.  I should probably 
change the log file output to separate the counts for the rejections vs the 
ones being ignored.  There are data sets where nearly half of the reflections 
end up being ignored, so I wouldn’t worry about that for this data set.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 11 Aug 2018, at 20:36, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Marcelo - there is something very wrong with the data. You dont to reprocess 
> in other space groups - the P2/mmm symmetry looks convincing but PHASER says 
> there are 3500 rejections! That is an awful lot - 10 is a more normal value. 
> I cant see the image clearly but something is causing problems. Split 
> crystal? How are you processing it?
> 
> The Moments go crazy at high resolution. Maybe try redoing the MR & 
> refinement  all at 3A? 
>  
> 
> The MR rotation function solution is wonderful, but you need to conside 
> spacegroups P 21 2 21. and P 21 21 21 .
> The translation vector of 0 0.5 0.13 would produce absences along the b axis. 
> The screw axes along a and c look safe. 
> 
> 
> 
> On 11 August 2018 at 19:36, Marcelo Liberato  
> wrote:
> I am sorry. I forgot to attach the image.
> 
> Cheers
> 
> Marcelo
> 
> Em sáb, 11 de ago de 2018 às 18:31, Marcelo Liberato 
>  escreveu:
> Dear Eleanor,
> 
> Thanks for you answer. 
> Indeed, there are clear ice rings in the images (example attached). So, I 
> integrated again (P1, P2 and P222) excluding the resolution ranges 2.28-2.22 
> and 3.70-3.64. I am attaching the log files from aimless, MR and refmac for 
> P2 (in two different cells) and P222 data. 
> I agree that MR seems very good (in all cases), but the final density maps 
> are always bad. Maybe the data has problems that I am not dealing with. 
> 
> Kind regards
> 
> Marcelo
> 
> Em sáb, 11 de ago de 2018 às 16:04, Eleanor Dodson 
>  escreveu:
> This MR looks good to me, but there are serious flaws with the data. Your 
> secon moment plot from the aimless log has most spectacular spikes which are 
> always a BAD THING, and the Wilson plot is not very smooth either.. 
> 
> As Randy says, try to sort those problems out first.
> 
> Then you have this message:
> 
> 
> TRANSLATIONAL NCS:
> 
> Translational NCS has been detected at ( 0.000,  0.500,  0.125).
> A translation of 0.5 along B will generate pseudo-absences along b so you can 
> be sure whether there

Re: [ccp4bb] tCNS and space group determination

[Randy Read]

Dear Marcelo,

First I would look at the data to see if you have ice rings, because the peak 
in mean intensity and second moment of the intensity at about 2.25A resolution 
suggests an ice ring problem.  If so, you should make sure you don't 
contaminate the data with spurious large intensities.

Second, the statistics (e.g. the second moments plot after tNCS correction in 
Phaser) would be consistent with a scenario in which you have pseudosymmetry 
along with a twin operator that parallels the pseudosymmetry.  If that's true, 
it's hard to be sure of the symmetry.  For instance, if the structure really is 
monoclinic, can you be sure you chose the correct axis to be the 2-fold?

Since you have a good model that gives clear MR solutions even in P21, you can 
probably process the data in P1 and solve it with 8 copies in the unit cell.  
Then you can look at the symmetry of the MR solution (e.g. in Zanuda) and see 
whether it obeys any higher symmetry than P1.

Good luck!

Randy Read

> On 7 Aug 2018, at 10:35, Marcelo Liberato  wrote:
> 
> Dear CCP4 members, 
> 
> I need your help to figure out what is going on with my data. 
> I've integrated my data set in space groups P1, P2 and P222. All with cell 
> parameters: 45 89 149 90 90 90.
> 
> Aimless always indicates P212121 as the correct space group and the 
> resolution is about 2.0 A.
> 
> I have tried MR (Phaser) with all these space groups (and the alternative 
> ones) using a trimmed model (I've cut some loops) with 37 % identity (besides 
> the low identity, this enzyme belongs to a protein family with high 
> structural similarity). In most cases MR resulted in TFZ between 12-19 and 
> LLG 200-400. 
> However, refinement with refmac (with and without twin) generally results in 
> Rwork/Rfree ~ 0.4400/0.4800. 
> The best scenario was obtained when data was integrated in P21. MR resulted 
> in TFZ=19 and LLG=450, and refmac (with twin) resulted in Rwork/Rfree = 
> 0.4129/0.4517. This data was used in Autobuild (Phenix), twin law h,-k,-l,  
> resulting in Rwork/Rfree = 0.3525/0.4040. The final model seems to fit the 
> electron density, but the map is not very good and there is a lot of bias. 
> 
> All data generated by aimless and MR in the different space groups were 
> analyzed by Xtriage. In all cases, "Translational pseudo-symmetry is very 
> likely present in these data" with an off-origin peak (38 % of origin peak) 
> at fractional coordinates (0.0, 0.5, 0.13). 
> 
> Now, I am stuck and I have no idea how to solve this problem. Could you 
> please help me with this?
> 
> The log files of aimless, MR, refmac and Xtriage of the "best scenario" are 
> attached. 
> 
> Many thanks
> 
> Marcelo Liberato
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1><49_refmac5.log><46_phaser_MR.log><45_aimless.log>

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

[ccp4bb] Call for targets (not-yet-released structures) for CASP13 modelling experiment

[Randy Read]

Posted on behalf of the CASP organizing committee:

 NB: There are only two weeks left to collect new targets for CASP13, so 
your contributions would be very gratefully received! 

Request for protein structures to test modeling methods - CASP13

CASP community experiments aim to establish and advance the state of the art in 
protein structure modeling.

To this end, CASP collects information on soon-to-be released experimental 
structures for a three month season every two years. Sequences and other 
information on these targets is distributed to the protein modeling community, 
and the resulting models assessed by independent experts in the field. Results 
are published in special issues of the journal PROTEINS (see all about CASP12 
at https://onlinelibrary.wiley.com/toc/10970134/86/S1). About 100 modeling 
groups from around the world take part.

The success of CASP depends completely on the generosity of the experimental 
community in providing modeling targets. We are now requesting targets for the 
CASP13 experiment, which will launch at the beginning of May. So, if you have 
anything suitable, we would be most grateful if you would go to the target 
entry page:
http://www.predictioncenter.org/casp13/targets_submission.cgi

The CASP community needs modeling targets over a wide range of difficulty, for 
modeling with and without the aid of templates. We are particularly interested 
in membrane proteins, protein complexes, and cryo-EM structures.  We are also 
extending CASP to include more modeling assisted by sparse experimental data, 
in collaboration with experimental groups in NMR, SAXS, chemical crosslinking, 
and FRET. For that, protein material is needed (of course this is not expected 
for most targets, but if its available, it would be much appreciated!).

For those of you who have not provided targets to CASP before, the procedure is 
simple – there is a web page for submitting targets, and an experienced staff 
to deal with any queries. We don’t need the structure in advance of its release 
by the PDB, and if we are notified early enough (a minimum of three weeks 
before release, more is better) there need be no delay in structure release. 
More details are available on the web site.

Thanks,
CASP organizing committee:
John Moult, University of Maryland, USA
Krzysztof Fidelis, University of California, Davis, USA
Andriy Kryshtafovych, University of California, Davis, USA
Torsten Schwede, University of Basel, Switzerland

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

[ccp4bb] Software Engineer position to work with Phaser team in Cambridge

[Randy Read]

We've just advertised a position for a Software Engineer to work with us on the 
development of Phaser.  Detailed information on the position and how to apply 
can be found on the University of Cambridge website 
(http://www.jobs.cam.ac.uk/job/17918/) and on jobs.ac.uk 
(http://www.jobs.ac.uk/job/BKS034/software-engineer).

The text of the job description is appended, if you'd like to look at that 
before clicking on one of the links!

Best wishes,

Randy Read

=

We are looking for a graduate Software Engineer to join the group of Prof Read 
in the Cambridge Institute for Medical Research (http://www.cimr.cam.ac.uk/), 
based at the Cambridge Biomedical Campus in Cambridge. The well-established 
project develops mathematical methods and software for the imaging of 
biological molecules. The software supports academic and pharmaceutical R 
worldwide.

You will be joining a small expert academic team. The successful applicant will 
have a careful approach to software development. Your software will be required 
to be reliable, algorithmically fast, and memory efficient. We have a slow 
development cycle and the code you write will be in use for the foreseeable 
future. Your individual contribution will be recognised with authorship of 
peer-reviewed articles; publications from our group have a track record of 
being highly cited.

The project is funded by the National Institute of Health (USA) and the 
successful applicant will have the opportunity to join bi-annual collaborative 
meetings that cycle among Lawrence Berkeley National Laboratory (Berkeley, 
California), Duke University (Durham, North Carolina) and the New Mexico 
Consortium (Santa Fe, New Mexico). We also have an active collaboration with 
IBMB- CSIC (Barcelona, Spain).

Software is open source 
(https://git.uis.cam.ac.uk/x/cimr-phaser/phaser.git/summary). We use the 
C++0x/C++11 standard to maintain cross-platform compilation and compatibility. 
You will have access to the University of Cambridge's computing resources, and 
a laptop will be provided.

Qualifications and Experience: University degree that includes computer science 
and/or mathematics, experience with C++ and the Linux OS. Skills in Python 
scripting, GPU programming and threading would be desirable. Knowledge of 
biology to A level standard or equivalent would be an advantage.

This appointment is full-time (37 hours/week Monday to Friday) but we welcome 
applications from individuals who wish to be considered for part-time working 
or other flexible working arrangements.

Royalty payments from software will extend beyond the lifetime of the contract. 
University of Cambridge employment benefits include USS pension, and reduced 
staff fees for University of Cambridge graduate courses (see 
http://www.jobs.cam.ac.uk/offer/ for a full list).

For more information on the research group see 
https://www.cimr.cam.ac.uk/research/principal-investigators/principal-investigators-q-z/read,
 and details of the Phaser project can be found at 
http://www.phaser.cimr.cam.ac.uk/.

Fixed-term: The funds for this post are available untill 30th April 2022.
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Predicting self-rotation function peaks from coordinates?

[Randy Read]

Eleanor Dodson also mentioned offline that the CCP4 program ROTMAT can do the 
calculation.  At first I thought that you would have to use a separate run for 
each NCS operator, but you can give multiple input angles with ROTMAT as well.

I will have to sort out which angle conventions I want in ROTMAT or glrf, and I 
haven't succeeded in matching the ones from Liang's sample job for glrf yet, 
but here's a sample job for ROTMAT:

  rotmat << eof > rotmat.log
  CRYSTAL NUMBER 1 CELL 30 30 40 90 90 90
  CRYSTAL NUMBER 1 ORTH 1
  CRYSTAL NUMBER 1 SYMM P422
  INPUT CCP4 ALPHA BETA GAMMA 40 70 180
  INPUT CCP4 ALPHA BETA GAMMA 90 90 90
  OUTPUT CCP4 PHI PSI KAPPA
  END
  eof

Best wishes,

Randy

> On 11 Jun 2018, at 13:33, Randy Read  wrote:
> 
> Dear Liang,
> 
> Thanks, that looks like that will do the job very well starting from a list 
> of NCS rotations!  
> 
> Best wishes,
> 
> Randy
> 
>> On 11 Jun 2018, at 12:23, Liang Tong > <mailto:lt...@columbia.edu>> wrote:
>> 
>> 
>> Hello Randy
>>   The SYMShow command in GLRF can take a set of rotation angles (up to 10) 
>> and output all the  angles related by the crystallographic symmetry. 
>> Hopefully this will help with what you need to do. I attach a script and 
>> print file as examples.
>> 
>> 
>> best regards
>> Liang Tong
>> Columbia University
>> 
>> 
>> 
>> On Mon, Jun 11, 2018 at 4:53 AM, Randy Read > <mailto:rj...@cam.ac.uk>> wrote:
>> Hi,
>> 
>> I'd prefer not to have to reinvent a wheel that I'm almost certain is 
>> already out there!  Is there a nice tool for taking a set of coordinates and 
>> space group information, working out the NCS relationships, applying 
>> symmetry and generating a list of peaks that should be expected on a 
>> self-rotation function?  Something that takes a list of NCS operators 
>> instead of coordinates would also be okay, though slightly less convenient.  
>> I've seen annotated self-rotation functions in papers, but I'm failing to 
>> find anything in the list of CCP4 programs that would do this.
>> 
>> Computing a self-rotation function from the Fcalcs is another option to 
>> prove that the structure agrees with the self-rotation function computed 
>> from the data, but it would sometimes be helpful to be able to say which 
>> pair of molecules led to a particular peak.
>> 
>> Thanks!
>> 
>> Randy Read
>> 
>> --
>> Randy J. Read
>> Department of Haematology, University of Cambridge
>> Cambridge Institute for Medical Research  Tel: + 44 1223 336500
>> Wellcome Trust/MRC Building   Fax: + 44 1223 336827
>> Hills RoadE-mail: rj...@cam.ac.uk 
>> <mailto:rj...@cam.ac.uk>
>> Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk 
>> <http://www-structmed.cimr.cam.ac.uk/>
>> 
>> 
>> 
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
>> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
>> 
>> 
> 
> --
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical Research  Tel: + 44 1223 336500
> Wellcome Trust/MRC Building   Fax: + 44 1223 336827
> Hills RoadE-mail: rj...@cam.ac.uk 
> <mailto:rj...@cam.ac.uk>
> Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk 
> <http://www-structmed.cimr.cam.ac.uk/>
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] Predicting self-rotation function peaks from coordinates?

[Randy Read]

Dear Liang,

Thanks, that looks like that will do the job very well starting from a list of 
NCS rotations!  

Best wishes,

Randy

> On 11 Jun 2018, at 12:23, Liang Tong  wrote:
> 
> 
> Hello Randy
>   The SYMShow command in GLRF can take a set of rotation angles (up to 10) 
> and output all the  angles related by the crystallographic symmetry. 
> Hopefully this will help with what you need to do. I attach a script and 
> print file as examples.
> 
> 
> best regards
> Liang Tong
> Columbia University
> 
> 
> 
> On Mon, Jun 11, 2018 at 4:53 AM, Randy Read  <mailto:rj...@cam.ac.uk>> wrote:
> Hi,
> 
> I'd prefer not to have to reinvent a wheel that I'm almost certain is already 
> out there!  Is there a nice tool for taking a set of coordinates and space 
> group information, working out the NCS relationships, applying symmetry and 
> generating a list of peaks that should be expected on a self-rotation 
> function?  Something that takes a list of NCS operators instead of 
> coordinates would also be okay, though slightly less convenient.  I've seen 
> annotated self-rotation functions in papers, but I'm failing to find anything 
> in the list of CCP4 programs that would do this.
> 
> Computing a self-rotation function from the Fcalcs is another option to prove 
> that the structure agrees with the self-rotation function computed from the 
> data, but it would sometimes be helpful to be able to say which pair of 
> molecules led to a particular peak.
> 
> Thanks!
> 
> Randy Read
> 
> --
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical Research  Tel: + 44 1223 336500
> Wellcome Trust/MRC Building   Fax: + 44 1223 336827
> Hills RoadE-mail: rj...@cam.ac.uk 
> <mailto:rj...@cam.ac.uk>
> Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk 
> <http://www-structmed.cimr.cam.ac.uk/>
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1 
> <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1>
> 
> 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk




To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

[ccp4bb] Predicting self-rotation function peaks from coordinates?

[Randy Read]

Hi,

I'd prefer not to have to reinvent a wheel that I'm almost certain is already 
out there!  Is there a nice tool for taking a set of coordinates and space 
group information, working out the NCS relationships, applying symmetry and 
generating a list of peaks that should be expected on a self-rotation function? 
 Something that takes a list of NCS operators instead of coordinates would also 
be okay, though slightly less convenient.  I've seen annotated self-rotation 
functions in papers, but I'm failing to find anything in the list of CCP4 
programs that would do this.

Computing a self-rotation function from the Fcalcs is another option to prove 
that the structure agrees with the self-rotation function computed from the 
data, but it would sometimes be helpful to be able to say which pair of 
molecules led to a particular peak.

Thanks!

Randy Read

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk



To unsubscribe from the CCP4BB list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB=1

Re: [ccp4bb] [ccp4bb] tNCS problem

[Randy Read]

Hi,

Phaser is pretty robust now if you have 2-fold tNCS (indicated by having only 
one large non-origin peak in the native Patterson map), and recent versions are 
better at dealing with higher order tNCS (i.e. 3 or more copies related by 
multiples of a single translation vector).  However, getting higher order tNCS 
to work properly can take a bit more thought and analysis outside the program, 
and if you're unlucky enough to have 2 or more different independent tNCS 
vectors it's not sophisticated enough to deal with that.

One common problem is that tNCS interferes with proper space group 
identification.  For instance, it can be difficult to tell whether you have a 
crystallographic 2(1) screw axis with a parallel NCS 2-fold, or 
crystallographic 2-fold with parallel NCS screw axis.  Both of these will lead 
to tNCS, but one will be much closer to reality.  Also, tNCS can obscure the 
detection of twinning, and sometimes a crystal has both pseudo-symmetry with 
twinning as well as tNCS.

If you want to send me a Phaser log file (off-list), I might be able to make 
some suggestions about what to consider.

Best wishes,

Randy Read

> On 8 May 2018, at 16:02, 苏纪勇 <sujy...@nenu.edu.cn> wrote:
> 
> Dear Herman, I tried phaser. It worked. But there are a lot of crashes in the 
> structure. Meanwhile, R factors could not be lowered. I tried to use P1 to 
> process the data, but I do not know how many monomers in the asymetric unit. 
> Bests, JiYG 
> On 05/08/2018 22:50, herman.schreu...@sanofi.com 
> <mailto:herman.schreu...@sanofi.com> wrote:
> Dear JiYG,
>  
> Unless the tNCS has caused processing problems, Phaser should automatically 
> deal with tNCS and I would recommend to just give it a try. If it fails, you 
> could try more sophisticated approaches.
>  
> Best,
> Herman
>  
> Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von ???
> Gesendet: Dienstag, 8. Mai 2018 16:41
> An: CCP4BB@JISCMAIL.AC.UK
> Betreff: [EXTERNAL] [ccp4bb] tNCS problem
>  
> Dear all, Recently, I collected a data set of a crystal of a big protein, 
> which contains 1000 amino acids. I processed the data to P2. But the data set 
> has tNCS problem. I want to do molecular replacement. Is there anyone know 
> how to deal with this problem? Thanks, JiYG
>  
>  

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

[ccp4bb] Call for targets (not-yet-released structures) for CASP13 modelling experiment

[Randy Read]

Posted on behalf of the CASP organizing committee:


Request for protein structures to test modeling methods - CASP13

CASP community experiments aim to establish and advance the state of the art in 
protein structure modeling.

To this end, CASP collects information on soon-to-be released experimental 
structures for a three month season every two years. Sequences and other 
information on these targets is distributed to the protein modeling community, 
and the resulting models assessed by independent experts in the field. Results 
are published in special issues of the journal PROTEINS (see all about CASP12 
at https://onlinelibrary.wiley.com/toc/10970134/86/S1). About 100 modeling 
groups from around the world take part.

The success of CASP depends completely on the generosity of the experimental 
community in providing modeling targets. We are now requesting targets for the 
CASP13 experiment, which will launch at the beginning of May. So, if you have 
anything suitable, we would be most grateful if you would go to the target 
entry page:
http://www.predictioncenter.org/casp13/targets_submission.cgi

The CASP community needs modeling targets over a wide range of difficulty, for 
modeling with and without the aid of templates. We are particularly interested 
in membrane proteins, protein complexes, and cryo-EM structures.  We are also 
extending CASP to include more modeling assisted by sparse experimental data, 
in collaboration with experimental groups in NMR, SAXS, chemical crosslinking, 
and FRET. For that, protein material is needed (of course this is not expected 
for most targets, but if its available, it would be much appreciated!).

For those of you who have not provided targets to CASP before, the procedure is 
simple – there is a web page for submitting targets, and an experienced staff 
to deal with any queries. We don’t need the structure in advance of its release 
by the PDB, and if we are notified early enough (a minimum of three weeks 
before release, more is better) there need be no delay in structure release. 
More details are available on the web site.

Thanks,
CASP organizing committee:
John Moult, University of Maryland, USA
Krzysztof Fidelis, University of California, Davis, USA
Andriy Kryshtafovych, University of California, Davis, USA
Torsten Schwede, University of Basel, Switzerland

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] According correct space group assignment...

[Randy Read]

I think that starting with a direct methods program like shelxt would be fun.  
If that doesn’t work, it could be interesting to try to solve it by molecular 
replacement with fragments varying from a tetraplex, a base pair or even just 
single bases.  (Assuming that Phoebe’s concern about twinning does not turn out 
to be correct…)

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 20 Apr 2018, at 20:09, Tim Gruene <tim.gru...@psi.ch> wrote:
> 
> Dear Rafal,
> 
> shelxt does not require the space group, it only needs the Laue group. If it 
> finds a decent solution, it'll find also the space group for you.
> 
> Best,
> Tim
> 
> On Friday, April 20, 2018 3:30:36 PM CEST Rafal Dolot wrote:
>> Dear CCP4BB,
>> 
>> I've recently collected data for 11mer build of DNA (9xG, 2xT). XDS, and
>> DIALS gave me similar solution - SG: I2(1)2(1)2(1) or I222, with cell
>> dimension 20.65, 22.96, 43.37, 90, 90, 90, what is too small for this
>> size of the molecule. 11mer is rich in G, so we expect the G-tetraplex
>> formation. Data were collected to almost 1 A, so it should be enough for
>> trials with direct methods/ab initio solution. What I should do first to
>> find correct SG and/or cell parameters?
>> 
>> Best regards,
>> 
>> Rafal
> 
> -- 
> --
> Paul Scherrer Institut
> Tim Gruene
> - persoenlich -
> OSUA/204
> CH-5232 Villigen PSI
> phone: +41 (0)56 310 5297
> 
> GPG Key ID = A46BEE1A

Re: [ccp4bb] Problems with pseudo-symmetry

[Randy Read]

Dear Renu,

If the intensity statistics (especially the L-test and Phaser's second moments 
tests after the tNCS correction) do not indicate twinning, then you probably 
would not be justified in assuming that the crystal is twinned, unless you find 
good evidence (e.g. from the symmetry of the MR solution in P1) that the 
crystal is pseudo-symmetric rather than exactly P21.

