Re: [ccp4bb] Experimental phasing Selenomethionine data collection etc. tips

[Randy John Read]

Dear Marco,

You don’t mention here or in the earlier thread whether you tried AlphaFold 
models and, if you did, how you prepared them for MR. I’m happy to hear of any 
case where we still need experimental phasing methods to solve new protein 
structures, but we’ve seen very few examples where AlphaFold models didn’t work!

I’m probably not the only one who would be delighted to take a look at the 
problem to see what we can learn from it, if AlphaFold models really aren’t 
working.

Now, back to your question: to prepare for a SAD phasing experiments one place 
I would look would be Tom Terwilliger’s recent papers on planning and analysing 
SAD experiments (https://doi.org/10.1107/S2059798315019269, 
https://doi.org/10.1107/S2059798315019403) as well as other information on this 
from the Phenix website, including the YouTube tutorials.

Best wishes,

Randy Read

> On 14 May 2024, at 01:17, Marco Bravo 
>  wrote:
>
> Hello all,
> I have a data collection trip next week and plan to collect data on 
> selenomethionine derivative crystals at the al831 beamline. Are there any 
> resources, tips, tutorials, literature etc. That you can recommend to help me 
> prepare for these experiments. Also is there a way to plug in the 
> experimental data into ccp4 cloud to do the automatic structure solution? Do 
> I need native and derivative data to solve the structure? Last trip I 
> collected a seemingly 2.8 angstrom resolution data on a crystal of the native 
> protein but could not get a solution depsite extensive molecular replacement 
> attempts. It seems that assigning a space group for the crystals has been 
> troublesome as well. here is my last thread I posted about the issue for 
> reference.
>
> https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=ind2402=CCP4BB=D=CCE6DFA19FA3D40346=mbrav005%40ucr.edu=112302
>
> 
>
> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Rescale merged data?

[Randy John Read]

I haven’t deposited any PDB entries for a while. The last time I did, I 
remember it not being completely trivial to add these loops. However, I was 
hoping that someone from wwPDB or CCP4 would weigh in with advice on how it can 
be done!

For those who use StarAniso, the current version makes a CIF file with the 
required loops, and they now have advice on their website about this: 
https://staraniso.globalphasing.org/deposition_about.html.