I think if you look closely at the xtriage output, you should find provisos 
warning you that the results from the tests using potential twin laws can 
indicate twinning when, in fact, the data are simpy merged in too low symmetry. 
 The tests using potential twin laws are based on the assumption that 
twin-related reflections should have independent intensities, which is not true 
*either* when there is twinning *or* when the data have been merged in too low 
symmetry and the reflections really are related by symmetry.

Best wishes,

Randy Read

> On 2 Feb 2018, at 15:27, Renuka Kadirvelraj <r...@ccrc.uga.edu> wrote:
> 
> Dear All,
> 
> I am trying to refine a 1.9 A structure that indexed with excellent 
> statistics (XDS) into a primitive monoclinic cell with cell lengths a=81.33, 
> b=90.11, c=113.25 and beta angle =102.15. Data analysis (Xtriage) indicated 
> strong pseudo-symmetry with an off-origin peak (53% of origin peak) at 
> fractional coordinates (0.07, 0.5, 0.19). The intensity statistics did not 
> indicate twinning and Pointless picked P2(1) as the space group with a 94% 
> probability. Molecular replacement (Phaser) gave a solution with 4 molecules 
> in the asymmetric unit with appropriate packing of symmetry-related molecules 
> and NCS-restrained refinements with TLS led to very good quality maps. 
> However, the Rfree for the completed model has stalled at 28% (Rwork is at 
> 24%).
> 
> I could not solve the structure in P2. Data processed in P1 has almost the 
> same dimensions as the monoclinic cell (a=81.37, b=90.14, c=113.29) and MR 
> with Phaser located 8 molecules in the asymmetric unit. The data scaled in P1 
> also has translational pseudo-symmetry at (0.06, 0.5, 0.185). Intensity 
> statistics do not indicate twinning but Xtriage detects the two-fold as a 
> pseudo-merohedral twin (-h,k,-l) andsuggests the likely point group is P2. 
> Initial refinement indicates the Rfree will likely remain high for the 
> structure in P1 as well (Rfree ~37% and Rwork at 34%).
> 
> I would really appreciate advice regarding the stalled Rfree and finding the 
> true space group. 
> 
> Many thanks,
> Renu
> 
> -
> Renu Kadirvelraj
> Biochemistry and Molecular Biology
> 120, Green Street
> University of Georgia
> Athens, GA 30605
> Tel: (706) 583 0303
> r...@ccrc.uga.edu <mailto:r...@ccrc.uga.edu>
--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] coordinate transformation

[Randy Read]

Hi,

In principle, when you solve a structure by molecular replacement with Phaser, 
you should end up with a reasonably compact assembly near the origin of the 
unit cell.  However, I usually use coot to shuffle around the symmetry copies 
(one of my favourite features!) to make the most sensible arrangement, 
particularly when there are multimers with NCS symmetry.

Unfortunately, Phaser has had a bug in this feature for the past year or so, as 
Robbie Joosten kindly pointed out to us recently.  We've now fixed this bug, 
and a corrected version of Phaser will appear soon in a new Phenix release and 
hopefully not too much later as a CCP4 update.  In the meantime, it would be a 
good idea to check any molecular replacement solutions you generated in Phaser 
over the last while and choose better symmetry copies if the solution has ended 
up far from the origin.

Sorry for the trouble!

Best wishes,

Randy Read

> On 14 Dec 2017, at 08:39, Kajander, Tommi A <tommi.kajan...@helsinki.fi> 
> wrote:
> 
> Dear Paul, 
> 
> Yes thank you. This was the best answer i think. Someone else already also 
> suggested that also.  Coot is very handy indeed.
> 
> (Would still be curious of knowing how to find the "closest to origin" copy 
> otherwise - but this solves my problem)
> 
> Thanks to all who responded.
> 
> Cheers,
> Tommi
> From: CCP4 bulletin board <CCP4BB@JISCMAIL.AC.UK> on behalf of Paul Emsley 
> <pems...@mrc-lmb.cam.ac.uk>
> Sent: Thursday, December 14, 2017 3:25:19 AM
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] coordinate transformation
>  
> On 13/12/2017 13:50, Kajander, Tommi A wrote:
> > Hello,
> > 
> > If someone could point this out would be very helpful... Wasnt there a 
> > simple script somewhere that would 
> > transfer coordinates close to origin - if they for some reason are not? 
> > Just cant find anything right away. 
> > 
> 
> At the risk of not answering the question because it's not a simple script, 
> my I recommend Coot?
> 
> File -> Open -> yourcoords.cif
> Draw -> Cell & Symm -> Master Switch -> Yes
> Show Unit Cells -> Yes
> OK
> Drag the View to the Origin # it's marked with an "O"
> Middle-mouse click on an Symmetry-related Atom # that's close to the origin
> Extensions -> Modelling -> Symm Shift Reference Chain Here

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] P212121 twinning

[Randy Read]

Hi,

In the presence of tNCS, the normal moments tests can indeed be difficult to 
interpret: twinning makes the 2nd moments smaller, whereas tNCS makes them 
bigger.  However, you can use Phaser to compensate for this effect, by 
characterising the tNCS and accounting for its statistical effect on the 
distribution of intensities.  In the case of 2-fold tNCS, this usually works 
pretty well and both the cumulative distributions and the 2nd moments can then 
reveal the presence of twinning.  This would be worth trying, and it works 
better than trying to test the distributions for a supercell.  Note that you 
want to provide the data in terms of intensities, not amplitudes, for this to 
work best.

The L-test should be reasonably unaffected by twinning and does, in this case, 
seem to suggest at least partial twinning.

If you have a really good model, I would suggest trying molecular replacement 
in P1 first, because if you get a clear solution the symmetry of the solution 
will tell you what the true space group is.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 3 Dec 2017, at 19:52, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Certainly you may have a lower symmetry spacegroup, but with NC translation 
> the twinning tests are a bit uncertain.
> 
> Your moments seem too high, whereas twinning usually makes them too low.  But 
> we need to see the plots v resolutio which should be pretty linear.. 
> Sometimes these go haywire at higher resolution.. 
> 
> 
> And also the spacegroup determination - absences along a* and C*could be 
> caused by the translation (1/2,0.4,1/2) so you need to test P2 21 2 and P 21 
> 21 2 and P 2 21 21 etc..
> 
> This is messy but you can test the data h'=h/2 k l'=l/2 ( Ie for a cell 
> 2a,b,2c and see if that seems twinned
> 
> (Use reindex and requset
> h/2,k,l/2 as the reindexing operator then input it to ctruncate ..) 
> 
> 
> 
> And of course if th integration is a bit ropy you can get strange effects.. 
> Do you see the same results for all crystals?
> 
> Eleanor
> 
> On 3 December 2017 at 18:42, Kay Diederichs <kay.diederi...@uni-konstanz.de> 
> wrote:
> Hi Carmela,
> 
> as Jacob suggests, the real space group may be P2 or P2(1). You need to 
> prepare 3 MTZ files, each corresponding to a different unique b axis.
> So just run XDS 3 times, each time with JOB=CORRECT and SPACE_GROUP_NUMBER=3, 
> and afterwards XDSCONV to create a MTZ file (or pointless/aimless).
> 
> First time:
> UNIT_CELL_CONSTANTS= 58 103 220 90 90 90
> Second time:
> UNIT_CELL_CONSTANTS= 103 220 58 90 90 90
> Third time:
> UNIT_CELL_CONSTANTS=220 58 103 90 90 90
> 
> Then, if you have a good model, do  MR with each MTZ file, taking care that 
> both P2 and P2(1) are tested. Or experimental phasing, but that requires good 
> data.
> 
> HTH,
> Kay
> 
> On Sun, 3 Dec 2017 18:29:57 +0100, Carmela Garcia <c.gar...@bioc.uzh.ch> 
> wrote:
> 
> >Hi,
> >
> >The dimensions for a native are 58 103 220, with small differences for the 
> >derivatives.
> >
> >Best,
> >
> >Carmela.
> >
> >> On 3 Dec 2017, at 18:18, Sridhar Prasad <spra...@calasiapharma.com> wrote:
> >>
> >> Hello,
> >>   Can you please share the unit cell dimensions.
> >>
> >> Cheers,
> >> Sridhar
> >>
> >> On Dec 3, 2017 9:13 AM, "Carmela Garcia" <c.gar...@bioc.uzh.ch 
> >> <mailto:c.gar...@bioc.uzh.ch>> wrote:
> >> Dear all,
> >>
> >> I know that some years ago a similar situation was discussed here, and I 
> >> wonder if someone has new insights about these problems.
> >>
> >> My protein is a dimer in solution. I tried several derivatives for SAD, 
> >> and all my datasets seem to have the same problem, including the native 
> >> crystals. I processed the data with XDS and the space group determination 
> >> was done with Pointless, being a P212121.
> >>
> >> Checking the quality of the data, I found several problematic results:
> >>
> >> - Translational NCS is detected. There is a peak at (0.50, 0.40, 0.50)
> >> - The L test suggests twinning (L statistic = 0.41)
> >> - The mean acentric moments I from input data have the following values:
> >>  <I^2

Re: [ccp4bb] Dubious slide aggregator slideplayer.com

[Randy Read]

Dear Robbie,

Thanks for posting that!  When I share my slides, I generally try to do so only 
as PDF files to make it harder for anyone to reuse them without permission, but 
I hadn’t realised how much of an issue this is.  It seems that I’ve nonetheless 
let one PowerPoint into the wild (“Using molecular replacement to exploit 
multiple crystal forms”).  Searching for its title in Google, I’ve found copies 
on not only slideplayer.com but also on documents.tips (aka DocSlide), 
slideserve.com and studylib.net.  

Best wishes,

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 30 Oct 2017, at 08:46, Robbie Joosten  wrote:
> 
> Dear CCP4BBers,
>  
> Yesterday I ran into one of my talks at Slideplayer.com, a slide aggregator 
> that makes money from advertising. You can watch presentations for free 
> (surrounded by ads), but to download them you have advertise for Slideplayer 
> by posting on social media.
> The presentations are seemingly uploaded by users, but as their names give no 
> real hits on Google, I assume these are proxies to abuse the protection that 
> the DMCA provides to aggregators where users can post stuff. This probably 
> why they are officially based in Virginia (but their communication is partly 
> in Russian). The presentations are just spidered from public websites.
> There are many talks from CCP4bb members on Slideplayer.com and I don’t like 
> it. There are already enough commercial enterprises that make money from our 
> scientific works, but at least we get credit for it. That is not the case 
> here.
> The good news is that if you send a copyright infringement notice they do 
> remove the presentations (while hiding behind the DMCA of course). Because of 
> US copyright laws, you have to actively protect your copyright. Below is a 
> template notice that worked for me.
>  
> All the best,
> Robbie
>  
>  
> Template:
> //
> Dear Slideplayer developers, 
>  
> This writing is to inform you that I reserve and own all rights to the 
> presentation titled "$TITLE". It has come to my attention that you have made 
> unauthorised use of this copyrighted work (http://slideplayer.com/slide/$ID) 
> by duplicating and sharing it on Slideplayer. 
> As you have not sought or requested authorisation to distribute this work, 
> you are hereby ordered to cease and desist any and all acts of unlawful 
> copyright infringement with respect to my work immediately. Failure to comply 
> with this notice will leave me no other option than to seek compensation for 
> your copyright infringement through legal action.
>  
> Sincerely, 
> $SIGNATURE

Re: [ccp4bb] High R/Rfree after MR

[Randy Read]

Just to add to this point.  The MR algorithms in Phaser are now able to make 
better use of intensity data, which is particularly important when you have any 
very weak data.  Having weak data can’t be avoided when you have serious 
anisotropy (or tNCS or a combination of the two).  Unfortunately, if you use 
amplitudes that have been through the French & Wilson (truncate) algorithm, the 
real variation in intensity is partially masked because the posterior amplitude 
values are computed on the prior assumption that all the reflections in a 
resolution shell have the same underlying intensity distribution.  

The UCLA server actually uses Phaser under the hood — what they add is to turn 
the anisotropic B-values into suggested resolution limits in the different 
directions.  However, I don’t think they allow you yet to submit intensities, 
which would be better.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 13 Oct 2017, at 10:29, vincent Chaptal <vincent.chap...@ibcp.fr> wrote:
> 
> Dear Gia and Paul, 
> 
> about anisotropy, one point to keep in mind is that it is not necessarily 
> linked to the difference in resolution limits.
> In fact I am at the moment working on one of these cases, with extremely 
> large difference in resolution limits, but relatively low anisotropy. 
> Anisotropy is more a deviation from "normal" intensity falloff as a function 
> of resolution. There is a thin difference/relationship between the two 
> concepts that I think is worth investigating. 
> 
> we have performed a statistical analysis of this phenomenon and the paper is 
> in revision at the moment, but if you want to know where your anisotropy 
> stands in respect to all the other PDBs out there, feel free to contact me 
> off list. 
> You mention MR: Phaser calculates the anisotropy so you can find the value in 
> the first lines of the log. 
> 
> Staraniso or the UCLA server are good to test if you have anisotropy. 
> Staraniso has a newer way of dealing with intensity falloff and accounting 
> for it. 
> 
> All the best
> Vincent
> 
> 
> 
> 
> On 13/10/2017 10:58, Paul Miller wrote:
>> I had a similar problem to what you describe. In my case the dataset was 
>> severely anisotropic (2.6A, 3.4A, 4.6A in each axis). The R factors were 
>> stuck similar to yours but the map looked good. I was told by someone with a 
>> much better appreciation of the theory than myself that the anisotropy was 
>> causing the problem.
>> 
>> It would be interesting to know from an expert in anisotropy e.g. the 
>> creators of UCLA anisotropy server or Startaniso whether anisotropy can 
>> cause this problem and whether there is any way around it.
>> 
>> Cheers, Paul
>> 
>> Paul Steven Miller (PhD)
>> Postdoctoral Researcher
>> University of Oxford
>> Wellcome Trust Centre for Human Genetics
>> Division of Structural Biology
>> Roosevelt Drive
>> Oxford
>> OX3 7BN
>> 
>> 
>>  Original message 
>> 
>>> Date: Fri, 13 Oct 2017 10:30:22 +0200
>>> From: CCP4 bulletin board 
>>> <CCP4BB@JISCMAIL.AC.UK> (on behalf of Gianluca Cioci 
>>> <gianluca.ci...@gmail.com>
>>> )
>>> Subject: [ccp4bb] High R/Rfree after MR  
>>> To: 
>>> CCP4BB@JISCMAIL.AC.UK
>>> 
>>> 
>>>   Dear All,
>>>   I am trying to refine a structure at 3.3A. Model has
>>>   60% identity to the target. Maps look OK (for 3.3A)
>>>   and rebuilding in Coot is relatively
>>>   straightforward. However, after some rebuilding
>>>   cycles the R factors are stuck at 0.37/0.39
>>>   (REFMAC). 
>>>   XTRIAGE tells me that everything is normal (no twin,
>>>   98% completeness, R=3.5% in the low resolution bin),
>>>   perhaps some anisotropy is present. 
>>>   I have already refined 2 homologous structures at
>>>   resolutions going from 3.2 to 3.8 and there were no
>>>   problems (final R ~ 0.21/0.24). 
>>>   Any advice ?
>>>   Thanks,
>>>   GIA
>>> 
> 
> -- 
> Vincent Chaptal, PhD
> Institut de Biologie et Chimie des Protéines
> Drug Resistance and Membrane Proteins Laboratory
> 7 passage du Vercors 
> 69007 LYON
> FRANCE
> +33 4 37 65 29 01
> http://www.ibcp.fr
> 
> 

Re: [ccp4bb] doubt regarding MR search model

[Randy Read]

Dear Emily,

In this case, I was referring to the difference between 1065 obtained placing 4 
copies of the monomer and 722 obtained placing 2 copies of the dimer.  A 
difference of over 300 is substantial when we feel that an increase of 60 is 
enough to be reasonably convinced that a solution for placing one copy of a 
molecule is correct.  I guess that this value of 60 would be my yardstick for 
“significant” in the sense of “big enough to matter”.  Putting a number to 
“significant” in the sense of “bigger than random” is harder.  There’s an 
argument that the random error in an LLG value is roughly equivalent to the 
square root of the LLG, which would still make this a significant difference.  
That argument could get much more complicated if you started to worry about 
things like the 2-dimer model just being a constrained version of the 4-monomer 
model because the monomers within the dimer are identical!

Not sure if that really helps!

Randy
 
-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 18 Sep 2017, at 17:39, Emilia C. Arturo (Emily) <ecgart...@gmail.com> 
> wrote:
> 
> Hello Randy,
> 
> I'm chiming in about the last sentence in your reply:
> 
> 
> Finally, I would suspect that getting a significantly lower LLG for two 
> copies of a dimer means that the dimer in your structure is slightly 
> different from the dimer in the model.
> 
> Will you please be more specific about what you consider "significantly 
> lower" (or higher) for LLG scores? And how does this difference translate to 
> differences among TFZ scores? I have wondered this especially when a Phaser 
> MR result from use of a smaller part (i.e. one domain, 64% of the sequence) 
> of a structure differs by a few units relative to a result from use of a 
> larger part (i.e. two domains, 90% of the sequence) of the protein; the 
> difference in my case was a top TFZ score of 42.5 using the one-domain search 
> model, but a lower TFZ score of 36.9 for the two-domain model. 
> 
> Even more interesting to me is when the TFZ scores vary by quite a bit more 
> between two phaser runs that used two different, but 100% sequence identical, 
> one-domain search models; in this case the top TFZ scores were 18.7 versus 
> 42.5, for the two models that differed only in the coordinate position of ~20 
> out of 290 residues (a lid for the enzyme). 
> 
> Regards,
> 
> Emily.
> 
> 
> Best wishes,
> 
> Randy Read
> 
> -
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical ResearchTel: +44 1223 336500
> Wellcome Trust/MRC Building Fax: +44 1223 336827
> Hills RoadE-mail: 
> rj...@cam.ac.uk
> Cambridge CB2 0XY, U.K.   
> www-structmed.cimr.cam.ac.uk
> 
> > On 18 Sep 2017, at 16:24, Andrew Leslie <and...@mrc-lmb.cam.ac.uk> wrote:
> >
> > Regarding the warning message "The top solution from a TF rescoring did not 
> > pack”, I get this on all the PHASER jobs that I have run recently, but 
> > looking through the PHASER log file I cannot see any evidence for packing 
> > failure.
> >
> > It may be that the failure is buried in an obscure place in the very long 
> > log file, but if I search for “pack” then I get the output summarised 
> > below, all of which, apart from the ADVISORY, suggests to me that there is 
> > no problem at all with the packing.
> >
> > Perhaps someone on the PHASER team can cast light on this.
> >
> > Andrew
> >
> >
> > Extracts from PHASER log file:
> >
> >
> > TRANSLATION FUNCTIONS
> > -
> >
> >Target Function: FAST LETF1
> >Translation Packing Function applied: top peak will pack
> >Translation Packing Cutoff: 50%
> >Sampling:  1.82 Angstroms
> >
> >
> > ===
> > 
> > PACKING FUNCTION
> > 
> >
> >There are 4 solutions to pack
> >Packing analysis
> >0%100%
> >|=| DONE
> >
> > 
> >Packing Table
> >-
> >Solutions accepted if pairwise clashes less than 10 % of trace atoms
> >#in  #ou

Re: [ccp4bb] doubt regarding MR search model

[Randy Read]

Yes, I can see how that message would be confusing!

This comes from a new feature.  Phaser now does a relatively liberal packing 
test as part of rescoring the fast translation results with the full likelihood 
function.  Peaks that are rejected at this stage aren’t counted for the top 
peak height used to define the cutoff for what is stored for the proper packing 
test and eventual rigid-body refinement.  This turns out to be very useful for 
models that are small fragments, particularly helices in Arcimboldo, where the 
initial searches tend to put molecules on top of symmetry axes.

The confusing part is that you get the advisory even if the particular 
translation search where a peak was rejected didn’t even yield a score as high 
as the one that ended up being chosen.

The bit in the log file about increasing the composition suggests that what is 
in your model has more scattering matter than what you defined for the 
corresponding part of the structure in the composition definition.  If the 
change wasn’t big, you don’t need to worry, but if there was a big discrepancy 
it’s worth sorting out where it comes from.

Finally, I would suspect that getting a significantly lower LLG for two copies 
of a dimer means that the dimer in your structure is slightly different from 
the dimer in the model.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 18 Sep 2017, at 16:24, Andrew Leslie <and...@mrc-lmb.cam.ac.uk> wrote:
> 
> Regarding the warning message "The top solution from a TF rescoring did not 
> pack”, I get this on all the PHASER jobs that I have run recently, but 
> looking through the PHASER log file I cannot see any evidence for packing 
> failure.
> 
> It may be that the failure is buried in an obscure place in the very long log 
> file, but if I search for “pack” then I get the output summarised below, all 
> of which, apart from the ADVISORY, suggests to me that there is no problem at 
> all with the packing.
> 
> Perhaps someone on the PHASER team can cast light on this.
> 
> Andrew
> 
> 
> Extracts from PHASER log file:
> 
> 
> TRANSLATION FUNCTIONS
> -
> 
>Target Function: FAST LETF1
>Translation Packing Function applied: top peak will pack
>Translation Packing Cutoff: 50%
>Sampling:  1.82 Angstroms
> 
> 
> ===
> 
> PACKING FUNCTION
> 
> 
>There are 4 solutions to pack
>Packing analysis
>0%100%
>|=| DONE
> 
> 
>Packing Table
>-
>Solutions accepted if pairwise clashes less than 10 % of trace atoms
>#in  #out Clash-% Symm TF-SET  ROT TFpk#TFTFZSpaceGroup 
>1Top1  0.741  --   1 11325.94 26.03  P 1
>22 1.058  --   1 21293.29 24.34  P 1
>33 0.952  --   1 31245.70 21.88  P 1
>44 0.847  --   1 41211.49 20.11  P 1
> 
>4 accepted of 4 solutions
>   4 pack of 4 accepted solutions
> 
> ==
> ---
> TRANSLATION PACKING
> ---
> 
>Translation Packing Cutoff: 50%
>All solutions have been packed
> 
> ==
>Packing Fast Search Translations...
>   9871 peaks
>   500 peaks over 1049.74 checked for packing
>   Translation peak 1 first to be kept
>   Done
> 
> 
>New Top Packing Fast Translation Function FSS = 2181.37 (TFZ=46.3) at 
> Trial #1
> 
> 
>Top Peaks Without Clustering
>Select peaks over 67.5% of top (i.e. 0.675*(top-mean)+mean)
>There was 1 site over 67.5% of top
>1 peak selected
>The sites over 67.5% are:
># Frac X Frac Y Frac ZFSS   Z-score
>1  0.719  0.122  0.890 2181.4 46.35
> 
>Top 1 translations before clustering will be rescored
>Calculating Likelihood for TF SET #1 of 1 TRIAL #1 of 1
>0% 100%
>|==| DONE
> 
>Packing LLG Translations: pass 1 of 11...
>   1 peaks
>   No peaks over 1878.46 - no packing check
>Packing LLG Translations: pass 2 of 11...
>   1 peaks
>   1 peaks over 1878.46 checked for packing
>   Translation peak 1 first to be ke

Re: [ccp4bb] NMR or Homology Model as a MR model

[Randy Read]

Hi,

Molecular replacement gets progressively more difficult as the sequence 
identity drops, though the correlation between sequence identity and model 
quality is not perfect.  With the best models having sequence identities around 
24%, success is by no means guaranteed, so you might have to try lots of 
strategies or resort to experimental phasing.  NMR models do tend to be worse 
for molecular replacement, so that will probably lower the chances of success 
further.

In addition to what others have said already, it's worth mentioning that, in a 
number of very difficult structure determinations (best models around 20% 
sequence identity), the key thing that made the difference between failure and 
success was pruning off the parts of the ensemble that do not agree among the 
members of the ensemble, leaving just the conserved core.  This can be done, 
for instance, in the Ensembler program with the "trim=True" flag.  For this to 
work best, it's helpful if you happen to have a number of potential models that 
are also relatively distantly related to each other, to really highlight what 
is conserved in the fold.

Best wishes,

Randy Read

> On 29 Aug 2017, at 12:42, Andreas Forster <docandr...@gmail.com> wrote:
> 
> Dear Nishant,
> 
> Rosetta is a good suggestion.  You can also use an ensemble of several 
> related (superposed) structures as your search model.  This will improve your 
> chances of success.
> 
> All best.
> 
> 
> Andreas
> 
> 
> 
> On Tue, Aug 29, 2017 at 12:50 PM, Nishant Varshney <nik...@gmail.com 
> <mailto:nik...@gmail.com>> wrote:
> Dear Crystallographers,
> 
> I am working to solve an human protein structure which has a domain sequence 
> identity of 24% with domain of another protein.  As Phaser as well as Molrep 
> failed to give any definite solution (TFZ=3.7 from MR), I want to ask, if 
> solution structure of another protein having domain sequence similarity of 
> 25% or a homology model can be used as a template for MR?
> Many thanks and Regards
> Nishant
> 
> -- 
> Dr. Nishant Kumar Varshney,
> Research Associate,
> C/O Dr. Sameena Khan,
> Drug Discovery Research Center,
> Translational Health Science and Technology Institute (THSTI)
> NCR Biotech Science Cluster,
> 3rd Milestone, Faridabad – Gurgaon Expressway,
> Faridabad – 121001 (HARYANA), India
> Ph: +91- 0129-2876477 <tel:+91%20129%20287%206477>
> Mob: 8390564690
> 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] ample

[Randy Read]

Dear Patrick,

Apologies for a partly non-CCP4 answer!  You can use the HHPRED alignments in 
some of the software available in Phenix.  One option is to provide the .hhr 
file to MRage, which will fetch the PDB entries, run Sculptor to modify the 
templates with one or more protocols, based on that alignment, and then test 
all the models that have been generated.  That's fully automated and will often 
work.  If it doesn't, an alternative is to run phenix.mr_model_preparation, 
which again will fetch the PDB entries and run the single best Sculptor 
protocol on all of them.  I tend to prefer this option myself, because then you 
can look at whether it would be sensible to combine various alternative models 
as ensembles.  One of the most powerful approaches is to run Ensembler 
(available in both CCP4 and Phenix, as is Sculptor) to make an ensemble, with 
the "trim=True" option turned on to trim any bits that don't agree among the 
structures in the ensemble.  Cutting back to a conserved core has allowed a 
number of very recalcitrant structures to be solved with collections of models 
around the 20% sequence identity range.

Best wishes,

Randy

> On 20 Jul 2017, at 23:13, Patrick Loll  wrote:
> 
> I’m intrigued by the prospect of using AMPLE to test multiple distant 
> homologs in a MR problem. I’ve used HHPRED to identify about 20 
> high-probability homologs of known structure, each of which has about 20-25% 
> identity with the unknown protein. However, it’s not clear to me from the 
> documentation whether the program will use the alignments from HHPRED, and, 
> if so, how I should provide that information. 
> 
> Or does AMPLE perform its own alignment? I.e., do I simply point the program 
> to a directory containing 20 different PDB files and stand back?
> 
> Thanks for any insights.
> 
> Cheers,
> 
> Pat 
> ---
> Patrick J. Loll, Ph. D.  
> Professor of Biochemistry & Molecular Biology
> Drexel University College of Medicine
> Room 10-102 New College Building
> 245 N. 15th St., Mailstop 497
> Philadelphia, PA  19102-1192  USA
> 
> (215) 762-7706
> pjl...@gmail.com
> pj...@drexel.edu

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] What are acceptable Rwork/Rfree for publication

[Randy Read]

Dear Gerard,

I was puzzled by the statement that the UCLA anisotropy server characterises 
anisotropy in terms of a combination of effects restricted to lie along the 
crystallographic axes.  That server is built on the anisotropy correction 
algorithm in Phaser, and from the beginning Phaser has used an anisotropic 
tensor that is constrained only by the Laue symmetry.  I've been in touch with 
Mike Sawaya to clarify whether they were doing anything different.  Perhaps he 
will comment directly, but he said that their server indeed uses the general 
ellipsoid for the anisotropic truncation.  The confusion arises from a 
simplified description on their results page, which describes the principal 
axes in terms of which reciprocal cell edge is closest.

In seeking to clarify exactly what the server is doing, I attempted to track 
down the example underlying the illustration on the STARANISO web page.  I 
believe that this is PDB entry 5j1i, but it is not possible to repeat the 
calculations with the data deposited in the PDB because what has been deposited 
appears to be the anisotropy-corrected, truncated data used in refinement.

This brings up an important point.  I am confident you would agree with me 
that, in cases such as this, the uncorrected, untruncated data should always be 
deposited as well as any massaged data used for refinement.  Once an anisotropy 
correction has been applied, the systematically weak data may not contribute 
any useful information to the refinement, but those data provided essential 
information to the characterisation of anisotropy.  Without access to those 
weak intensities, it becomes impossible to make use of new information (e.g. 
from the atomic model) or to apply any new, improved algorithms for analysing 
anisotropy.  Similar issues are even more important for the treatment of the 
systematically weak intensities arising from translational non-crystallographic 
symmetry (tNCS).

By the way, we have a rather different reflection-by-reflection approach to 
data truncation in Phaser, implemented last year but not yet published.  After 
characterising the anisotropy and (if present) tNCS, we determine how much 
information (measured in bits) each intensity measurement contains, relative to 
what was expected from the Wilson distribution.  By default, reflections that 
contribute less than 0.01 bits of information are omitted from any subsequent 
calculations, though we do not truncate them from the output reflection list.  
In some data sets with severe anisotropy and/or tNCS, over 40% of the 
reflections are omitted.  Importantly, we could not have characterised the 
anisotropy and tNCS without having those extremely weak measurements included 
in the data set!  

We feel that the reflection-by-reflection approach has the advantage that the 
significant number of reflections with some signal that will be found beyond 
any hard limit can still be included, without a large increase in computational 
cost.

It's also worth reminding the community that if the French & Wilson algorithm 
has been applied without accounting for anisotropy and tNCS, the resulting 
amplitudes will significantly over-estimate the systematically weak 
reflections.  That's another reason to strongly encourage everyone to deposit 
their data in terms of intensities and their standard deviations (as discussed 
here: http://scripts.iucr.org/cgi-bin/paper?S2059798315013236 
).

Best wishes,

Randy

> On 18 Jun 2017, at 13:58, Gerard Bricogne  wrote:
> 
> Dear Khoa,
> 
> You are asking a very pertinent question, that will resonate in
> (too) many people's minds.
> 
> Somehow anisotropy is very much "an inconvenient truth" in MX, as
> many things have long been set up and operated on the basis of having
> resolution be a single number and of all statistics being calculated
> in spherical shells, throwing in everything that has an HKL associated
> with it. We are repeatedly told that resolution *has* to be a single
> number, because the title of a Nature paper only has room for one
> number - if that isn't a case of the tail wagging the dog, I don't
> know what is ;-) . 
> 
> There are more serious contexts in which the requirement for a
> single number can be a more complex issue, such as contractual
> arrangements where the achievement of certain milestones and the 
> remuneration attached to them depend on reaching a certain resolution.
> Here, whoever writes and/or accepts resolution criteria of this kind
> will simply have to undergo further education and learn about that
> inconvenient fact called anisotropy. Probably - especially in the case
> of membrane protein structures - it is the highest resolution limit
> that matters; but you can't have good completeness to that highest
> resolution, by the very definition of anisotropy! Therefore, some
> quality criteria that are compatible in the case of 

Re: [ccp4bb] Se-Met and Se-Cys double labelling

[Randy Read]

Yes, I agree that MR is worth a shot, though depending on the resolution your 
life may be vastly easier with experimental phases!  We've had several cases 
where the following kind of procedure works: find related structures with a 
very sensitive homology search (we like HHPRED, though other options probably 
work), use the corresponding sequence alignment to prune the models (with 
sculptor or chainsaw or other tools) and make a trimmed ensemble with 
ensembler.  The last step can be very important, in trimming off any bits of 
the collection of models that do not constitute a conserved core.  Then provide 
the ensemble to Phaser as a molecular replacement model.  The trimmed ensemble 
is usually the best model, in our experience, but one of the individual models 
may be better in some cases, so that's also worth testing.

Also, the MR-Rosetta pipeline has had a substantial number of successes when 
the highest sequence identity was in the range of 15-25%.

Best wishes,

Randy Read

> On 21 Jun 2017, at 17:08, Mark J van Raaij <mjvanra...@cnb.csic.es> wrote:
> 
> If your data is good enough, your SeMets alone might well be enough.
> Soaking native crystals in Hg compounds may also work, avoiding SeMet 
> altogether. We have had a lot of success with methylmercury chloride binding 
> to free Cys. You may have to experiment with different soaking times and 
> protocols.
> Finally, don't give up on MR too early, what matters is the structural 
> similarity, not directly the sequence identity. We've had success once with 
> 19% identity.
> Native protein may be much easier to produce than the SeMet and SeMet/SeCys 
> versions (and may differ a lot between proteins).
> 
> Mark J van Raaij
> Dpto de Estructura de Macromoleculas
> Centro Nacional de Biotecnologia - CSIC
> calle Darwin 3
> E-28049 Madrid, Spain
> tel. (+34) 91 585 4616
> http://wwwuser.cnb.csic.es/~mjvanraaij 
> <http://wwwuser.cnb.csic.es/~mjvanraaij>
> 
>> On 21 Jun 2017, at 17:46, Vito Calderone <calder...@cerm.unifi.it 
>> <mailto:calder...@cerm.unifi.it>> wrote:
>> 
>> I am working on a protein having 360 residues. In its sequence there are 3
>> Met and 5 free Cys.
>> I will need MAD to solve the structure since based on the sequence the
>> closest homologue has 20% identity匢 suppose MR would be very unlikely to
>> work卻o I would like to express a selenium derivative to exploit MAD.
>> Looking in the literature 1 Se-Met every 120 residues seems not to comply
>> the threshold to get a good anomalous signal. For this reason I would like
>> to exploit both Met and Cys so I would have 8 seleniums per 360 residues.
>> Could somenone suggest a reference to a protocol to express the double
>> mutant protein in NON auxotrophic strains of E. coli which you have
>> experienced working efficiently?
>> Thanks
> 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] No improvement in R-factor after Refmac.

[Randy Read]

Hi,

I was just going to make the same point!  The only thing to add is that, if 
there really is translational NCS (which is certainly possible with 4 copies in 
the a.u.), then it’s essential both to account for it (which current versions 
of Phaser should do automatically, if you search for all 4 copies in one job) 
and to try all possible space groups.  The situation Craig describes, in which 
it’s not immediately obvious whether your crystal has a crystallographic 2(1) 
and a pseudosymmetric non-crystallographic 2-fold or the reverse, is not 
uncommon.  However, we’ve found that the likelihood score accounting for the 
effect of tNCS is pretty good at discriminating the two possibilities.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 18 Mar 2017, at 06:12, CRAIG A BINGMAN <cabing...@wisc.edu> wrote:
> 
> You really need to approach such situations with caution.  Examination of the 
> relatively small number of axial reflections probably show that there might 
> be twofold screw axes in all three directions.  But a non-crystallographic 
> microscopic translation of nearly 0.5 in the direction of a crystallographic 
> axis will give the same pattern of strong and weak reflections as a 
> crystallographic twofold screw axis.  If I were you, I would be very sure to 
> try molecular replacement in all possible orthorhombic space groups.  Several 
> programs, including Phaser, will organize that exhaustive search across all 
> eight possibilities for you.
> 
>> On Mar 17, 2017, at 11:56 PM, Polisetty Satya Dev <pvss...@gmail.com> wrote:
>> 
>> Hi,
>> 
>> We checked all possible space groups of orthorhombic crystal system using 
>> Scala and Pointless but the statistics show that P212121 is the possible 
>> space group.
>> 
>> Thank You,
>> Satya Dev
>> 
>> On Fri, Mar 17, 2017 at 8:03 PM, Teplyakov, Alexey [JRDUS] 
>> <atepl...@its.jnj.com> wrote:
>> Check the space group. It may be orthorhombic with a pure rotational axis 
>> (e.g. P21212) or even monoclinic.
>> 
>>  
>> 
>> From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of 
>> Polisetty Satya Dev
>> Sent: Friday, March 17, 2017 9:51 AM
>> To: CCP4BB@JISCMAIL.AC.UK
>> Subject: [EXTERNAL] [ccp4bb] No improvement in R-factor after Refmac.
>> 
>>  
>> 
>> Dear all,
>> 
>> I solved a structure at 2.0 A resolution with R-work and R-free values 0.25 
>> and 0.32 respectively and I am stuck at Refmac step where there is no 
>> further reduction in R-factor. 
>> 
>> The above stated values were obtained after several rounds of manual 
>> refinement followed by refmac. There are also areas where electron density 
>> is missing around peptide backbone in one of the monomer in ASU. 
>> 
>> Can anyone please tell me how can I improve the electron density and 
>> R-factor.
>> 
>> 
>> The structure solution was obtained using Phaser MR and here are the data 
>> statistics:
>> 
>>  
>> 
>> Average unit cell: 81.95, 100.40, 156.96,   90.00, 90.00, 90.00,
>> Space group: P212121,
>> Completeness 99.5,
>> Multiplicity 6.4,
>> Four monomers per ASU.
>> Solvent content: 47%.
>>  
>> Thank you everyone,
>> Satya Dev,
>> JNCASR.
>> 
> 

Re: [ccp4bb] Estimating the amount of missing electron density for a model

[Randy Read]

Hi,

I rarely disagree with Ethan, but there are in fact ways of getting some idea 
of how much of the ordered structure from your crystal is missing in your 
model.  It's not based on the difference between Fo and Fc but rather more on 
the correlation between them and how that varies as a function of resolution.  
How well this works depends very much on why your model is incomplete — in many 
circumstances Ethan will be right in practice!

Some time ago, I proposed the use of something called a SigmaA plot 
(http://www-structmed.cimr.cam.ac.uk/Personal/randy/pubs/a25349r.pdf).  SigmaA 
is a parameter that measures what fraction of the calculated structure factor 
is correlated to the observed structure factor in a particular resolution 
shell.  The correlation depends on both the fraction of the crystal that has 
been modelled and on the quality of the model.  Since the effect of errors in 
the model is greater at higher resolution, SigmaA is a function of resolution.  
The way it varies with resolution tells us something about the size of the 
errors in the model, and the limiting value at low resolution tells us 
something about the completeness of the model.  If we make some assumptions 
(most importantly that the errors are primarily coordinate errors that are 
drawn from the same isotropic 3D Gaussian distribution of errors for all atoms, 
and that the missing part of the structure has the same overall B-factor as the 
modelled part of the structure), then we expect a plot of the logarithm of 
SigmaA against 1/d^2 to be a straight line, where the slope is proportional to 
the mean-square coordinate error and the intercept is proportional to the 
logarithm of the model completeness.

By the way, the original SigmaA plot also assumed that the measurement errors 
in the diffraction data were negligible.  With the development of our new LLGI 
(intensity-based likelihood) target, we now have a way of correcting the 
estimation of SigmaA for the effect of errors, so we could do better if there 
was a resurgence of the SigmaA plot.  

We don't model bulk solvent nearly as well as ordered atoms, so these plots are 
not very linear at low resolution.  But the plots do tend to be surprisingly 
linear at higher than about 5-6A resolution.  If you're doing molecular 
replacement on a complex and you place a model for part of the complex, the 
intercept of a SigmaA plot computed before refining your model will usually 
give you a reasonable idea of how much is still missing.  Refining the model 
could make things worse because of the effects of overfitting your data.

However, the case of a complex was assuming that the missing part is as well 
ordered as the modelled part.  Frequently we have trouble completing a model 
because a domain or an NCS copy has much higher overall B-factors.  What that 
means is that the fraction of the scattering we can model also varies with 
resolution instead of being a constant, because the better ordered part 
accounts for an increasing fraction of the scattering at high resolution.  This 
violates the assumption about the coordinate error being the only thing that 
makes SigmaA vary with resolution.

So whether this might be useful to you will depend on where you are with the 
structure determination and why your model is incomplete!

Best wishes,

Randy Read

> On 21 Feb 2017, at 21:53, Hunter Moseley <hunter.mose...@gmail.com> wrote:
> 
> Is there a straight-forward way to estimate the amount of missing electron 
> density that a particular protein structure is missing based on the 
> difference between Fo and Fc?
> 
> It appears that the normalization of the Fc due to the employing of a maximum 
> entropy method that keeps Fo and Fc comparable to the standard deviation of 
> Fo would make this difficult. 
> Or am I missing something?
> 
> Cheers,
> Hunter
> 
> -- 
> Hunter Moseley, Ph.D. -- Univ. of Kentucky
> Associate Professor, Dept. of Molec. & Cell. Biochemistry / Markey Cancer 
> Center
> / Resource Center for Stable Isotope Resolved Metabolomics
>  Not just a scientist, but a fencer as well.
>My foil is sharp, but my mind sharper still.
> ---
> Email: hunter.mose...@uky.edu <mailto:hunter.mose...@uky.edu> (work)   
> hunter.mose...@gmail.com <mailto:hunter.mose...@gmail.com> (personal)   
> Phone: 859-218-2964 (office)   859-218-2965 (lab)   859-257-7715 (fax)
> Web: http://bioinformatics.cesb.uky.edu/ <http://bioinformatics.cesb.uky.edu/>
> Address: CC434 Roach Building, 800 Rose Street, Lexington, KY 40536-0093

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Phaser MR pairwise percent packing criterion

[Randy Read]

Hi,

The old approach of counting the number of clashes was changed to the 
percentage of trace atoms (typically CA atoms) involved in clashes largely 
because, as people like Eleanor pointed out to us, the number of clashes grows 
with the size of the problem and solutions can be rejected when the fraction of 
the structure involved in clashes is still very small.  The default now is to 
accept up to 10% of the trace atoms being involved in clashes.  So increasing 
that to a larger number should have the same basic effect as increasing the 
number of allowed clashes did in the old version.

If you want to get sophisticated, there's even another new feature, where you 
can use the ENSEMBLE…TRACE command to define a TRACE molecule.  That will be 
carried along with the actual search model and used to check for clashes.  This 
might be useful if, for instance, you think a domain might change  conformation 
slightly so you don't want it in the search model, but you want to exclude 
solutions that would create clashes with that domain.

Best wishes,

Randy Read

> On 17 Feb 2017, at 08:28, Xiao Lei <xiaolei...@gmail.com> wrote:
> 
> Dear CCP4bb members,
> 
> I found in the newer CCP4 Phaser MR version (I use CCP4 7.0.0 in Win7), there 
> is no "allow maximal clash..." option anymore. There are only two options 
> "pairwise percent packing" and "accept all solutions".  I used to increase 
> the number in the "allow maximal clash" (let's say from 100 to 200) to do MR 
> and this works fine in a couple of cases.
> 
> With the new Phaser version, I could not do this anymore, can I understand 
> that I just need to increase the pairwise percent packing cutoff (let's say 
> from 5% to 10%) to loose the tight packing criteria and allow more clashes in 
> finding solutions? 
> 
> 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Glass etching of structure

[Randy Read]

I’ve been very happy with Luminorum too, getting them to make blocks for PhD 
students when they finish.  Patrick Goldsmith there has been incredibly helpful 
with making sure that all the little things, like the bonds to 
post-translational modifications that I wanted to be visible, were just right.

Of course, this may not be the most convenient choice if you’re not located in 
the UK or even in Europe!

Best wishes,

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 28 Jan 2017, at 15:47, Paula Salgado  wrote:
> 
> I recently used Luminorum and was very happy with the result - and they 
> managed to do it last minute as well!
> 
> http://www.luminorum.com/shop/
> 
> 
> Paula
> 
> ===
> 
> Dr Paula S. Salgado
> Lecturer in Macromolecular Crystallography
> Institute for Cell and Molecular Biosciences
> Faculty of Medical Sciences
> 3rd Floor Cookson Building 
> Newcastle University   
> Newcastle upon Tyne, NE2 4HH, UK   
> 
> Tel: +44 (0)191 208 7432
> Fax: +44 (0)191 208 7424
> Email: paula.salg...@ncl.ac.uk
> From: CCP4 bulletin board  on behalf of Eleanor Dodson 
> 
> Sent: 28 January 2017 15:46:10
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] Glass etching of structure
>  
> I would ask Steve Gamblin or Phil Walker - they had some done last year.
> 
> philip.wal...@crick.ac.uk
> or 
> steve.gamb...@crick.ac.uk
> 
> 
> On 28 January 2017 at 15:26, Nicholas Larsen 
>  wrote:
> Dear All,
> I'm sorry if this is off topic, but it might be of interest to general 
> audience.  I want to get one of our crystal structures etched in glass.  Can 
> anyone help recommend a company that can provide this service?  BioEtch had 
> come highly recommended to me, but unfortunately their website currently 
> shows they are experiencing equipment malfunction and it's unclear when they 
> would be up and running again.
> 
> Thanks for any suggestion!
> Nick 
> 
> [This e-mail message may contain privileged, confidential and/or proprietary 
> information of H3 Biomedicine. If you believe that it has been sent to you in 
> error, please contact the sender immediately and delete the message including 
> any attachments, without copying, using, or distributing any of the 
> information contained therein. This e-mail message should not be interpreted 
> to include a digital or electronic signature that can be used to authenticate 
> an agreement, contract or other legal document, nor to reflect an intention 
> to be bound to any legally-binding agreement or contract.]

Re: [ccp4bb] phaser

[Randy Read]

Hi,

The syntax of the PACK command has changed, but the commands generated by the 
GUI have also changed in synch, so this problem shouldn't normally arise.

Our best guess is that the user is attempting to rerun an old job, and ccp4i is 
getting confused when trying to interpret the saved old GUI variables in terms 
of the new possibilities.  Checking through the GUI for a PACK choice that has 
been toggled without a valid choice being made might fix it, or alternatively 
the user could set up the whole job from scratch.

If that doesn't explain the problem, please get in touch directly with us!

Best wishes,

Randy Read

> On 6 Jan 2017, at 14:01, Sidhu, Khushwant (Dr.) <k.si...@leicester.ac.uk> 
> wrote:
> 
> Dear all,
>  
> A user has upgraded to ccp4-7 on her Mac. Previously phaser was running fine.
> She is now getting the following error. Could somebody help ?
>  
> The error log is attached.
>  
> Regards
>  
> Sid
>  
>  
>  
> 
>  
> Dr Khushwant Sidhu
> Senior Experimental Officer / I. T. Professional
>  
> Department of Molecular and Cell Biology, 
> 1/61 Henry Wellcome Building , University of Leicester, 
> Lancaster Road, 
> Leicester,
> LE1 7RH
>  
> T: 0116 229 7237
> E: k.si...@le.ac.uk <mailto:k.si...@le.ac.uk>
> 
> Elite without being elitist
> Follow us on Twitter <https://twitter.com/uniofleicester> or visit our 
> Facebook <http://www.facebook.com/uniofleicester> page
>  
>  
> <20_phaser_MR[2][3].log>

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] unusual monoclinic relation?

[Randy Read]

Dear Andy,

You've mentioned that MR doesn't work in any of the space groups you've tried, 
but there are different scenarios that you might be able to recognise from 
where the MR fails.

1. You might have the wrong protein, e.g. crystallised from a contaminant.  In 
that case, I would expect to see no significant signal in either the rotation 
or translation searches.

2. You might have the right protein, but there could be some pathology 
(including some relatively obscure things like reticular merohedral twinning, 
where twin-related reflections interleave and you incorrectly infer long cell 
dimensions).  In this case, I would still expect to see a strong signal for the 
rotation search, given the nearly-perfect sequence identity, but then there 
would be no signal in the translation search (or alternatively an unrealistic 
number of strong translation peaks).

3. There might be some unexpectedly large conformational change, but you 
probably have a reasonable idea of what kinds of motions occur in your system.

Good luck, and have a happy Christmas regardless!

Randy Read

> On 19 Dec 2016, at 08:43, Andrew Lovering <a.lover...@bham.ac.uk> wrote:
> 
> Dear All,
>  
> I have just collected data on a mutant of a protein that should be facile to 
> solve by molrep (one residue/320 changed, approx 2Ang resolution) but is 
> proving problematic. Data merging stats look good.
>  
> The spacegroup is monoclinic, P21, the cell:
>  
> a=39.47 b=157.36 c=74.9 beta=98.26
>  
> I spotted the relevant monoclinic twin laws on ccp4 twinning page and all 
> relate multiples of axes a and c with one another (na +nc etc) but in the 
> above case it would appear b~ = 4a
>  
> There are other datasets, all index in this way, some hint at issues by 
> indexing with the alternate a=74 b=157 c=79 (where a and c “swap” with a 
> doubled, and thus our b=4a turns into b=2c)
>  
> I would appreciate any advice on how to progress! Be nice to solve it 
> pre-xmas.
>  
> Best & thanks in advance,
> Andy

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

[ccp4bb] Recent improvements in Phaser for systematically weak data

[Randy Read]

Dear all,

Between the last official release of Phenix (version 1.11.1-2575, released on 
26 October) and a recent CCP4 update (7.0.023 released on 23 November), we 
fixed a problem in which the fast rotation function was not properly accounting 
for the effect of measurement errors in intensities.  Under many circumstances 
(i.e. with well-measured data to the resolution needed to solve the structure 
by MR), the change is not noticeable.  However, if you have systematically weak 
data (particularly from high levels of anisotropy or potentially from the 
effects of translational non-crystallographic symmetry), the fix can mean the 
difference between failing to solve the structure, because the correct 
orientation has not been generated, or solving it with a default run of Phaser.

We recommend that anyone trying to solve a structure with substantial 
anisotropy (anisotropic deltaB of 50 A^2 or greater) and/or translational NCS 
should update to a newer version of either package.  If you're using CCP4, you 
should get the latest update, because there was a problem with the version of 
coot distributed with the 7.0.023 update.  If you're using Phenix, you should 
get a recent nightly build, i.e. dev-2608 or something even newer.  Please note 
that you want to provide data to Phaser in the form of intensities rather than 
amplitudes to get the best results for such data!

So far, I've found 3 of our test cases where this makes a real difference, but 
I would be interested in hearing from anyone who finds that this makes a 
difference for other cases.

Best wishes,

Randy Read

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Phaser: tNCS present but correction not applied

[Randy Read]

Hi,

That message should mean that you didn't ask for a number of molecules 
divisible by the order of the tNCS. With that vector you need to search for an 
even number of copies.

Let me know if that doesn't explain what you're saying. 

Best wishes,
Randy Read 


Randy J. Read

> On 25 Nov 2016, at 21:27, KAUSHIK H.S. 
> <02b42c18251e-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hello,
> 
> I have a crystal in C2 spacegroup (94.2073 148.003 72.9967 90 97.6686 90) and 
> Xtriage predicts 923 residues in ASU.  My protein could be anywhere between 
> 95 to 110 residues long (assuming the terminals could be cleaved).  Using a 
> homolog, Phaser gives me a solution with LLG=1072 and TFZ=45.  Molrep placed 
> 4 protomers in the ASU however, the crystal packing is poor.
> 
> I understand from Phaser that the tNCS is not accounted for (because the 
> structure is linear?).  Is there a way the exact location of the tNCS related 
> protomers be predicted using information from SfCheck and Xtriage?  
> 
> Xtriage information:
> Fraqc. coord. : 0.500 0.000 0.000
> Distance to origin: 47.104
> Height relative to origin: 79.688%
> p_value(height) : 5.283e-07
> 
> SFCHECK:
> Pseudo-translation is detected: 62.6%
> Pseudo-translation vector: 0.500 0.000 0.000
> 
> Sorry if my question is too naive.
> 
> Thanks in advance,
> Kaushik Hatti,
> Molecular Biophysics Unit,
> Indian Institute of Science,
> India.
> 

Re: [ccp4bb] MrBump and Balbes Molecular Replacement of multiple copies in ASU Failed even with high homology model

[Randy Read]

Dear Alex,

We’ve had very good luck using Phaser to place large numbers of copies of good 
models.  It’s in this kind of case where the increased sensitivity of the 
likelihood approach really helps.  I would suggest trying the different choices 
of model as alternatives, and you might also want to make an ensemble in which 
loops that deviate among the different models have been trimmed off.

If you need more advice on how to do this, you can get in touch off-line.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

> On 31 Oct 2016, at 18:16, Alex Lee <alexlee198...@gmail.com> wrote:
> 
> Hi All,
> 
> I have a protein which contains 72 amino acids. The crystal of this protein 
> diffracts to 2.5A with SG P31 (CELL Dimension 65.9590 65.9590 164.3900 
> 90. 90. 120.). Pointless indicates no twinning.
> 
> Mathew coefficient as below:
> For estimated molecular weight 7919.
> Nmol/asym  Matthews Coeff  %solvent   P(2.49) P(tot)
> _
>   126.0795.29 0.00 0.00
>   213.0490.57 0.00 0.00
>   3 8.6985.86 0.00 0.00
>   4 6.5281.14 0.00 0.00
>   5 5.2176.43 0.00 0.00
>   6 4.3571.71 0.00 0.01
>   7 3.7267.00 0.02 0.02
>   8 3.2662.28 0.05 0.05
>   9 2.9057.57 0.11 0.11
>  10 2.6152.85 0.19 0.19
>  11 2.3748.14 0.25 0.25
>  12 2.1743.42 0.22 0.22
>  13 2.0138.71 0.11 0.12
>  14 1.8633.99 0.03 0.03
>  15 1.7429.28 0.00 0.00
>  16 1.6324.56 0.00 0.00
>  17 1.5319.85 0.00 0.00
>  18 1.4515.13 0.00 0.00
>  19 1.3710.42 0.00 0.00
>  20 1.30 5.70 0.00 0.00
>  21 1.24 0.99 0.00 0.00
> _
> 
> I have at least 10 structures in PDB with homology 60% or above.
> I tried the CCP4online MrBump and Balbes for automatic molecular replacement. 
> No solutions at all.
> I do not know if this is because that the high number of copies in ASU (maybe 
> 10 copies as shown by Mathew coefficient) hamper the MrBump and Balbes to do 
> MR?
> 
> Dose anyone experience similar difficulties with high number of copies to do 
> automatic MR?
> 
> Thanks ahead!

Re: [ccp4bb] MR phasing using Negative Stain EM reconstruction

[Randy Read]

Dear Pascal,

I'm assuming that you're talking about using the negative stain image as an MR 
model.  I don't recall hearing of this having ever worked (though I would be 
very interested of course if anyone has managed to do this!), but my intuition 
is that it's not going to work.  Negative stain just gives you an external 
shape, whereas a cryo-EM reconstruction has internal features as well.

Presumably you don't have atomic models of the individual components of the 
complex?  If you did, using those directly for MR would be my first choice, but 
you could also consider making a pseudo-atomic model by docking them into the 
shape of the negative stain image.  Such a model would add substantial 
higher-resolution information.

Best wishes and good luck,

Randy Read

> On 25 Oct 2016, at 02:49, Pascal Egea <pas...@msg.ucsf.edu> wrote:
> 
> Dear All,
> 
> I would like to know if it is possible to use a low resolution EM 
> reconstruction of a complex obtained in negative stain EM (not cryo EM) to 
> help molecular replacement in a 4.5A resolution X-ray diffraction data set of 
> the same complex
> I am aware of the possibility of using low resolution cryoEM maps for MR as 
> described in the review from Jackson et al in Nature Protocols but I was 
> wondering if there is an intrinsically impossibility for negative stain 
> reconstructions.
> 
> Any thoughts or advice will be greatly appreciated.
> 
> Best,
> 
> -- 
> Pascal F. Egea, PhD
> Assistant Professor
> UCLA, David Geffen School of Medicine
> Department of Biological Chemistry
> Boyer Hall room 356
> 611 Charles E Young Drive East
> Los Angeles CA 90095
> office (310)-983-3515
> lab  (310)-983-3516
> email pegea at mednet.ucla.edu <http://mednet.ucla.edu/>

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Twinning/Spacegroup Woes

[Randy Read]

Dear Rhys,

There's one important consideration to start from, which is that the space 
groups P6322 and C2221 with their associated cell dimensions are two different 
ways to describe the shapes of the unit cells and the positions of the spots.  
You can see this by running phenix.explore_metric_symmetry on the two cases.  
(Perhaps there's a tool in CCP4 for this as well?)  So dehydrating your crystal 
hasn't really changed the cell, and the underlying symmetry will be basically 
the same (give or take the possibility that a symmetry operator that is 
crystallographic in one case might be pseudosymmetry in the other, if things 
move a bit in the crystal).

As you say, there are no twin operators for P6322 so, if that crystal is 
twinned, this implies that the true symmetry is lower and that part of the 
apparent symmetry arises from the twinning.  You don't mention anything about 
how well the data merge in P6322 for the first data set, but it is also 
possible to computationally twin your data by merging in too high symmetry!

Have you run pointless on the unmerged data from both crystals?  That may give 
you a better idea of the actual symmetry of the diffraction pattern (part of 
which may come from twinning).  If the two data sets give different results, 
then maybe the true symmetry is closer to the lower symmetry you see from the 
two data sets.

Do you have a molecular replacement model?  If you do, then molecular 
replacement in lower symmetry subgroups can work very well to discover the true 
symmetry while solving the structure at the same time.  Phaser has some nice 
new features to list the possible subgroups and to expand the data to lower 
symmetry (though the latest version, which is available in Phenix nightly 
builds but not quite yet in CCP4 works considerably better than the current 
CCP4 version).  Also, the tool Zanuda in CCP4 can be very effective for working 
out the true symmetry.

Best wishes,

Randy Read

> On 4 Oct 2016, at 00:26, Rhys Grinter <rhys.grin...@monash.edu> wrote:
> 
> Dear All,
> 
> As I have approached my crystallography from a biological perspective, 
> sometimes so of the more mathematical/geometrical aspects sometimes perplex 
> me. I was wondering if anyone would be able to clarify what is going on with 
> some problematic crystals I'm working on.
> 
> I've grown crystals of a protein which forms a concentration dependent 
> oligomer. This is almost certainly a physiological oligomer and probably is a 
> hexamer at maximum oligomerisation (although maybe a trimer). These crystals 
> diffract poorly, however after some optimisation I managed to collect data to 
> around 3.6 A, with a predicted space group of P6322 with unit cell dimensions 
> of 177, 177, 150 . In order to improve diffraction I performed dehydration on 
> these crystals. This seemed to improve diffraction to around 3A (although as 
> the crystals are quite variable attribution of effect is a little difficult), 
> however the best space group I can find for indexing is C2221 with a unit 
> cell of 177, 310, 151, XDS doesn't process the data when I force the previous 
> P6322 SG. It seems also that the C2221 space group isn't the correct choice 
> as the merging stats are worse than I would expect from looking at the 
> diffraction pattern. 
> 
> Additionally, the intensity statistics from both space groups suggests 
> twinning. Although for the P6322 space group it says twinning is not 
> possible. If this is the case what is causing these abnormal intensities and 
> is this related to my SG ambiguity?
> 
> Also what is the best what to proceed with processing in this case?
> 
> Cheers,
> 
> Rhys
> 
> -- 
> Dr Rhys Grinter
> Sir Henry Wellcome Fellow
> Monash University
> +61 (0)3 9902 9213
> +61 (0)403 896 767

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] [phenixbb] accept MR solution even if the number of clashes is above the threshold

[Randy Read]

Dear Almu,

In the development version of Phaser currently available in nightly builds of 
Phenix (and soon in an upcoming stable release of Phenix, as well as in CCP4), 
solutions that fail to pack, even though they have such high TFZ scores that 
they should normally have been convincing, are subjected to a pruning 
algorithm, where the local occupancies of residues along the chain are refined 
and ones that refine to low values are trimmed out.  This will often sort out 
problems from loops, or even domains, that are in the wrong conformation.  By 
default, potential solutions have to have a TFZ8 to be subjected to this 
algorithm, but this threshold can be changed.

Please give this a try and tell us how it works for you!  If it doesn’t, there 
are other possible issues, generally coming down to having the wrong 
spacegroup, so it might be worth double-checking the choice of spacegroup.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

 On 2 Jul 2015, at 13:55, Almudena Ponce Salvatierra maps.fa...@gmail.com 
 wrote:
 
 Hi everyone, 
 
 I am running Phaser (from Phenix) and while checking the .log file (it is 
 still running) I realize that it found some solutions with a TFZ score over 
 7, but it won't take them I guess because the number of clashes is higher 
 than allowed (12, I guess they're not so many either).
 
 My models are poor and for so I would like to check these solutions that for 
 now Phaser will reject. However in the input I don't want to increase the 
 allowed number of clashes because this will then affect all the other 
 solutions and this will take forever... 
 
 So my question is: can I tell phaser somehow to write down/save the solutions 
 for which the TFZ score will be higher than 7? Even though the clashes
 
 Thanks a lot in advance
 
 All the best, 
 
 Almu
 
 -- 
 Almudena Ponce-Salvatierra
 Macromolecular crystallography and Nucleic acid chemistry
 Max Planck Institute for Biophysical Chemistry
 Am Fassberg 11 37077 Göttingen
 Germany
 
 ___
 phenixbb mailing list
 pheni...@phenix-online.org
 http://phenix-online.org/mailman/listinfo/phenixbb
 Unsubscribe: phenixbb-le...@phenix-online.org

Re: [ccp4bb] reference needed for TRUNCATE NO equation.

[Randy Read]

Hi Bernhard,

Thanks for that, but it reinforces my impression (also reinforced offline by 
George Sheldrick) that this is something that was floating around at some 
point, and no-one ever bothered to put it in a proper publication.

The background is that we've just submitted a paper on a new way to deal with 
intensity errors (the paper with all the equations that Word has been giving me 
grief over, in the earlier thread) and we wanted to compare it to what people 
use now.  The French  Wilson treatment is the most common, and this equation 
probably comes next out of a number of alternatives that we found, but it would 
have been nice to give a reference.  

Phil Evans gave me the same derivation (not a surprise, as he seems to have 
been the person who put that code in TRUNCATE).  The thing that bothers me 
about this is that it's not immediately obvious why it's right to solve the 
equation

  (F+SIGF)^2 == I+SIGI

for SIGF and not

 (F-SIGF)^2 == I-SIGI

Both are ways of looking at the effect of a perturbation, but the latter 
equation gives a complex value for SIGF when ISIGI.  Phil said something about 
the skewness of the probability of the amplitude, when the probability 
distribution for the corresponding intensity is an unskewed Gaussian, and maybe 
there's some kind of argument for which direction of perturbation is 
appropriate for a particular sign for the skewness?

It turns out that an equivalent equation is used in the ADDREF program in the 
Xtal package of programs for small molecule crystallography.  ADDREF was 
written by Syd Hall and George Davenport, so I've been in touch with Syd Hall, 
who also can't remember who first came up with this equation, but who feels 
that there was more to the derivation than Phil's explanation!

Perhaps it will now remain a mystery, unless someone else on the BB has a 
recollection of the origin of this equation decades ago!

Randy

 On 26 May 2015, at 23:51, Bernhard Rupp b...@ruppweb.org wrote:
 
 Derivation but no reference except truncate is in BMC page 330 sidebar.
 Probably read the source code or some prog. ref….
 BR
  
 From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK 
 mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Jonathan Cooper
 Sent: Tuesday, May 26, 2015 4:33 PM
 To: CCP4BB@JISCMAIL.AC.UK mailto:CCP4BB@JISCMAIL.AC.UK
 Subject: [ccp4bb] reference needed for TRUNCATE NO equation.
  
 Dear all
  
 I have been asked by Randy for a reference for the following equation which 
 appears in Sherwood  Cooper:
  
 http://www.ucl.ac.uk/~rmhajc0/truncate_no.jpg 
 http://www.ucl.ac.uk/~rmhajc0/truncate_no.jpg
  
 ... or in text form: 
  
 sigma(F) = sqrt(I + sigma(I)) - sqrt(I)
  
 It is the equation which TRUNCATE uses to calculate sigma(|F|) if the user 
 switches off the French-Wilson treatment, although that is never usually done!
  
 Can anyone shed any light on who derived it and where/whether it has been 
 published, etc?
  
 Cheers for now.
  
 Jon Cooper

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] SUMMARY: Equation Editor woes with Office 2011 for Mac

[Randy Read]

Thanks, as always, to everyone for a thoughtful discussion!

Martin Montgomery suggested trying the beta release of the upcoming Office for 
Mac suite.  I installed this (which required carrying out the update to 
Yosemite that I had been putting off), and it turns out that rather than fixing 
the bug with Equation Editor (i.e. the one based on the MathType software), 
Microsoft removed that option and all that is left is their built-in equation 
tool (which it turns out is based on MathML markup)!  So it looks like anyone 
who likes the Equation Editor will either have to cough up for a licence for 
the fully-featured MathType (which creates its own problems collaborating with 
people who don't have a licence) or learn to use an alternative.

A few people (Robbie Joosten, Klaus Fütterer and Shing Ho) suggested that I 
should bite the bullet and learn how to use the built-in equation tool.  I had 
tried this and found it unintuitive compared to the Equation Editor, but maybe 
that's just because I haven't put in anything like the same amount of time 
learning it.  This is probably going to be the way I'll go, mostly for 
compatibility with collaborators (who are almost all using Word without special 
plugins) but partly for journal typesetting considerations mentioned below.

As I said before, collaboration considerations rule out a full-scale adoption 
of LaTeX for me, but maybe the suggestions of different LaTeX environments will 
be useful for others.  Perhaps Lyx (George Reeke's suggestion) might be 
sufficiently WYSIWYG to convince collaborators to install it, although Blaine 
Mooers and Murpholino Peligro pointed out some potential complications with 
this solution, particularly in compatibility with bibliography tools.

However, there also seem to be some interesting tools that allow you to prepare 
just the equations in LaTeX and paste at least images of those equations into 
Word documents.  Your collaborators wouldn't be able to edit the equations 
without getting the same tools, which is a downside of this option.  Several 
people (Jonathan Davies, Nicolas Soler, Blaine Mooers) suggested LaTeXit and 
Bill Scott suggested TeX-fog.  I'm going to play with LaTeXit to see how I like 
that, particularly for PowerPoint where there's less of an issue with 
collaboration.

Similarly, Adrian Goldman suggested that the Mac Grapher program could be used 
to prepare pictures of equations.

Then there was some lateral thinking.  Jens Thomas suggested that I could turn 
AutoSave off and implement my own replacement from the Terminal window with the 
following script:

—— 
#!/bin/bash
[[ $# -ne 1 ]]  echo Usage: $0 path_to_file  exit 1
while true;
do
   cp $1 ${1}.bak
   sleep 60
done
—— 

Kay Diederichs suggested installing Word for Windows on a Windows virtual 
machine, though I balk at paying Microsoft for two more licences to work around 
a bug in their software!

Another issue, which I didn't raise in my original post, is what happens when 
your file gets to the publisher.  Putting my Acta D hat on (and for all of us 
as members of the IUCr), we should try to make it easier for our journals.  
Also, proofreading is much easier if some poor person hasn't had to retype all 
the equations.  So I asked at Acta Cryst what happens when papers are submitted 
in different formats.  One possibility is to submit a LaTeX document for 
everything, in which case the equations should be fine.  For documents 
submitted in Word format, Simon Westrip described their workflow as follows:

——
1) All documents (Word or OpenOffice) are processed using Word on Windows 
machines

2) MathType is used to convert any equation objects in the document to Plain TeX

3) If MathType fails (which it often does with Microsofts 'docx MathML' 
equations,
then my own software attempts to convert MathML-based equations objects to 
Plain Tex.

4) Any remaining equations have to be typeset manually (e.g. those that are 
included as
images)

So ideally, in the absence of MathType or any other plugin that can embed 
equation 'objects'
in Word documents, we would prefer that the equations be prepared using Word's 
built-in editor.
—— 

As Simon said in a followup, if other solutions become popular, they'll 
accommodate them at the IUCr journals.  In the meantime, I guess I'll be 
learning the built-in editor.

Randy Read

 On 18 May 2015, at 09:10, Randy Read rj...@cam.ac.uk wrote:
 
 Rather off-topic, but maybe someone on the list has found a way to work 
 around this!
 
 There’s a problem with the Equation Editor in Office 2011 for Mac (i.e. the 
 one that is based on a stripped-down version of MathType, which you get with 
 Insert-Object-Microsoft Equation).  You can insert an equation, re-open it 
 and edit it several times, and then suddenly (and seemingly randomly) the 
 equation object will be replaced by a picture showing the equation, which can 
 no longer be edited.  I’m writing a rather equation-heavy paper at the 
 moment

Re: [ccp4bb] Low Phaser RFZ

[Randy Read]

Dear Eric,

I've just clarified the section in our documentation on how to tell if Phaser 
has solved your structure:
http://www.phaser.cimr.cam.ac.uk/index.php/Molecular_Replacement#Has_Phaser_Solved_It.3F

Briefly, the RFZ score is not that diagnostic of success or failure, 
particularly for high-symmetry space groups.  For your space group, P6(2)22, 
the rotation function is evaluating the agreement between the observed data and 
the structure factors that you might get by adding up the contributions of 12 
symmetry-related molecules.  Because, at the rotation function stage, the 
relative phase angles of the contributions from symmetry-related molecules is 
not known, there's a lot of uncertainty so the signal-to-noise is low.  That's 
why we emphasise the TFZ as the thing to look at.  In your case, a TFZ of 8.4 
for the first molecule is already pretty convincing, but even if the first 
molecule had a poor signal on its own, the TFZ of 13.7 for the second copy 
would tell you that the first one must have been right.

I would say that you can be confident that this is basically right.  (With the 
normal provisos: that doesn't rule out possible complications like 
pseudosymmetry, but as Phil Evans says the space group is a hypothesis, and you 
always have to be prepared to reconsider the hypothesis later if, say, it turns 
out to be difficult to refine the structure.)

Best wishes,

Randy Read

 On 19 May 2015, at 01:36, Eric Karg 
 052044071b36-dmarc-requ...@jiscmail.ac.uk wrote:
 
 Hi all,
 
 Running Phaser using the apo protein as search model on a ~2.5 A dataset of a 
 protein-DNA complex, I get a single solution but with low RFZ. The map looks 
 reasonable but I was wondering why the RFZ is so low. Would this solution be 
 acceptable?
 
   SOLU SET  RFZ=3.2 TFZ=8.4 PAK=0 LLG=66 TFZ==10.6 RFZ=2.9 TFZ=13.7 PAK=0 
 LLG=203
TFZ==14.0 LLG=1440 TFZ==34.2
   SOLU SPAC P 62 2 2
   SOLU 6DIM ENSE ensemble1 EULER 181.8 55.7 74.8 FRAC 0.27 0.26 -0.40 BFAC 
 -7.38
   SOLU 6DIM ENSE ensemble1 EULER 4.5 120.2 7.9 FRAC -0.31 0.20 -0.09 BFAC 
 12.61
   Ensemble ensemble1 RMS variance(s): 0.87
 
 Thank you for your help!

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

[ccp4bb] Equation Editor woes with Office 2011 for Mac

[Randy Read]

Rather off-topic, but maybe someone on the list has found a way to work around 
this!

There’s a problem with the Equation Editor in Office 2011 for Mac (i.e. the one 
that is based on a stripped-down version of MathType, which you get with 
Insert-Object-Microsoft Equation).  You can insert an equation, re-open it 
and edit it several times, and then suddenly (and seemingly randomly) the 
equation object will be replaced by a picture showing the equation, which can 
no longer be edited.  I’m writing a rather equation-heavy paper at the moment, 
and this is driving me crazy.

This seems to be a known bug, which has existed from the release of Office 
2011.  Apparently it happens, unpredictably, when an AutoSave copy of the 
document is saved, so you can avoid it by turning off the AutoSave feature.  
The last time this drove me crazy, several years ago, I did try turning off 
AutoSave.  For a while, I was very good about manually saving frequently, but I 
got into bad habits and eventually Word crashed after I had worked for several 
hours on a grant proposal without manually saving.  So I turned AutoSave back 
on.

At the moment, the least-bad solution seems to be to turn off AutoSave while 
I’m working on a document with lots of equations and then (hopefully) remember 
to turn it back on after that document is finished.  But it would be great if 
someone has come up with a better cure for this problem.

No doubt someone will suggest switching from Word to LaTeX, but I need to be 
able to collaborate on paper-writing, and even though I might be willing to 
invest the effort in learning LaTeX, I can’t really expect that of my 
collaborators.  Most people in our field do use Microsoft Word, regardless of 
its failings.  I’ve also tried using the professional version of MathType, but 
that requires your collaborators to install it as well — and I don’t think that 
cured the equation to picture problem anyway.

Thanks!

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Phaser going into infinite loop in Ample

[Randy Read]

Hi Dale,

It must actually be AMPLE deciding how many copies to search for.  Phaser will 
give you some information about how consistent the specified composition is 
with the Matthews volume, but it just searches for the number of copies that 
it's instructed to look for.  We haven't put the intelligence into it to 
correlate the model(s) with the composition information and try out different 
possibilities.  At the moment, we're leaving that level of analysis to 
pipelines like MRage.

We're well aware of the tension between looking hard enough to find a solution 
in a difficult case and not throwing good CPU cycles after bad when it's 
hopeless.  We're gradually introducing new features to make these decisions 
better, but we do tend to prefer wasting CPU time to missing solutions.  
However, we've introduced a couple of ways to limit the amount of time that 
Phaser spends pursuing very difficult or hopeless solutions, partly for the 
benefit of people such as the developers of Arcimboldo, AMPLE and the 
wide-search molecular replacement pipeline.  One is the KILL command, which is 
a rather blunt instrument saying to give up if a solution isn't found in a 
certain length of time.  In AMPLE, if you set quick mode, then the KILL time is 
set to 15 minutes.  Another option (which I don't think AMPLE uses) is the 
PURGE command, where you can say (for instance) that Phaser should pursue a 
maximum of 25 partial solutions when adding the next component.

If you're seeing an infinite loop, it would be handy if you could send me a 
copy of the logfile showing what is going on.  There have been some bugs 
leading to such infinite loops under some circumstances, and if you're running 
into one of those there's a good chance that it has been fixed in a recent 
nightly build of Phaser available through Phenix.  You can instruct CCP4 to use 
the Phaser executable from Phenix, and I think this should work fine in AMPLE, 
though I haven't tested it — I don't think any relevant syntax has changed.  
It's been a while since we've had a new stable release of Phaser in either CCP4 
or Phenix, so we're aiming to get one out in the not too distant future.

All the best,

Randy

 On 22 Apr 2015, at 05:56, Dale Tronrud de...@daletronrud.com wrote:
 
 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1
 
 
   We are having a problem with AMPLE and hope someone can help.
 
   The protein is about 70 amino acids long and we suspect it forms a
 coiled-coil.  Our previous attempts at molecular replacement have
 failed so we hoped that AMPLE, with its ability to generate a variety
 of potential models, would do the trick.
 
   Our problem is that all of our CPU cores are consumed by Phaser
 jobs that are not making progress.  With this protein Phaser decides
 that it will look for 11 copies in the asymmetric unit.  For a few of
 the possible ensembles it fails to find even one copy and gives up.
 That's fine with us.  For other ensembles it finds a handful of
 possible first positions, goes on to look for a second and fails, then
 goes back to try to place a second copy again.  We presume that the
 intent is to lower the acceptance criteria in the second pass, but in
 actuality Phaser simply repeats the same search that failed before and
 fails again.  The leads to an infinite loop.
 
   Once all the cores are occupied in this futile endeavor AMPLE makes
 no further progress.
 
   How can we get Phaser to either try harder to place a molecule or
 to give up?
 
   We are using CCP4 6.5.008 and the copy of Phaser that came with it.
 We used CCP4i to create a script which we modified slightly and ran
 using the at command.  The command is:
 
 /usr/local/ccp4-6.5/bin/ccp4-python -u /usr/local/ccp4-6.5/bin/ample
 - -mtz /user/sarah/xray/1Apr_Athena/SD6004_2_001_mergedunique14.mtz
 - -fasta /user/sarah/xray/1Apr_Athena/swaseq.fa -mr_sequence
 /user/sarah/xray/1Apr_Athena/swaseq.fa -nmodels 500 -name MVD0
 - -run_dir /home/sarah/AMPLE -nproc 6 -make_models True -rosetta_dir
 /usr/local/rosetta-3.5 -frags_3mers
 /user/sarah/xray/1Apr_Athena/aat000_03_05.200_v1_3 -frags_9mers
 /user/sarah/xray/1Apr_Athena/aat000_09_05.200_v1_3 -make_frags False
 - -F F -SIGF SIGF -FREE FreeR_flag  -early_terminate True   -use_shelxe
 True -shelx_cycles 15 -use_arpwarp False
 
 Any help is appreciated,
 Dale Tronrud
 Sarah Clark
 -BEGIN PGP SIGNATURE-
 Version: GnuPG v2.0.22 (MingW32)
 
 iEYEARECAAYFAlU3KhEACgkQU5C0gGfAG117JACfXahEX8z1k3ev043a7V2SzhNh
 p6UAnAmKkUZe46zbiXckdDTNgUQ0dgWq
 =A/Sj
 -END PGP SIGNATURE-

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Sortwater NCS Matrix input

[Randy Read]

Dear Shane,

It's not CCP4, but Tom Terwilliger's find_ncs program works conveniently.  Like 
coot, it's only returning 4 decimal places, if that's a problem (though adding 
a 5th decimal place would only typically change the result of a rotation by 
hundredths of an Angstrom).  I've just tried it:

phenix.find_ncs 1bos.pdb

returns a file called find_ncs.ncs_spec, with information like the following:

rota_matrix0.32230.7547   -0.5714
rota_matrix   -0.72820.58330.3597
rota_matrix0.60480.30020.7376
tran_orth-2.1022   47.6880  -37.6220

center_orth   24.4889   44.51719.

Best wishes,

Randy Read

 On 15 Apr 2015, at 17:21, Shane Caldwell shane.caldwel...@gmail.com wrote:
 
 good enough to pass the orthonormal test.
 
 .. scratch that, it passes sometimes and still fails for other 
 structures/chains. So I'm still in search of a higher-precision export
 
 Shane 
 
 On Wed, Apr 15, 2015 at 11:44 AM, Shane Caldwell shane.caldwel...@gmail.com 
 mailto:shane.caldwel...@gmail.com wrote:
 Hi Bernhard, thanks for the pointer. It looks like most of the coot ncs 
 functions don't print the matrix values, which made exporting them tricky, 
 especially because I'm pretty new to scripting in Python. 
 
 I have 4 chains, and 
 ncs_ghosts(0) just spits out 
 
 result is [['NCS found from matching Chain B onto Chain A', 'B', 'A', False, 
 False], ['NCS found from matching Chain C onto Chain A', 'C', 'A', False, 
 False], ['NCS found from matching Chain D onto Chain A', 'D', 'A', False, 
 False]]
 
 However, after playing around, I was able to extract the matrix from coot 
 using single_manual_ncs_ghosts, and while it's still at only limited 
 precision, it's good enough to pass the orthonormal test. 
 
 Running coot  coot.log and then running the python:
 
 single_manual_ncs_ghosts(0,start,end,A,B)
 
 where start and end were my residue numbers, wrote out :
 
 INFO:: coordinates transformed by orthonal matrix: 
 |   0.07272,0.9909,   -0.1131|
 |0.9973,  -0.07361, -0.003729|
 |  -0.01202,   -0.1125,   -0.9936|
 ( 41.44,-39.63, 77.24)
 INFO:: fractional coordinates matrix:
 |   0.06948, 1.063,   -0.4001|
 |0.9013,  -0.07361,   -0.2467|
 |-0.012,   -0.1243,   -0.9903|
 (0.6896,   -0.3967,0.8545)
 
 Which is good enough for now, though it does feel like repairing the dining 
 room chair with duct tape. If anyone knows a cleaner way to get these values, 
 It'd be great to know!
 
 Shane
 
 On Wed, Apr 8, 2015 at 6:08 AM, B.Lohkamp b.lohk...@gmail.com 
 mailto:b.lohk...@gmail.com wrote:
 
 Just for completeness, of course (!?) you can get something like this in 
 Coot. In terms of return value and accuracy (3 digits) I would use the 
 scripting function:
 
 ncs_ghosts(imol)  - pythonic
 
 (ncs-ghosts imol)  - schemey
 
 imol - your protein with NCS
 
 
 Bernhard
 
 Alright, thanks! It's a good thing, then, I spent the afternoon brushing
 up on matrices.
 
 I guess the next, probably more general question for the bb is: which
 utilities export an NCS transformation matrix with more precision?
 *superpose* and *gesamt* only export three decimals, though I'm sure
 they use greater precision under the hood. I'm not opposed to exporting
 from coot or pymol either, I just haven't figured out how to do this yet
 - what would be the simplest way to calculate and export an NCS
 transformation matrix?
 
 Shane
 
 On Wed, Apr 1, 2015 at 7:34 PM, Dale Tronrud de...@daletronrud.com 
 mailto:de...@daletronrud.com
 mailto:de...@daletronrud.com mailto:de...@daletronrud.com wrote:
 
 
 I think you are on the right track - There are not enough decimal
 points in your matrix elements to pass the orthonormal test.  This test
 checks that the length of each row (x^2+y^2+z^2) is equal to one and the
 dot product of each row with every other row is equal to zero.  If the
 values on your NCS statement are in row order I calculate 0.999337 for
 the length of the first row.  If the program is testing if this is equal
 to one to four decimal points you lose.
 
 You have to add more digits, but just adding zeros isn't going to
 accomplish much.  The best solution is to get your ncs program to report
 its matrix with more digits -- three is pitiful.  Failing that you could
 calculate one element of each row from the other two to ensure the
 length is equal to one at a higher level of precision and hope this
 doesn't mess up the dot product test.  You'll end up with one number in
 each row having more than three decimal places.
 
 Dale Tronrud
 
 On 4/1/2015 2:52 PM, Shane Caldwell wrote:
  Hi ccp4bb,
 
  I'm trying to solve a problem I never quite figured out in the past. I'd
   like to use the *sortwater* utility to send my picked waters to
 various
  protein chains, and to give them the same residue number if they are
  NCS-equivalent, as the manual outlines

Re: [ccp4bb] adjusting bad Ramachandran angles

[Randy Read]

Hi,

Sorry, can’t take the credit for that!  The names that come to mind are Gerard 
Kleywegt and Alwyn Jones.

I’ve actually said in the past that, the more criteria you add to optimising 
your model, the less likely that you can satisfy all the criteria and still 
have a wrong model!  But I agree with the point I think Gert was making, which 
is that the Ramachandran plot doesn’t lend itself to a simple gradient-driven 
optimisation strategy, i.e. the correct answer isn’t likely to simply be 
downhill in a potential function that includes some kind of Ramachandran 
restraints, because you will almost certainly have to do something more 
exhaustive that will involve jumping local barriers.

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 26 Feb 2015, at 04:37, Jeremy Tame jt...@tsurumi.yokohama-cu.ac.jp wrote:

 I think Goodhart's Law applies here (see the Wikipedia page):
 When a measure becomes a target, it ceases to be a good measure.
 
 From memory I believe Randy Read and George Sheldrick have commented that
 Ramachandran plots are a good measure of structure quality, and therefore 
 should
 not be used explicitly at the modelling stage. Some residues may be difficult
 if they have more than one backbone conformation or are just mobile, but
 expressly holding them in favoured regions of the Ramachandran plot is not a 
 good 
 idea. The most interesting proteins are of course enzymes, and the 
 Ramachandran
 outliers are often among the most interesting active site residues. So you 
 may be trying
 to eliminate something which is actually more important than a low Rfactor.
 
 The idea of limiting data use may seem counter-intuitive, but to take another
 example from economics, John Cowperthwaite was in charge of Hong Kong's
 financial affairs in the 1960s. He attributed the success of the economy 
 under his
 tenure to his adamant refusal to collect any economic data whatsoever!
 
 
 On Feb 26, 2015, at 2:08 AM, Michael Murphy wrote:
 
 Does anyone know of a way to adjust Ramachandran angles so that they fall 
 within the preferred range? Either in Coot or possibly some online server? I 
 have been trying to do it manually without much success, I was wondering 
 whether there might another way to do it. -Thanks

Re: [ccp4bb] Poor experimental phases for a largish structure...

[Randy Read]

Dear William,

In a similar circumstance, we cut out the density from one crystal, and used 
that to solve the structures of non-isomorphous crystals by molecular 
replacement.  At this point, we had an envelope (needed to cut out the 
density), initial maps, and operators to place one copy of the density in the 
other crystals, which is all you need for multi-crystal averaging.  The more 
non-isomorphous your crystals are, the more powerful this is.  What’s important 
(Kevin Cowtan explained this in a talk at the IUCr meeting in Glasgow) is that 
the size of the difference in cell dimensions should preferably be comparable 
to the resolution of the data (dmin) or larger.  Our case was a bit difficult 
because the non-isomorphism was marginal — enough to stop us from using MIR(AS) 
but barely enough to help the phasing.

There are other cases of people using similar approaches in the literature.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 18 Feb 2015, at 12:51, William Chao william.c...@cancer.org.uk wrote:

 Dear all,
 
 I am trying to phase a largish novel structure of 130 kDa with P21 (sometimes 
 P222) space group. So far I have collected a few 3.5ish-Å 3-wavelength MAD 
 datasets from SeMet derivatives as well as an anisomorphous 6Å Hg peak 
 dataset (with detectable anomalous signal). As my crystal is rather 
 anisotropic, CC1/2 of one direction of the Se data drops below 0.5 at 4Å. I 
 can generate a map from a MAD dataset that gives a overall shape of the 
 molecule with clear molecular boundary​ after solvent flattening, resembling 
 the shape of some EM class averages that I obtained earlier. However, the 
 density of this map is very discontinuous and is impossible to build any 
 helix in by machine or by eye. ​
 
 As I use an insect-cell expression system and the occupancy of Se is expected 
 to be low, the programmes that I used could only find a 2 reasonable sites 
 out of 24 Se per molecule. I have about 10 Se crystals and 50 native crystals 
 (which I shall use for derivatisation) left for one last trip before the 
 synchrotron shuts down for two months. Could someone advise me on a 
 reasonable data collection strategy that could maximise my chance on this 
 upcoming trip? I am sure that many people have encountered difficult data 
 like this one and have solved their structures successfully. Would anyone be 
 able to advise me on how it'd be best to improve my phases/density given the 
 limitations of the data?
 
 Also to mention that the best native crystal can diffract to 3A with CC1/2 of 
 the worst direction dropping to 0.5 at 3.5A.
 
 Many thanks in advance!
 
 William
 ---​
 
 NOTICE AND DISCLAIMER
 This e-mail (including any attachments) is intended for the above-named 
 person(s). If you are not the intended recipient, notify the sender 
 immediately, delete this email from your system and do not disclose or use 
 for any purpose. 
 
 We may monitor all incoming and outgoing emails in line with current 
 legislation. We have taken steps to ensure that this email and attachments 
 are free from any virus, but it remains your responsibility to ensure that 
 viruses do not adversely affect you. 
 Cancer Research UK
 Registered charity in England and Wales (1089464), Scotland (SC041666) and 
 the Isle of Man (1103)
 A company limited by guarantee. Registered company in England and Wales 
 (4325234) and the Isle of Man (5713F).
 Registered Office Address: Angel Building, 407 St John Street, London EC1V 
 4AD.

Re: [ccp4bb] Absence of contact between layers in a crystal

[Randy Read]

Actually, if you go back through the archive of CCP4-BB from the first time 
this came up, I think you'll find that there are real crystals with apparent 
gaps in the packing.  This can arise because of statistical disorder, where 
there are two or more ways that a statistically-disordered layer in the crystal 
can mediate the interaction between ordered layers.  So not finding a connected 
packing is something to look closely at and worry about, but it doesn't 
necessarily indicate that somebody did a bad job of making up a structure.

Randy

On 6 Feb 2015, at 11:09, Robbie Joosten robbie_joos...@hotmail.com wrote:

 Not in real crystal structures ;)
 
 Cheers,
 Robbie
 
 Sent with my Windows Phone
 Van: Kerff Fred
 Verzonden: ‎6-‎2-‎2015 12:02
 Aan: CCP4BB@JISCMAIL.AC.UK
 Onderwerp: [ccp4bb] Absence of contact between layers in a crystal
 
 Hello,
 
 Looking at structure 2HR0 (The structure of complement C3b provides insights 
 into complement activation and regulation. »,Abdul Ajees, A.,  Gunasekaran, 
 K.,  Volanakis, J.E.,  Narayana, S.V.,  Kotwal, G.J.,  Krishna Murthy, H.M.;  
 (2006) Nature 444: 221-225), I noticed the absence of contacts between layers 
 in the crystal. Is it something that has already been observed in other 
 crystals?
 
 Best regards,
 
 Fred
 -
 Frédéric Kerff
 Chercheur qualifié F.R.S.-FNRS
 Cristallographie des protéines
 Centre d'Ingénierie des Protéines
 Université de Liège
 17, Allée du 6 Août - Bat B5a
 4000 Liège (Belgium)
 Tel.: +32 (0)4 3663620
 Fax: +32 (0)4 3663772
 
 
 
  Le 6 févr. 2015 à 10:12, Tim Gruene t...@shelx.uni-ac.gwdg.de a écrit :
  
  -BEGIN PGP SIGNED MESSAGE-
  Hash: SHA1
  
  Dear Smith,
  
  The sca file most likely does not contain flags. pointless can read
  the sca file, standardise it to ccp4 standards and freerflag marks
  random reflections. You should use the maximum of 500 unique
  reflections or 5% of the unique reflections, whichever is larger.
  
  Best,
  Tim
  
  On 02/06/2015 09:49 AM, Smith Lee wrote:
  Dear All, I have a sca file. Will you please tell me by which
  software or how I can know whether the sca file contains R-free
  tags? If not, by which software or how I can add the R-free tags?
  And how much of the reflections I add the R-free tags? I am looking
  forward to getting your reply. Smith
  
  
  - -- 
  - --
  Dr Tim Gruene
  Institut fuer anorganische Chemie
  Tammannstr. 4
  D-37077 Goettingen
  
  GPG Key ID = A46BEE1A
  
  -BEGIN PGP SIGNATURE-
  Version: GnuPG v1.4.12 (GNU/Linux)
  
  iD8DBQFU1IWVUxlJ7aRr7hoRAmZHAJ4+6wREnwkFN0EhfErAA0tPSopKKwCgiLdi
  j0JFZac4kAh8twpov71MG84=
  =XN57
  -END PGP SIGNATURE-

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Bulk solvent correction in Phaser MR LF

[Randy Read]

Hi Bernhard,

You're right, it's a Babinet-style correction to the SigmaA curve for the 
effect of neglecting bulk solvent in the model.  We gave a formula in equation 
19 of the main Phaser reference (McCoy et al., 2007), but actually we've 
changed the form of this since then.  It looks like the equation wasn't given 
in the documentation, so I've just updated that:

http://www.phaser.cimr.cam.ac.uk/index.php/Keywords#SOLPARAMETERS

Best wishes,

Randy Read

On 3 Feb 2015, at 13:49, Bernhard Rupp (Hofkristallrat a.D.) 
hofkristall...@gmail.com wrote:

 Hi Fellows,
  
 I cannot find the proper reference for the implementation of bulk solvent 
 corrections in the
 Phaser Molecular replacement likelihood functions. For the unplaced model, it 
 can only
 be a Babinet model to improve the scaling, and I believe that is implemented 
 via a
 Babinet rescaled function for sigma A including the coordinate error.
  
 Can anybody help me where to find that?
 Thx, BR

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] CCP4 for twinned crystals?

[Randy Read]

Dear Jacob,

For the molecular replacement part, twinning is not a great barrier.  We've 
generally found that, as long as you get the space group right (which can be an 
issue with twinned crystals), the signal in the MR search is not that badly 
degraded.  You could easily set up a search using different possible calmodulin 
conformers as alternatives and see which gives some signal (if any).

The signal in SAD phasing is probably affected worse, but there are still 
plenty of examples where people have succeeded with twinned crystals.  I don't 
have a list of references at hand, but people on the BB probably can point you 
at some of them.  However, if you're able to get a convincing MR solution then 
MR-SAD should work.

You didn't mention whether or not your crystal suffers from translational NCS.  
With the new version of Phaser that handles tNCS this is arguably an advantage, 
as there are fewer truly independent copies to find.  By coincidence, we 
published a paper last year, together with Mariusz Jaskolski, Zbyszek Dauter 
and their colleagues, on a structure that combined tetartohedral twinning and 
7-fold tNCS: http://journals.iucr.org/d/issues/2014/02/00/yt5061/index.html.

Best wishes,

Randy Read

On 2 Feb 2015, at 16:16, Keller, Jacob kell...@janelia.hhmi.org wrote:

 Dear Crystallographers,
 
 Is there any software in CCP4 which can solve twinned structures? I have 
 several datasets which appear to be tetartohedrally twinned, with possible 
 spacegroup of p31/p32 masquerading as p6222/p6422. I think this is 
 approximately equivalent to 1/8-fold NCS?
 
 The data quality (Rpim, CC1/2) is great and resolution is good at 1.9 Ang, 
 although cut off by detector edge I/sig = ~4. Also, I have a possible SAD 
 dataset with CsI soaked in, measured at CuKa, as well as [way too many] 
 possible MR models (calmodulin, which has a ton of different conformers in 
 the PDB). A major problem is that the cell is pretty big, ~100x100x130, so 
 this means ~8 copies, and with tetartohedral twinning, this is more like 64 
 monomers I think. I am thinking to try MR-SAD, with some brute-force 
 search-model testing, but I am not sure whether/where this could be done in 
 CCP4. Any suggestions in terms of approaches and/or software?
 
 BTW, additive screens to try to grow untwinned crystals did not pan out, and 
 I am pretty much out of protein. I do have a few more nice crystals to play 
 with, but they all have the same morphology.
 
 Jacob
 
 
 ***
 Jacob Pearson Keller, PhD
 Looger Lab/HHMI Janelia Research Campus
 19700 Helix Dr, Ashburn, VA 20147
 email: kell...@janelia.hhmi.org
 ***

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] PHASER MR solution

[Randy Read]

Hi all,

Actually, what I think has happened here is that Phaser has limited the 
resolution to what it estimated was needed to get a clear enough solution, and 
then the final refinement was carried out with all the data.  The 
signal-to-noise is lower with limited data, but it’s a price that seems worth 
paying to get the answer more quickly.  The final refined LLG of 1848 is a very 
large number, and the TFZ-equivalent calculation indicates that, if the 
translation search had been carried out with the final refined orientation 
using all the data, the TFZ would have been a very respectable 30.2.  That’s 
probably a better number to look at now, with the new adaptive Phaser searches.

Of course, I agree completely that you would also like to see that the correct 
space group gave a much higher score with the same data than the alternative 
space groups!  Phaser chose a single solution, so it was probably at least 
reasonably clear.

We’re stlll working on how to give the clearest possible indication at the end 
of a run as to whether you should be confident that it is correct.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 28 Jan 2015, at 19:33, Matthew Franklin mfrank...@nysbc.org wrote:

 Hi Jeorge -
 
 Something seems to have changed for the worse in this MR run.  In your 
 earlier posting, where you failed to find a solution in P222, your log file 
 had solutions with RFZ=23.4, generally a clear indication of a correct 
 rotation function solution.  (You didn't include this log file as text, so I 
 can't extract the relevant lines to show you.)  But in this run, your log 
 file shows a solution with RFZ=5.7, which is much more marginal.  The space 
 group choice for the translation function shouldn't affect the results of the 
 rotation function.  What else did you vary - is the search model different?
 
 In addition to what Roger suggested, other indications that you have the 
 right solution come from the translation function tables.  You should have a 
 very clear distinction between the correct solution and incorrect solutions, 
 like this:
 
Fast Translation Function Table: Space Group C 1 2 1

#SET #TRIAL  Top(Z)Second(Z) Third(Z)Ensemble
   1  1  4328.56 (29.62)-  - -  -3gi8_fab
   1  2  4294.35 (26.16)-  - -  -3gi8_fab
   1  3  4252.29 (27.13)-  - -  -3gi8_fab
   1  4  3746.99 (13.34)  3725.62 (12.73)-  -3gi8_fab
   1  5  3467.72 ( 5.02)  3467.43 ( 5.02)  3464.58 ( 4.94)   3gi8_fab
   1  6  3545.59 ( 6.93)  3482.22 ( 5.27)-  -3gi8_fab
 
 Trials 1-3 are correct, while trials 5-6 are incorrect.  (Trial 4 is probably 
 also correct, but wasn't tested further by Phaser.)
 
 You also want to check that the packing test did not throw out one of these 
 high-scoring translation solutions (e.g. your RFZ=23.4 solution).  If that 
 happened, this could mean you need to trim away some loops in your search 
 model.
 
 Good luck,
 
 Matt
 
 
 
 On 1/28/15 12:25 PM, jeorgemarley thomas wrote:
 Hi, all
 
 As per the suggestions, I hv done with the phaser MR and the solution has 
 come with screw axes P 21 21 21. here I am attaching the output text from Mr 
 and sol file. So Now should I go ahead with this? Please suggest. 
 
 Thank you very much in advance !
 
 On Tue, Jan 27, 2015 at 9:33 PM, jeorgemarley thomas kirtswab...@gmail.com 
 wrote:
 Thank you so much to all for your kind concern.
 
 
 
 Jeorge
 
 On Mon, Jan 26, 2015 at 5:55 PM, Kay Diederichs 
 kay.diederi...@uni-konstanz.de wrote:
 Dear Jeorge,
 
 you'll find some information about this in 
 http://strucbio.biologie.uni-konstanz.de/xdswiki/index.php/Space_group_determination
  . A practical and easy way is described in 
 http://strucbio.biologie.uni-konstanz.de/xdswiki/index.php/Pointless
 
 HTH,
 
 Kay
 
 On Mon, 26 Jan 2015 11:24:27 +0100, Tim Gruene t...@shelx.uni-ac.gwdg.de 
 wrote:
 
 Dear Jeorge,
 
 XDS make no claim to determine the SPACE GROUP but rather the LAUE
 GROUP, as only the latter is taken into account during data integration.
 
 This is definitely so during the indexing step (IDXREF.LP), but even in
 CORRECT, when systematic absences are sometimes indicated, XDS does not
 really choose the space group.
 
 Best,
 Tim
 
 On 01/26/2015 05:46 AM, jeorgemarley thomas wrote:
  Hello Dr. Randy
  Here is the IDXREF.LP I got in which, on the basis of quality of fit, I
  went for this space group well I would also try

Re: [ccp4bb] phase shift vs non-isomorphism

[Randy Read]

Hi,

Bart Hazes' sftools program does map correlation in resolution shells in this 
way, and may even have been doing it before 1996 if I'm remembering correctly.

Randy

On 27 Jan 2015, at 21:12, Alexandre OURJOUMTSEV sa...@igbmc.fr wrote:

 Dear Phil,
 
 Gerard Bricogne pointed out a long time ago that the clearest comparison 
 between two sets of phases is the complex correlation coefficient between two 
 best structure factors ( m F exp(i phi) ) - this is equivalent to map 
 correlation, but can be analysed in resolution bins.
 
 I'm not sure whether there are programs which calculate this, though it's 
 straightforward, and I've had various jiffy programs in the past (which I've 
 probably lost)
 
 A while ago this was formally reported, including technical details, by Lunin 
  Lunina (1996), Acta Cryst., A52, 365-368, and they had at that moment some 
 fortran program. In particular, this algorithm makes a key part of the 
 low-resolution direct phasing approaches.
 
 With best regards,
 
 Sacha Urzhumtsev

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] PHASER MR solution

[Randy Read]

Dear Jeorge,

I can’t see anything in that output saying whether the systematic absences that 
would allow you to detect screw axes were covered in your data collection.

In any case, it’s not necessary to reintegrate.  All that happens when going 
from P222 to any one of the other 7 possible orthorhombic space groups is that 
some of the reflections will become systematically absent so they will be 
discarded, but Phaser will do that internally when testing all the space groups.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 26 Jan 2015, at 04:46, jeorgemarley thomas kirtswab...@gmail.com wrote:

 Hello Dr. Randy 
 Here is the IDXREF.LP I got in which, on the basis of quality of fit, I went 
 for this space group well I would also try for the other screw axes. So 
 should I Integrate the data from beginning with all possible screw axes of 
 orthogonal space group?  I am attaching the IDXREF.LP screen shot here.
 
 Jeorge
 
 On Sun, Jan 25, 2015 at 5:00 PM, Roger Rowlett rrowl...@colgate.edu wrote:
 Did you search all 8 possibilities of screw axes, e.g. P2221, P21212, 
 P212121, etc?
 
 Roger Rowlett
 
 On Jan 25, 2015 4:50 AM, jeorgemarley thomas kirtswab...@gmail.com wrote:
 Hi, I have processed the data using XDS and space group found to be P 2 2 2 
 (16) and I used the phaser MR for first phase determination. The model I have 
 used has has more than 70 % sequence identity, when I run the phaser I got 
 the message which I have attached here. And only sum. file I got as an 
 output. Does any one have suggestion what should I do ? I would highly 
 appreciate your kind suggestions. Thank you in advance. 
 
 
 
 IDXREF.PNGSpaceG.PNG

[ccp4bb] Free X-ray diffraction setup

[Randy Read]

Dear all,

We are disposing of an old data collection system comprising a Rigaku RUH3R 
rotating anode X-ray generator, a Mar345 image plate detector and Osmic Blue 
optics, which was installed in 1999 but is still in working order.  If anyone 
is interested in it, either as a working setup or as a source of spare parts, 
you just need to arrange for its removal from our building and shipping to 
yours, at your expense.

Space is at a premium, so if nothing is arranged in the next week we will be 
disposing of it through recycling.

If you are interested, please contact Mark Lethbridge at mw...@cam.ac.uk.

Best wishes,

Randy Read

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] CCP4 Scalepack2mtz problem: Anisotropy correction failed

[Randy Read]

Hi,

Since you posted the whole data set to the BB (a link to a file-sharing website 
might have been better), I took a look at it with other tools.  Perhaps I’m 
running a later version of phenix.xtriage, as in my hands a moderate amount of 
anisotropy is found.  Similarly, Phaser finds a moderate amount of anisotropy.

I suspect what you have uncovered here is a bug in the version of ctruncate 
that you were using, leading to the negative eigenvalues.  Your data set is 
relatively incomplete, and the incompleteness itself is highly anisotropic, 
which can be seen for instance with hklview in CCP4.  However, I’m not getting 
the same results, using version 1.15.10 of ctruncate from CCP4-6.4 on a Mac (I 
haven’t updated yet to 6.5).  Which version of ctruncate and which operating 
system were you using?

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 13 Jan 2015, at 00:34, Xiao Lei xiaolei...@gmail.com wrote:

 Hi Eleanor,
 
 Thank you for the advice.  The .sca file is here. The data may not be useful 
 as you said, but the anisotropic analysis for this data is confusing. From 
 the phenix xtriage output, it seems the data has no anisotropic issue (I may 
 interpret wrong the log file, I attach part of the xtriage log file here). 
 But from CCP4, I know the data has anisotropic problem (Negative 
 Eigenvalues). 
 
 Below is part of log file from Phenix.
 
 Anisotropy analyses 
 
 Anisotropy( [MaxAnisoB-MinAnisoB]/[MaxAnisoB] ) :  3.233e-01
   Anisotropic ratio p-value :  0.000e+00
 
  The p-value is a measure of the severity of anisotropy as observed in 
 the PDB.
  The p-value of 0.000e+00 indicates that roughly 100.0 % of datasets 
 available in the PDB have
  an anisotropy equal to or worse than this dataset.
 
 Xiao
 
 On Sun, Jan 11, 2015 at 7:51 AM, Eleanor Dodson eleanor.dod...@york.ac.uk 
 wrote:
 Well - I really need to see the output.sca file with the reflection list to 
 test the problem properly, but this log file does show your reflections at 
 the highest resolution are very weak, and maybe not useful. Try cutting the 
 resolute a bit and see what happens.
 Eleanor
 
 On 11 January 2015 at 00:08, Xiao Lei xiaolei...@gmail.com wrote:
 Hi Eleanor,
 
 Thanks for the input, I attach the scale log file here. 
 
 On Sat, Jan 10, 2015 at 2:16 PM, Eleanor Dodson eleanor.dod...@york.ac.uk 
 wrote:
 Certainly a negative eigen value is bad - can you attach the data? There may 
 be some obvious problem..
 Eleanor
 
 On 10 January 2015 at 04:42, Xiao Lei xiaolei...@gmail.com wrote:
 Dear All,
 
 I tried to convert my x-ray diffraction sca data from HKL200 (.sca file) to 
 mtz in CCP4 using Scalepack2mtz and it failed, I do not know what should I 
 supposed to do next to correct the problem, any suggestions are appreciated. 
 I pasted the message from part of log file below:
 
 
 ANISOTROPY ANALYSIS (using intensities):
 
 
 Eigenvalues: -0.5668 0.1479 0.3584
 
 Eigenvalue ratios: -1.5815 0.4128 1.
 
 ctruncate: Anisotropy correction failed - negative eigenvalue.
 
 Times: User: 0.4s System: 0.0s Elapsed: 0:00
 
 ***
 
 * Information from CCP4Interface script
 
 ***
 
 The program run with command: /Applications/ccp4-6.4.0/bin/ctruncate -hklin 
 /LAB/CCP4/TEMP/RelADD_12212014ALS502_3_1_mtz.tmp -hklout 
 /LAB/CCP4/TEMP/RelADD_12212014ALS502_3_3_mtz.tmp -colin 
 /*/*/\[IMEAN,SIGIMEAN\] -colout P1_12
 
 has failed with error message
 
 ctruncate: Anisotropy correction failed - negative eigenvalue.
 
 ObjectCache: Leaked 0005 refs to
 
 ***
 
 
 
 #CCP4I TERMINATION STATUS 0  ctruncate:  Anisotropy correction failed - 
 negative eigenvalue. ObjectCache: Leaked 0005 refs to 
 
 
 
 #CCP4I TERMINATION TIME 09 Jan 2015  19:55:50
 
 #CCP4I MESSAGE Task failed
 
 
 
 
 
 
 P1_12_P1_rej2.6_output.sca

Re: [ccp4bb] molecular replacement with poor model

[Randy Read]

Dear Ursula,

18% identity is really in the twilight zone for molecular replacement, where it 
may work but there are certainly no guarantees.

However, there’s a feature in Phaser that is useful for this kind of problem, 
i.e. the “rotate around” option, where you take advantage of knowing the 
approximate orientation of a domain.  In the current version of Phaser in 
either CCP4 or Phenix, this can be done as part of an automated search (instead 
of doing separate rotation/translation/packing/refinement steps as in earlier 
versions).  Let’s assume that the model from the first fragment was cut out of 
a larger model, from which you would now like to place the second fragment.  
The rotation to apply to the second one should be similar to what was applied 
to the first.  In ccp4i, you can go to the “Additional Search Parameters” 
folder, and choose “Brute” for the “Rotation function target”.  In the “Search” 
pulldown that appears then, choose “Around an angle”, give the Euler angles 
that were found in the search for the first domain, and choose something like 
20-30 degrees for “Range”, i.e. allow the second domain to differ in 
orientation by up to 20 or 30 degrees from the orientation for the first 
domain.  It’s probably better to do translation searches with all orientations 
generated by this rotation search, so open the “Expert parameters” folder, 
under “Rotation search peak selection” choose “All peaks”, and turn “Rotation 
clustering” “Off”.

We’ve had good luck in a number of cases where the hinge angle changes by more 
than even the best rigid-body refinement would manage.

Good luck!

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 12 Dec 2014, at 21:38, Ursula Schulze-Gahmen uschulze-gah...@lbl.gov wrote:

 I am trying molecular replacement with a very poor model. The model consists 
 mainly of 1 long helix and two slightly bent antiparallel helices. After 
 dividing it into 2 fragments, I was able to find a solution for one of the 
 fragments ( at least I think so after looking at maps, packing, refinement 
 etc). But even if I place the first solution as fixed ensemble in phaser, I 
 cannot find a solution for the second fragment ( 18% sequence identity). From 
 the structure of the model and the packing it seems clear where the fragment 
 should go roughly.
 
 Are there any other programs other than phaser that might be able to solve 
 this problem? I tried already epmr and mr_rosetta without success. 
 I also tried to just superimpose the complete model onto the partial 
 solution. This results in quite nice packing, but doesn't refine. Is there a 
 rigid program refinement program with very large convergence?
 
 Ursula
 
 -- 
 Ursula Schulze-Gahmen, Ph.D.
 Project Scientist
 UC Berkeley, QB3
 360 Stanley Hall #3220
 Berkeley, CA 94720-3220
 (510) 643 9491

Re: [ccp4bb] Challenging Mol.Rep. Problem

[Randy Read]

Hi,

We’ve had good luck with Phaser in placing many copies of proteins or domains, 
if the model is reasonably good.  Because of the variable elbow angle, you 
probably want to place the domains separately.  However, as soon as you find 
one convincing assembly you can switch to searching for the whole Fab.  
Alternatively, you can try a bunch of different models and hope that one is 
right.

Something that might be useful is a new “MR_GYRE feature in Phaser.  This is a 
likelihood-based version of the PC refinement that can be done in CNS, to 
optimise the relative orientations of domains between the rotation and 
translation searches.  If you get stuck, we might be able to give a hand — we 
haven’t had much experience ourselves with using this on unsolved structures.

Of course, the advice you’ve had from others to be certain first of the space 
group is very good!

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 9 Dec 2014, at 18:31, Mo Wong mowon...@gmail.com wrote:

 Hi all,
 
 I’m stuck on a rather complex molecular replacement problem. The crystals are 
 of an antigen-Fab complex totaling ~67 kDa (waiting to confirm using PAGE 
 gel). They diffract to ~3.5A at the synchrotron and process in C2 with 
 dimensions 220x130x230 and beta at 103 so it looks like there are round 
 9to12-ish copies in the ASU. The overall Rmerge is high at ~25% with I/sig 
 cutoff ~2 and redundancy of 5; however, at 4.5A this drops to ~15%. 
 Furthermore, processing in P1 gives similar Rmerge values too.
 
 Self-Patterson doesn’t suggest translational symmetry, but the self-rotation 
 function (SRF) suggests high NCS (see below/attached).
 
 I’m hoping the SRF might be helpful in trying to confirm/dismiss C2 is the 
 likely space group, and perhaps suggest a logical assembly with the ASU (I 
 see strong 2-fold and 3-fold NCS operators suggesting to me dimeric trimers 
 or vice versa - however, I’ve never had to really analyze SRFs in the absence 
 of a mol. rep. solution so my interpretation could be wrong).
 
 Anyway, any help to bringing a molecular replacement solution closer to 
 reality would be appreciated.
 
 Thanks!
 
 SELF_RF.gif

Re: [ccp4bb] Intensities and amplitudes

[Randy Read]

Dear George,

Yes, we’ve developed new likelihood functions that work with intensity data.  
They’re already available for the molecular replacement calculations in recent 
nightly-build versions of Phaser (though it’s been a while since a new nightly 
was released, and we’ve fixed a few problems with outliers that were more 
extreme than we had anticipated encountering).  I’ll be presenting our work on 
the intensity-based SAD likelihood target at the upcoming CCP4 Study Weekend.

It’s possible to define exact intensity-based likelihood functions (at least 
“exact” when the measurement errors are Gaussian in the observed intensity), 
but we haven’t found a way of evaluating them without either numerical 
integration (expensive) or approximation.  However, we’ve got a new 
approximation that turns out to be excellent over the whole range from small to 
extremely large intensity errors, and which is very efficient to work with.

Best wishes,

Randy

On 3 Dec 2014, at 15:07, George M. Sheldrick gshe...@shelx.uni-ac.gwdg.de 
wrote:

 Dear Randy,
 
 I could not agree more. Statistical methods for phasing and refinement
 must be better using the observed intensities and their esds than with
 (c)truncated F-values. In particular one should merge intensities, not
 truncated Fs!
 
 To elaborate on Harry's comment, when SHELXL started refining only
 against intensities 22 years ago, I received many complaints from irate
 small molecule crystallographers whose papers had been rejected because
 the unweighted R-factors R2 (based on intensities) were too high. I even
 sent a letter to editors of the journals involved to explain why
 R-factors based on intensities are at least twice as high as those based
 on F, but to no avail. So I had to output R1 (the unweighted R-value
 based on F) even though the structure had been refined against
 intensitites, then everyone was happy.
 
 Do I correctly understand that you have developed new (better) maximum
 likelihood criteria for use with I rather than F?
 
 Best wishes, George
 
 
 
 On 12/02/2014 10:26 PM, Randy Read wrote:
 Dear Mohamed,
 
 At the moment, a lot of programs require amplitudes, but I believe that they 
 should all eventually be updated to use intensities.  In fact, we’re in the 
 end stages of a large project to switch Phaser from using amplitudes to 
 using intensities.  There are a number of reasons why, in principle, it’s 
 better to work in terms of intensities.  One is that it’s perfectly 
 reasonable to have a negative observed intensity, which can come from 
 subtracting a background estimate with measurement errors from a very weak 
 peak with its own measurement errors.  That, of course, is where the French 
 and Wilson algorithm comes in, allowing an amplitude to be estimated without 
 simply taking a square root.  However, the problem with the French and 
 Wilson algorithm is that it loses information, i.e. you can’t reconstruct 
 the intensity and its standard deviation.  What you get out of French  
 Wilson depends on the estimate of the expected intensity for a reflection, 
 which is typically taken from t
 he mean in the resolution shell but should vary with direction for crystals 
 suffering from anisotropic diffraction and should be modulated for crystals 
 with translational non-crystallographic symmetry.
 
 Another reason it’s better to work in terms of intensities is that it’s 
 reasonable to assume that the measurement errors for intensities are 
 Gaussian, but then less reasonable to assume that for amplitudes 
 (particularly with the problem that amplitudes can’t be negative).
 
 For now, you need amplitudes for a lot of purposes and then the French  
 Wilson algorithm is useful.  But what I would strongly recommend is that you 
 hang on to the intensities and you make sure that the intensities are 
 deposited at the PDB.  It’s a pity that many PDB depositions only have 
 amplitudes that have been through French  Wilson, so that new procedures 
 based on intensities won’t be able to be applied with their full power.
 
 Best wishes,
 
 Randy Read
 
 -
 Randy J. Read
 Department of Haematology, University of Cambridge
 Cambridge Institute for Medical ResearchTel: +44 1223 336500
 Wellcome Trust/MRC Building Fax: +44 1223 336827
 Hills Road
 E-mail: rj...@cam.ac.uk
 Cambridge CB2 0XY, U.K.   
 www-structmed.cimr.cam.ac.uk
 
 On 1 Dec 2014, at 20:49, Mohamed Noor mohamed.n...@staffmail.ul.ie wrote:
 
 Dear crystallographers
 
 Is there any reason for using one data type over the other? Are there any 
 errors associated with the French and Wilson I-to-F conversion step?
 
 Thanks.
 Mohamed
 
 
 -- 
 Prof. George M. Sheldrick FRS
 Dept. Structural Chemistry,
 University of Goettingen,
 Tammannstr. 4,
 D37077 Goettingen, Germany
 Tel. +49-551-39-33021 or -33068
 Fax. +49-551-39-22582

--
Randy J. Read
Department

Re: [ccp4bb] Intensities and amplitudes

[Randy Read]

Dear Mohamed,

At the moment, a lot of programs require amplitudes, but I believe that they 
should all eventually be updated to use intensities.  In fact, we’re in the end 
stages of a large project to switch Phaser from using amplitudes to using 
intensities.  There are a number of reasons why, in principle, it’s better to 
work in terms of intensities.  One is that it’s perfectly reasonable to have a 
negative observed intensity, which can come from subtracting a background 
estimate with measurement errors from a very weak peak with its own measurement 
errors.  That, of course, is where the French and Wilson algorithm comes in, 
allowing an amplitude to be estimated without simply taking a square root.  
However, the problem with the French and Wilson algorithm is that it loses 
information, i.e. you can’t reconstruct the intensity and its standard 
deviation.  What you get out of French  Wilson depends on the estimate of the 
expected intensity for a reflection, which is typically taken from the mean in 
the resolution shell but should vary with direction for crystals suffering from 
anisotropic diffraction and should be modulated for crystals with translational 
non-crystallographic symmetry.

Another reason it’s better to work in terms of intensities is that it’s 
reasonable to assume that the measurement errors for intensities are Gaussian, 
but then less reasonable to assume that for amplitudes (particularly with the 
problem that amplitudes can’t be negative).

For now, you need amplitudes for a lot of purposes and then the French  Wilson 
algorithm is useful.  But what I would strongly recommend is that you hang on 
to the intensities and you make sure that the intensities are deposited at the 
PDB.  It’s a pity that many PDB depositions only have amplitudes that have been 
through French  Wilson, so that new procedures based on intensities won’t be 
able to be applied with their full power.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 1 Dec 2014, at 20:49, Mohamed Noor mohamed.n...@staffmail.ul.ie wrote:

 Dear crystallographers
 
 Is there any reason for using one data type over the other? Are there any 
 errors associated with the French and Wilson I-to-F conversion step?
 
 Thanks.
 Mohamed

Re: [ccp4bb] Intensities and amplitudes

[Randy Read]

Hi Pavel,

We were chatting with Phil Evans the other day about things like this, and 
generally we were in agreement that any programs that need amplitudes (and 
you’re right of course, you have to have some sort of amplitude to calculate a 
map!) should be able to compute them on the fly.  That may well involve the 
French  Wilson algorithm, but can take advantage of whatever is understood by 
the program (e.g. anisotropy and translational non-crystallographic symmetry, 
both of which in principle can be modeled better as the atomic model improves).

I haven’t really worried about R-factors.  We could learn to embrace the 
R-factor on intensity that small molecule crystallographers are comfortable 
with but, as you say, people are not used to these.  If we compute amplitudes 
on the fly, with a French  Wilson algorithm that is calibrated better as the 
model improves, the R-factors will be calculated with a changing set of Fobs.  
This would probably be a minor effect, but it’s slightly disconcerting.

Randy

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 2 Dec 2014, at 21:44, Pavel Afonine pafon...@gmail.com wrote:

 Hi Randy,
 
 I can see all good reasons for using intensities! What about maps and 
 R-factors? I guess you still need F to compute them (I realize you can 
 compute R(I) but this is not what people are used to do in general), and if 
 that's the case then I-F is still inevitable (at least for some purposes).
 
 Thanks,
 Pavel
 
 On Tue, Dec 2, 2014 at 1:26 PM, Randy Read rj...@cam.ac.uk wrote:
 Dear Mohamed,
 
 At the moment, a lot of programs require amplitudes, but I believe that they 
 should all eventually be updated to use intensities.  In fact, we’re in the 
 end stages of a large project to switch Phaser from using amplitudes to using 
 intensities.  There are a number of reasons why, in principle, it’s better to 
 work in terms of intensities.  One is that it’s perfectly reasonable to have 
 a negative observed intensity, which can come from subtracting a background 
 estimate with measurement errors from a very weak peak with its own 
 measurement errors.  That, of course, is where the French and Wilson 
 algorithm comes in, allowing an amplitude to be estimated without simply 
 taking a square root.  However, the problem with the French and Wilson 
 algorithm is that it loses information, i.e. you can’t reconstruct the 
 intensity and its standard deviation.  What you get out of French  Wilson 
 depends on the estimate of the expected intensity for a reflection, which is 
 typically taken from the mean in the resolution shell but should vary with 
 direction for crystals suffering from anisotropic diffraction and should be 
 modulated for crystals with translational non-crystallographic symmetry.
 
 Another reason it’s better to work in terms of intensities is that it’s 
 reasonable to assume that the measurement errors for intensities are 
 Gaussian, but then less reasonable to assume that for amplitudes 
 (particularly with the problem that amplitudes can’t be negative).
 
 For now, you need amplitudes for a lot of purposes and then the French  
 Wilson algorithm is useful.  But what I would strongly recommend is that you 
 hang on to the intensities and you make sure that the intensities are 
 deposited at the PDB.  It’s a pity that many PDB depositions only have 
 amplitudes that have been through French  Wilson, so that new procedures 
 based on intensities won’t be able to be applied with their full power.
 
 Best wishes,
 
 Randy Read
 
 -
 Randy J. Read
 Department of Haematology, University of Cambridge
 Cambridge Institute for Medical ResearchTel: +44 1223 336500
 Wellcome Trust/MRC Building Fax: +44 1223 336827
 Hills RoadE-mail: 
 rj...@cam.ac.uk
 Cambridge CB2 0XY, U.K.   
 www-structmed.cimr.cam.ac.uk
 
 On 1 Dec 2014, at 20:49, Mohamed Noor mohamed.n...@staffmail.ul.ie wrote:
 
  Dear crystallographers
 
  Is there any reason for using one data type over the other? Are there any 
  errors associated with the French and Wilson I-to-F conversion step?
 
  Thanks.
  Mohamed
 

[ccp4bb]

[Randy Read]

Hi,

Working with Phaser in the presence of tNCS is not necessarily as automatic as 
in the absence of tNCS.  It can be rather complicated, and not all decisions 
are necessarily made correctly.  So I would suggest reading some new 
documentation on this: 
http://www.phaser.cimr.cam.ac.uk/index.php/Molecular_Replacement#Translational_Non-crystallographic_Symmetry.

In your case, if a) the accepted solution has a higher TFZ and LLG than the 
rejected ones and b) there isn’t the possibility of more than two tNCS-related 
copies, then there’s a good chance the accepted solution is correct.  You could 
send me a full logfile offline and I could tell you whether there’s anything 
you should note in there.

Best wishes,

Randy Read

On 28 Nov 2014, at 13:15, rohit kumar rohit...@gmail.com wrote:

 Dear all,
 
 i am solving a structure of 2.5 A resolution by PHASER but every time i got 
 this following warning.
 
 $TEXT:Warning: $$ Baubles Markup $$
 -
 Large non-origin Patterson peak indicates that translational NCS is present:
 correction factors applied
 -
 $$
 $TEXT:Warning: $$ Baubles Markup $$
 -
 Solutions with Z-scores greater than 49.2 (the threshold indicating a definite
 solution) were rejected for failing packing test
 #1 TFZ=65.6 PAK=42
 #2 TFZ=65.5 PAK=68
 -
 however every time PHASER run successfully and it give a output PDB as a 
 solution. is it a right solution or not?  or To get the right solution what 
 should i do? 
 
 if i used accept all the solution in expert parameter, i also got the 
 solution but which does not follow the Packing test.
 
 Can anyone suggest...
 
 -- 
 WITH REGARDS
 Rohit Kumar Singh
 Lab. no. 430,
 P.I. Dr. S. Gourinath,
 School of Life Sciences,
 Jawaharlal Nehru University
 New Delhi -110067

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] R free flag missing after ARP/wARP?

[Randy Read]

When you chose the option “Do not use the Free R flag”, then you were telling 
Refmac not to use cross-validation and therefore to use all reflections as the 
working set.

My experience (doubtless very limited compared to the ARP/wARP developers) is 
that it’s significantly better to use the Rfree flag, which makes sense since 
you need cross-validation data in order for likelihood targets to be calibrated 
properly.  But maybe the developers have some set of test cases that led them 
to choose the opposite as the default.

Best wishes,

Randy Read

On 20 Oct 2014, at 14:24, luzuok luzuo...@126.com wrote:

 
 Dear Colin and Tim,
 But ARPwARP uses REFMAC5 for refinement, does this means that REFMAC5 
 uses all reflections as working set?
 How to validate that the refinement is not over fit?
 
 Best wishes!
 
 Lu  Zuokun
 
 
 
 --
 卢作焜
 南开大学新生物站A202
 
 
 At 2014-10-20 17:38:58, Tim Gruene t...@shelx.uni-ac.gwdg.de wrote:
 Dear Lu Zuokun,
 
 since you can use the output mtz-file from ArpWarp for model building,
 but not for refinement, there is no need to include the Rfree flag in
 its output file. Maybe the omission is a deliberate caveat to the users
 to pay attention to this.
 
 Best regards,
 Tim
 
 
 On 10/20/2014 04:41 AM, luzuok wrote:
  Dear all,
  I was using ARP/wARP in ccp4i, the input mtz file certainly has free R 
  flag, but the output mzt file doesn't have the free R flag label? 
  I chose  do not use the Free R flag in the ARP/wARP GUI. Can anyone tell 
  me what's wrong with this? 
  
  
  best reagrds!
  Lu Zuokun
  
  
  
  --
  卢作焜
  南开大学新生物站A202
  
 
 -- 
 Dr Tim Gruene
 Institut fuer anorganische Chemie
 Tammannstr. 4
 D-37077 Goettingen
 
 GPG Key ID = A46BEE1A
 
 
 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Molecular Replacement model preparation

[Randy Read]

Dear Scott,

By “AND” searching, presumably you mean adding another SEARCH command? (Or 
clicking “Add another search” in the ccp4i interface.)  For Steven’s strategy, 
you want to specify separate searches for separate ensembles, i.e. give several 
SEARCH commands in a script or add extra searches in ccp4i.

The “OR” option (to specify several ensembles to search for in one SEARCH 
command, available through ccp4i by opening the “Additional Search parameters” 
pane and turning on “Allow search with alternative ensembles...” uses 
alternative models to search for the same component, which can be useful but 
isn’t what you’re interested in for defining a hinge angle.

At some point in the distant past, you had to tell Phaser what order to search 
for the components in, but then we implemented a method to automatically choose 
a good search order and, more recently, implemented an improved method to 
choose the optimal search order.  If the information you give Phaser about the 
quality of the model (RMSD or sequence identity, which Phaser turns into an 
RMSD estimate) is correct, then the strength of the signal for different models 
can be estimated.  Phaser will use this to define an initial search order.  If 
each search gives an unambiguous solution, then everything will proceed in the 
predefined order.  However, whenever there is ambiguity about whether a correct 
solution has been found, then Phaser will automatically try strategies like 
choosing a different search order or changing the resolution of the data used 
for the calculation.  

So, in most cases, it’s best to tell Phaser everything you’re looking for and 
set up one job to find everything, because this will allow the greatest level 
of optimisation and the most flexibility in trying different strategies.  It’s 
only if this fails that you’ll want to choose the strategy and parameters 
manually.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 7 Oct 2014, at 15:16, Scott Thomas Walsh swals...@umd.edu wrote:

 Hi Steven,
 
 Thank you for the information and guidance.  When you search for all the 
 ensembles with Phaser do you use “AND” or “OR”
 searching?
 
 Cheers,
 
 Scott
 
 
 While Phil Jeffrey attributed to me the “trick” of aligning the hinge axis 
 of an Fab along the Z direction, I, in turn, must give credit to Mirek 
 Cygler, who explained this to me at the Diffraction Methods in Molecular 
 Biology [now Structural Biology] Gordon Research Conference in 1986.
  
 To Scott’s query about searching for “1 domain sequentially and then other 
 domains OR searching for multiple domains all at once?”
  
 Inevitably a program like PHASER searches for domains (“ensembles” in its 
 terminology) sequentially. However, as to the practical question of whether 
 to “feed” PHASER all of the domains in one run, that is certainly how I 
 start and it is usually (almost always?) successful. In fact, while I no 
 longer bother to align the hinge axis of an Fab along the Z axis, I now 
 break Fabs into three parts: CL:CH1, VH, and VL to allow molecular 
 replacement to accommodate the “tilt” angle between VH and VL (tilt angle is 
 a term I learned from Gary Gilliland’s talk at the Diffraction Methods in 
 Structural Biology GRC in 2014). This also allows me to search for the 
 highest identity VL and, separately, VH in the PDB to use as probe models.  
 N.B. since I’m usually studying antigen/Fab complex, I’m usually searching 
 for 4 ensembles in one PHASER run: CL:CH1, antigen, VH and VL.
  
 
 
 
 Scott T. R. Walsh, PhD
 Assistant Professor
 University of Maryland
 IBBR/CBMG
 3127E CARB-2
 9600 Gudelsky Drive
 Rockville, MD  20850  USA
 phone: (240) 314-6478
 fax: (240) 314-6225
 email: swals...@umd.edu
 
 

Re: [ccp4bb] Phaser question, twinning, DIALS, suggestions welcome

[Randy Read]

Hi Jurgen,

You could send me a logfile off-list, and maybe I would spot something in there.

We’ve put some effort into putting more intelligence into the Phaser search, so 
that it adapts to the initial perceived difficulty of the problem in setting 
the initial parameters, and then adapts to indicators of success or failure 
during the search.  Much of the time this works very well, but there’s 
obviously room for improvement.  For one thing, it appears that we’re 
frequently too optimistic about how good the model will be and how easy the 
search will be.

One question: when you say that you cut at 2.5A for MR, do you do that by 
setting the resolution within the Phaser run, or do you have an MTZ file with 
only data to 2.5A?  If the former, then you’re over-riding some of Phaser’s 
automation, which will choose the initial resolution limit based on the 
perceived difficulty of the problem (a function of model completeness, expected 
RMS error of the model, and the number of reflections to different resolution 
limits).  It’s this initial automated choice that can go wrong if we’re too 
optimistic, because then a clear solution isn’t found, and then Phaser repeats 
the search with data to the full resolution, which can take longer than just 
choosing an intermediate resolution from the start.

Anyway, if the problem is expected to be easy but turns out to be difficult, 
this implies that some of the information used to decide it should be easy is 
wrong or too optimistic.  One top possibility is that the model is not as good 
as expected, e.g. because of conformational changes.  If there’s a potential 
hinge-bending motion, then you’re usually better off searching with separate 
domains.  If the change is something that can’t be described with rigid-body 
motions, then it would be better to increase the expected RMS error from what 
Phaser deduces from the sequence identity.  Increasing the RMSD by 10-20% would 
be a good first bet in such a case.

The other top possibility is that the space group is wrong.  Is there any 
ambiguity in the space group?  In particular, do any of the twinning tests 
indicate that the data may be twinned (which can lead to choosing too high 
symmetry)?

Best wishes,

Randy

On 1 Oct 2014, at 03:02, Jurgen Bosch jbos...@jhu.edu wrote:

 Dear BB, or in particular Phaser developers :-)
 
 This must be part of British humor right (or was that the Canadian influence 
 Randy) ?
 
 eLLG indicates that placement of ensemble ensemble_1 will be straightforward
The data are sufficient to exceed the eLLG target
 
 The search space is finite 143 x 143 x 80 Å with 3 or 4 molecules per asu, 
 but Phaser has been burning CPU cycles quite a bit, we are approaching 24h by 
 now. Data extends to 1.7 Å - for MR we cut at 2.5 Å.
 
 I can hear Garib, yes, Molrep was done in few minutes but I’m not super 
 convinced about the solution either. Same with BALBES and MrBump (which took 
 a few more minutes, actually also days for MrBump)
 
 The space group appears to be P63 22 as judged by XDS, pointless, xtriage, 
 however the crystals are split (at least in some areas it’s visible) but I 
 thought XDS would take care of these “aliens” and eliminate them mostly. My 
 graduate student, Lauren, found a nifty program called DIALS that we wish to 
 explore further to rescue the “nice” data we have and hopefully solve the 
 structure.
 
 Lower symmetry space groups were tried down to P21 with increasing number of 
 molecules per asu and applying twin laws if necessary. The minor problem with 
 the twins is how do I really know that it is a higher symmetry space group 
 and not a 50% twin in a lower symmetry ? Rwork/Rfree in this particular case 
 do not seem to help at all for distinguishing between the solutions. Maps 
 looks sort of right but R-factors are not reflecting what you see on the 
 screen.
 
 We appreciate any suggestions and ideas what else we could do. 
 
 Thanks,
 
 Jürgen
 
 ..
 Jürgen Bosch
 Johns Hopkins University
 Bloomberg School of Public Health
 Department of Biochemistry  Molecular Biology
 Johns Hopkins Malaria Research Institute
 615 North Wolfe Street, W8708
 Baltimore, MD 21205
 Office: +1-410-614-4742
 Lab:  +1-410-614-4894
 Fax:  +1-410-955-2926
 http://lupo.jhsph.edu
 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Phaser question, twinning, DIALS, suggestions welcome

[Randy Read]

Hi again,

I should have mentioned that, if you have a good enough model, it’s often 
possible to solve the structure in P1.  The molecular replacement solution will 
settle on one of the twin domains (or you may end up with more than one 
solution, related by the twin law(s)).  Then the symmetry of the solution will 
tell you the true symmetry of the crystal.  Otherwise, you’ll have to test all 
the subgroups, and there will be a lot of those for P6322, especially if you’re 
not sure about screw axes.

But the true point group symmetry will only be lower than the symmetry in which 
the data merge if there’s twinning, which is why I’m wondering whether there 
are indications of twinning.  It probably bears repeating that, if you test 
data merged in too low symmetry, the twinning tests that depend on twin laws 
can be misleading, because they look for reflections that are similar over a 
possible twin law.  You need to have independent evidence from tests that do 
not depend on twin laws (i.e. statistical tests such as the moment tests or the 
L-test) that the data are perturbed in a way that implies twinning.

Randy

On 1 Oct 2014, at 03:02, Jurgen Bosch jbos...@jhu.edu wrote:

 Dear BB, or in particular Phaser developers :-)
 
 This must be part of British humor right (or was that the Canadian influence 
 Randy) ?
 
 eLLG indicates that placement of ensemble ensemble_1 will be straightforward
The data are sufficient to exceed the eLLG target
 
 The search space is finite 143 x 143 x 80 Å with 3 or 4 molecules per asu, 
 but Phaser has been burning CPU cycles quite a bit, we are approaching 24h by 
 now. Data extends to 1.7 Å - for MR we cut at 2.5 Å.
 
 I can hear Garib, yes, Molrep was done in few minutes but I’m not super 
 convinced about the solution either. Same with BALBES and MrBump (which took 
 a few more minutes, actually also days for MrBump)
 
 The space group appears to be P63 22 as judged by XDS, pointless, xtriage, 
 however the crystals are split (at least in some areas it’s visible) but I 
 thought XDS would take care of these “aliens” and eliminate them mostly. My 
 graduate student, Lauren, found a nifty program called DIALS that we wish to 
 explore further to rescue the “nice” data we have and hopefully solve the 
 structure.
 
 Lower symmetry space groups were tried down to P21 with increasing number of 
 molecules per asu and applying twin laws if necessary. The minor problem with 
 the twins is how do I really know that it is a higher symmetry space group 
 and not a 50% twin in a lower symmetry ? Rwork/Rfree in this particular case 
 do not seem to help at all for distinguishing between the solutions. Maps 
 looks sort of right but R-factors are not reflecting what you see on the 
 screen.
 
 We appreciate any suggestions and ideas what else we could do. 
 
 Thanks,
 
 Jürgen
 
 ..
 Jürgen Bosch
 Johns Hopkins University
 Bloomberg School of Public Health
 Department of Biochemistry  Molecular Biology
 Johns Hopkins Malaria Research Institute
 615 North Wolfe Street, W8708
 Baltimore, MD 21205
 Office: +1-410-614-4742
 Lab:  +1-410-614-4894
 Fax:  +1-410-955-2926
 http://lupo.jhsph.edu
 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] Phaser MR problem

[Randy Read]

Just to add to what Herman said:

The statistics are good for placing the domain represented by ensemble 1 
(TFZ=14.3) and the first copy of the domain represented by ensemble 2 
(TFZ=11.9), but not for the two possible solutions for placing the second copy 
of ensemble 2 (TFZs of 4.5 and 4.7).  So it’s possible that only the placement 
of the first two components is correct.  If you turn on symmetry in coot, do 
those two domains come together to form a sensible whole protein?  If so, you 
could try using that composite model to look for more copies of the whole 
protein (keeping in mind Herman’s point about the possible numbers of copies).

That said, it’s going to be a challenge to finish up the structure rebuilding 
and refinement when you have limited resolution and relatively low sequence 
identity starting models.  With 3-fold or greater NCS, averaging would help a 
great deal (2-fold helps somewhat but is less powerful); otherwise, you’re 
likely to need some additional phase information.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 19 Sep 2014, at 10:33, Veerendra KUMAR (IMCB) 
veerend...@imcb.a-star.edu.sg wrote:

 Dear CCP4 members,
 
 Recently I have collected native data at 3.3 A resolution. The structure of 
 the protein should have two domains. The structure of c terminal domain from 
 homologous (30 seq similarity) is known. I took n terminal domain from 
 another homologous protein. I ran the phaser using these two ensembles and 
 got the following solutions.
 #   [No title given]
 SOLU SET RFZ=3.9 TFZ=14.3 PAK=4 LLG=144 TFZ==10.5 RFZ=2.4 TFZ=11.9 PAK=9 
 LLG=259 TFZ==7.0 RFZ=4.0 TFZ=4.5 PAK=24 LLG=207 TFZ==7.4
 SOLU SPAC I 41 3 2
 SOLU 6DIM ENSE ensemble1 EULER 43.364 14.638 271.356 FRAC 0.81509 0.22245 
 0.92422 BFAC -7.23070
 SOLU 6DIM ENSE ensemble2 EULER 173.495 55.758 234.557 FRAC 1.25797 -0.71934 
 0.10135 BFAC 1.81823
 SOLU 6DIM ENSE ensemble2 EULER 351.235 68.606 254.802 FRAC 0.64931 0.81795 
 0.27592 BFAC 5.82534
 SOLU ENSE ensemble1 VRMS 1.157
 SOLU ENSE ensemble2 VRMS 1.540
 SOLU SET RFZ=3.9 TFZ=14.3 PAK=4 LLG=144 TFZ==10.5 RFZ=2.4 TFZ=11.9 PAK=9 
 LLG=259 TFZ==7.0 RFZ=2.7 TFZ=4.7 PAK=19 LLG=202
 SOLU SPAC I 41 3 2
 SOLU 6DIM ENSE ensemble1 EULER 43.364 14.638 271.356 FRAC 0.81509 0.22245 
 0.92422 BFAC -7.23070
 SOLU 6DIM ENSE ensemble2 EULER 173.495 55.758 234.557 FRAC 1.25797 -0.71934 
 0.10135 BFAC 1.81823
 SOLU 6DIM ENSE ensemble2 EULER 41.452 118.850 315.630 FRAC 0.61025 0.05816 
 1.60902 BFAC 5.47988
 SOLU ENSE ensemble1 VRMS 1.157
 SOLU ENSE ensemble2 VRMS 1.540
 
 The TFZ score 14.3 suggests a solution. Then I did refinement using Ploy Ala 
 phaser model. The R values are 0.590/0.62. I use the map to build the model 
 in Buccaneer but it did not build anything.
 
 Is the pahser solution correct? Why are the R values so high despite the good 
 TFZ score? Any suggestions are greatly appreciated.
 
 Thank you
 
 Veerendra
 
 
 
 Note: This message may contain confidential information. If this Email/Fax 
 has been sent to you by mistake, please notify the sender and delete it 
 immediately. Thank you.

Re: [ccp4bb] Multiple pseudotranslations vectors in P1

[Randy Read]

Dear Thomas,

At the moment, Phaser only deals fully automatically with 2-fold tNCS (i.e. one 
unique peak in the native Patterson), although we’re hoping to make it more 
general.  With some tailored input, it can also deal with the case of several 
copies related by multiples of the same NCS translation.  If there are 
independent NCS translations, then it can’t deal with that.  However, you can 
override the automatic definition of the NCS translations and, for instance, 
provide the translation vector for the dominant one, or you can set the 
threshold for automatica detection of Patterson peaks to just pick one.

When you said Phaser went into an endless loop, do you mean that it was 
exploring a very large number of potential solutions with no signal to 
distinguish among them, or do you mean that it went into an infinite loop 
repeating the same thing?  If the latter, that would be a bug and we’d want to 
see the log file!

If you’d like, you could send the logfile off-line and I could tell you how I 
might approach dealing with tNCS in your case.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 30 Jul 2014, at 17:09, Thomas Krey tk...@pasteur.fr wrote:

 Dear all,
 
 we are currently in the process of solving the structure of a Fab/peptide 
 complex. We have first determined the structure of the native Fab and refined 
 it to ~2A, so we have a decent MR model. The complex crystals that we are 
 working on right now diffract to ~3.2A and are unfortunately in P1 (a 
 different spacegroup than the native Fab) with the cell parametersa=70.72 
   b=91.53  c=206.07   α=95.14   β=99.01   γ=90.83. This large AU contains 
 between 8 and 12 Fab molecules and unfortunaly these are related by pseudo 
 symmetry as suggested by the 4 non-origin distinct peaks detected in the 
 Patterson. Unfortunately this leads to Phaser turning off the 
 pseudotranslation correction and running into one of those endless loops that 
 I stopped eventually after 40 hours.
 Is there anything we can do in order to get around these different 
 pseudotranslations ?
 
 Thank you so much for any help or suggestions.
 
 Best wishes
 
 Thomas
 
 
 
 
 Dr. Thomas Krey
 Institut Pasteur
 Structural Virology Unit
 25-28 Rue du Docteur Roux
 75015 Paris
 France
 tk...@pasteur.fr

Re: [ccp4bb] Problem in molecular replacement

[Randy Read]

In addition to that good suggestion, with a model at such high sequence 
identity and with reasonable resolution data, it would be a good idea to search 
for smaller models, like a single dimer or even a monomer, because the assembly 
may have changed conformation.  Another possibility you should consider is that 
the space group may not be correct.

Good luck!

Randy Read

On 24 Jul 2014, at 13:06, Antony Oliver antony.oli...@sussex.ac.uk wrote:

 Try using DNA as a search model - this has worked very successfully in our 
 hands before.
 
 Tony.
 
 - - - - - - - - - - - - - - - - - -
 Dr Antony W Oliver
 Senior Research Fellow
 CR-UK DNA Repair Enzymes Group
 Genome Damage and Stability Centre
 Science Park Road
 University of Sussex
 Falmer, Brighton, BN1 9RQ
 - - - - - - - - - - - - - - - - - - 
 email: antony.oli...@sussex.ac.uk
 
 tel (office): +44 (0)1273 678349
 tel (lab): +44 (0)1273 677512
 
 http://www.sussex.ac.uk/lifesci/oliverlab
 http://tinyurl.com/aw-oliver
 - - - - - - - - - - - - - - - - - -
 
 On 24 Jul 2014, at 12:50, dusky dew duskyde...@gmail.com wrote:
 
 DEAR ALL
 I am trying to solve the structure of a protein DNA complex with molecular 
 replacement. The resolution in about 2.5 angstrom.  The spacegroup is P21. 
 The search model has about 50% identity and is a dimer of a dimer. The 
 problem is the unit cell is huge and so the number of molecules in asu 
 becomes 4 or 3. I am not finding any solution with phaser/molrep.
 Please suggest.
 Thanks
 Yang zhang
 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] New PDB validation reports

[Randy Read]

Dear Katherine,

Thanks for pointing this out!  As far as I can tell the community has not 
reached a consensus on dealing with disordered side chains.  I’m afraid it 
simply didn’t occur to us, when we were writing the validation task force 
report, that one approach would be favoured over the other, and we certainly 
didn’t intend to influence community practice by stealth!

This unintended consequence really should have occurred to us, since Gerard 
Kleywegt and I had noticed earlier that structural genomics structures 
(particularly those from the SGC) have a systematically lower completeness than 
other structures, apparently because the choice was made more often to 
completely omit poorly-ordered residues rather than include them in the model.  
This leaves out many of the residues that would have poor Ramachandran and 
rotamer scores, thus raising those scores above the average in a similar way.  
(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2631636/)

Some indication of completeness probably would be a good thing to include in 
the validation reports.

Best wishes,

Randy Read

On 10 Jul 2014, at 20:26, Katherine Sippel katherine.sip...@gmail.com wrote:

 Hi all,
 
 I've been playing with the new PDB validation service. It is very pretty and 
 kudos to all the hard work that has clearly gone into it. I did notice 
 however that the way the information is presented, there seems to be a bias 
 towards truncating side chains versus modeling them with higher b-factors. 
 The disordered side chains have higher RSRZs (rightfully so), but there 
 doesn't seem to be any indicator for missing atoms. As a results I can make 
 my validation report prettier by truncating versus modeling with high Bs. 
 
 I don't want to kick an ant pile here, but given this rather significant 
 difference in quality reporting, I was wondering if the community had reached 
 a consensus on this issue that I had missed. 
 
 Cheers,
 Katherine
 
 -- 
 Nil illegitimo carborundum - Didactylos

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] packing test PHASER

[Randy Read]

I agree that looking at the packing is a good idea, but I also agree that 
having the wrong space group is a likely possible explanation.  That’s the most 
common scenario when there are many clashing solutions with high TFZ scores.

Randy Read

On 17 Jun 2014, at 15:52, Roger Rowlett rrowl...@colgate.edu wrote:

 Increase the number of allowed clashes in Phaser, re-run it then look at the 
 packing of the solution found and identify the source of the clashes. 
 Possibilities for the clash issue include:
 Wrong space group
 Flexible loops or termini in search model not present  or differently 
 arranged in your crystal target
 Once you look at the packing in Coot or Pymol, you will have a good idea of 
 what to do next. If the problem is flexible loops or misplaced N- or 
 C-termini, you can delete these regions from your search model (they are not 
 helping you phase anyway) and re-run Phaser with an appropriately truncated 
 search model.
 Cheers,
 
 ___
 Roger S. Rowlett
 Gordon  Dorothy Kline Professor
 Department of Chemistry
 Colgate University
 13 Oak Drive
 Hamilton, NY 13346
 
 tel: (315)-228-7245
 ofc: (315)-228-7395
 fax: (315)-228-7935
 email: rrowl...@colgate.edu
 
 On 6/17/2014 10:44 AM, Almudena Ponce Salvatierra wrote:
 Dear ccp4 users, 
 
 I get the following message from Phaser when I do a molecular replacement 
 with two ensembles. One of the ensembles is placed but the second one is not 
 placed, and then it says this:
 
 Solutions with Z-scores greater than 13.0 (the threshold indicating a 
 definite
 solution) were rejected for failing packing test
 #1 TFZ=16.8 PAK=487
 #2 TFZ=17.4 PAK=440
 #3 TFZ=17.0 PAK=312
 #4 TFZ=16.7 PAK=294
 #5 TFZ=16.8 PAK=227
 #6 TFZ=16.6 PAK=284
 #7 TFZ=16.2 PAK=219
 #8 TFZ=16.3 PAK=287
 #9 TFZ=16.1 PAK=186
 #10 TFZ=16.6 PAK=277
 #11 TFZ=16.7 PAK=204
 #12 TFZ=16.8 PAK=404
 #13 TFZ=15.9 PAK=271
 #14 TFZ=15.8 PAK=194
 #15 TFZ=14.9 PAK=229
 #16 TFZ=16.4 PAK=368
 #17 TFZ=17.3 PAK=194
 #18 TFZ=15.2 PAK=240
 #19 TFZ=16.1 PAK=325
 #20 TFZ=15.3 PAK=455
 #21 TFZ=16.5 PAK=298
 #22 TFZ=16.6 PAK=290
 #23 TFZ=15.2 PAK=259
 #24 TFZ=16.2 PAK=194
 #25 TFZ=16.1 PAK=314
 #26 TFZ=16.3 PAK=194
 #27 TFZ=16.3 PAK=387
 #28 TFZ=15.6 PAK=193
 #29 TFZ=15.7 PAK=219
 #30 TFZ=15.6 PAK=474
 #31 TFZ=15.1 PAK=194
 #32 TFZ=15.4 PAK=339
 #33 TFZ=15.5 PAK=210
 #34 TFZ=15.2 PAK=300
 #35 TFZ=15.6 PAK=186
 #36 TFZ=16.2 PAK=216
 #37 TFZ=14.9 PAK=182
 #38 TFZ=15.7 PAK=279
 #39 TFZ=15.5 PAK=285
 #40 TFZ=15.0 PAK=374
 #41 TFZ=15.4 PAK=404
 #42 TFZ=15.3 PAK=185
 #43 TFZ=15.6 PAK=227
 #44 TFZ=14.8 PAK=364
 #45 TFZ=15.7 PAK=193
 #46 TFZ=14.5 PAK=448
 #47 TFZ=14.6 PAK=219
 #48 TFZ=15.5 PAK=210
 #49 TFZ=15.7 PAK=189
 #50 TFZ=15.0 PAK=193
 #51 TFZ=15.0 PAK=281
 #52 TFZ=15.2 PAK=202
 
 And the list still continues for a bit. How should I think about this? I 
 would assume that the clashes are too many only by seeing those numbers, but 
 maybe there is something else to take into account?
 
 Thanks a lot in advance.
 
 Best wishes, 
 
 Almudena
 
 
 -- 
 Almudena Ponce-Salvatierra
 Macromolecular crystallography and Nucleic acid chemistry
 Max Planck Institute for Biophysical Chemistry
 Am Fassberg 11 37077 Göttingen
 Germany
 
 

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] CNS MR Help

[Randy Read]

Dear Appu,

If it is possible to solve your structure with currently available MR models 
using any currently existing technology, one of these programs will be able to 
do it.  Throwing yet another program at the problem won’t help, especially an 
older program that was good in its time (and is still, to be fair, cutting edge 
for some things) but is not up-to-date with current molecular replacement 
algorithms.  

Some structures can’t be solved by molecular replacement.  We can now do a 
better job of predicting which ones will be very hard, and this is given in the 
expected LLG output of Phaser.  If that says that the problem is expected to be 
very difficult, you haven’t made any mistakes in running the programs, and 
you’ve tried all the strategies recommended by the authors of those programs, 
then your efforts will be better spent in trying to grow better crystals or 
doing experimental phasing.

Best wishes,

Randy Read

On 12 Jun 2014, at 20:14, Appu kumar appu.kum...@gmail.com wrote:

 Hello All,
I have tried MR in Phaser, MRage, Molrep , Mrbump but i am 
 not getting the true solution which it supposed to be. Although the resoution 
 of data is 6.1A, but i want to give a try to CNS and see if I can find right 
 solution. I have read the manual of CNS but i am unable to get the CNS 
 running.  Would you please help me in running the CNS program.
 
 Thank you 
 Appu

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 44 1223 336500
Wellcome Trust/MRC Building   Fax: + 44 1223 336827
Hills RoadE-mail: rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   www-structmed.cimr.cam.ac.uk

Re: [ccp4bb] [ccp4bb] phaser issue MR

[Randy Read]

Dear Almudena,

We actually have a tutorial that goes through the process of defining a 
solution step by step: 
http://www.phaser.cimr.cam.ac.uk/index.php/MR_using_keyword_input.

The .sol file is a small text file that just contains the description of how a 
defined ensemble should be rotated and translated.  If you look at that file, 
you’ll see that it doesn’t contain the ENSEMBLE definitions, so those still 
have to be provided in the script that refers to the .sol file.

Best wishes,

Randy Read

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk

On 3 Jun 2014, at 09:55, Almudena Ponce Salvatierra maps.fa...@gmail.com 
wrote:

 Dear all, 
 
 thank you for your suggestions, I finally fix the problem as follows. I'll 
 write a summary here for future reference. 
 
 1) Phaser run 1# 
 
 - I provide different ensembles
 - I get a solution in which some of them have been placed
 
 2) Phaser run 2# ERROR
 
 - I provide the ensembles that were not placed in the previous run and REMOVE 
 THE ONES THAT HAD BEEN PLACED
 - I provide the .sol file
 - The program gives the error I sent yesterday
 
 3) Phaser run 3# it works again
 
 Since you said what happened was that Phaser was not finding something and 
 the .sol file contained information about the ensembles that I removed...
 
 - I provided again all the ensembles I gave in run 1#
 - I provide de .sol file (I understand it will fix the solution of the 
 ensembles from run 1#)
 - It runs
 
 So, this was my fix :-) I hear many times the sentence fix a solution, but 
 then whenever I asked someone (before yesterday), each person refers to it in 
 an ambiguous way! some say you need to provide phaser with the pdb from the 
 solution you want to fix, some say that you need to provide the .sol file 
 from what you want to fix and then provide only one new ensemble at the 
 time... maybe it would be useful to have some script/example online of the 
 command line itself, so that one knows what to input exactly. Maybe each 
 one's way of explaining the same thing can lead to misunderstandings at the 
 end.
 
 Best wishes, 
 
 Have a nice day.
 
 Almudena
 
 
 -- Forwarded message --
 From: Almudena Ponce Salvatierra maps.fa...@gmail.com
 Date: 2014-06-03 9:49 GMT+02:00
 Subject: Re: [ccp4bb] phaser issue MR
 To: Airlie McCoy ajm...@cam.ac.uk
 
 
 Dear Airlie, 
 
 don't worry, actually it was you who gave me the clue when you told me what 
 the error was standing for. I will try to be a clear as possible:
 
 1) Phaser run 1# 
 
 - I provide different ensembles
 - I get a solution in which some of them have been placed
 
 2) Phaser run 2# ERROR
 
 - I provide the ensembles that were not placed in the previous run and REMOVE 
 THE ONES THAT HAD BEEN PLACED
 - I provide the .sol file
 - The program gives the error I sent yesterday
 
 3) Phaser run 3# it works again
 
 Since you said what happened was that Phaser was not finding something and 
 the .sol file contained information about the ensembles that I removed...
 
 - I provided again all the ensembles I gave in run 1#
 - I provide de .sol file (I understand it will fix the solution of the 
 ensembles from run 1#)
 - It runs
 
 So, this was my fix :-) I hear many times the sentence fix a solution, but 
 then whenever I asked someone (before yesterday), each person refers to it in 
 an ambiguous way! some say you need to provide phaser with the pdb from the 
 solution you want to fix, some say that you need to provide the .sol file 
 from what you want to fix and then provide only one new ensemble at the 
 time... maybe it would be useful to have some script/example online of the 
 command line itself, so that one knows what to input exactly. Maybe each 
 one's way of explaining the same thing can lead to misunderstandings at the 
 end.
 
 I have another question, if you don't mind. How good does the TFZ score get 
 in one of this iterative Molecular replacement routines? I am now above 5 
 (5.4) which I know is unlikely but I need to take into account that I start 
 with fragments that maybe are not exactly as they are in the structure (I 
 mean in reality) and also that the model is incomplete, meaning that even 
 though I have now most of it places there are still some regions that are not 
 connected, some others with empty electron density... What would you 
 recommend in this case? 
 
 I also have the following warinings:
 
 - Best model and data are insufficient to reach half expected LLG target: 
 placement of one/first ensemble will be very difficult. 
 
 - Search request requires more scattering than defined in composition. 
 Composition increased