Randy

> On 18 Apr 2024, at 12:25, Frank von Delft  wrote:
>
> Is it easy/non-arcane or indeed automatic for non-experts to add these loops? 
>  Because that's the only way this will be achieved systemically.
>
> Presumably ccp4i2 can be wrangled into making it happen magically. (Apologies 
> if it does already, if so, a comment here would help the discussion...)
>
> Frank
>
>
>
> On 18/04/2024 11:02, Randy John Read wrote:
>> I’d like to add my strong agreement to what Robbie said, but also point out 
>> a wrinkle. When the PDB runs validation, it just takes the data that are in 
>> the first reflection loop of the reflections.cif file. So if you want the 
>> validation statistics to match your reported refinement statistics, that 
>> loop should contain the set of data you gave the refinement program, 
>> especially if you’ve done something like apply an elliptical truncation, 
>> correct for anisotropy, or convert intensities to amplitudes, all of which 
>> change the data in ways that can’t be reversed later. Whenever you’ve done 
>> any of this (and many people are using StarAniso these days, which does all 
>> of those things), please put in a second reflection loop containing the 
>> whole set of intensities to the highest resolution you used, without any 
>> anisotropic scaling or elliptical cutoffs. Then anyone wanting to re-refine 
>> your structure or check your data for artefacts will have more information 
>> available.
>>
>> Of course, I hope we’re moving to a world in which we all also deposit the 
>> intensities before merging, which in principle allows even more quality 
>> control to be done.
>>
>> Best wishes,
>>
>> Randy Read
>>
>>> On 18 Apr 2024, at 05:04, Robbie Joosten  wrote:
>>>
>>> If I may add to that: Please deposit the full dataset, not just the set of 
>>> reflections you end up using. This allows people to use all the data if 
>>> they are interested.
>>>
>>> Cheers,
>>> Robbie
>>>
>>> On 17 Apr 2024 22:35, "Hekstra, Doeke Romke"  
>>> wrote:
>>> Hi Matt,
>>>  I appreciate disagreement and comments from colleagues. My two cents are 
>>> that it seems unnecessary to repeat scaling and merging, or any earlier 
>>> step. If you want to remove structure factor amplitudes or merged 
>>> intensities from the MTZ file you can do so using MTZUTILS or similar 
>>> functionality in CCP4 
>>> (https://www.ccp4.ac.uk/html/mtzutils.html#generalresolution). For 
>>> refinement, you can specify the desired resolution range in your favorite 
>>> refinement program.
>>>  My personal convention is to use CC1/2 = 0.30 as the point to which retain 
>>> data and  = 2 as the nominal resolution of the dataset. If you have 
>>> the HKL2000 scaling log, you should be able to retrieve this information. I 
>>> frankly wish we’d just deposit all data in the PDB rather than truncate 
>>> based on some criterion or another.
>>>  Best, Doeke
>>>  From: Matt Mcleod 
>>> Sent: Wednesday, April 17, 2024 4:12 PM
>>> To: Hekstra, Doeke Romke 
>>> Cc: CCP4BB@JISCMAIL.AC.UK
>>> Subject: Re: [ccp4bb] Rescale merged data?
>>>  Sure thing.
>>>  A former student left somewhere between 30-50 datasets but they scaled the 
>>> data to the detector corners (or maybe edge) in HKL2000.  There are many of 
>>> the high-resolution bins with no reflections in them.  He then went forward 
>>> and merged this data, presumably in HKL2000 again and did his model 
>>> building/refinement.   We now need to re-refine the models against this 
>>> data for publication but we need a more suitable resolution cutoff for the 
>>> data.
>>>  Rather than go back and index/integrate all the data and then rescale the 
>>> data to a more appropriate place (then merge), I was wondering if there was 
>>> a way to take the merged reflections as either .sca or .mtz (from 
>>> scalepacktomtz output) and then rescale to a more appropriate resolution.  
>>> It doesn't seem like the student left unmerged data.
>>>  So, nothing f

Re: [ccp4bb] Rescale merged data?

[Randy John Read]

I’d like to add my strong agreement to what Robbie said, but also point out a 
wrinkle. When the PDB runs validation, it just takes the data that are in the 
first reflection loop of the reflections.cif file. So if you want the 
validation statistics to match your reported refinement statistics, that loop 
should contain the set of data you gave the refinement program, especially if 
you’ve done something like apply an elliptical truncation, correct for 
anisotropy, or convert intensities to amplitudes, all of which change the data 
in ways that can’t be reversed later. Whenever you’ve done any of this (and 
many people are using StarAniso these days, which does all of those things), 
please put in a second reflection loop containing the whole set of intensities 
to the highest resolution you used, without any anisotropic scaling or 
elliptical cutoffs. Then anyone wanting to re-refine your structure or check 
your data for artefacts will have more information available.

Of course, I hope we’re moving to a world in which we all also deposit the 
intensities before merging, which in principle allows even more quality control 
to be done.

Best wishes,

Randy Read

> On 18 Apr 2024, at 05:04, Robbie Joosten  wrote:
> 
> If I may add to that: Please deposit the full dataset, not just the set of 
> reflections you end up using. This allows people to use all the data if they 
> are interested.
> 
> Cheers,
> Robbie
> 
> On 17 Apr 2024 22:35, "Hekstra, Doeke Romke"  
> wrote:
> Hi Matt,
>  
> I appreciate disagreement and comments from colleagues. My two cents are that 
> it seems unnecessary to repeat scaling and merging, or any earlier step. If 
> you want to remove structure factor amplitudes or merged intensities from the 
> MTZ file you can do so using MTZUTILS or similar functionality in CCP4 
> (https://www.ccp4.ac.uk/html/mtzutils.html#generalresolution). For 
> refinement, you can specify the desired resolution range in your favorite 
> refinement program.
>  
> My personal convention is to use CC1/2 = 0.30 as the point to which retain 
> data and  = 2 as the nominal resolution of the dataset. If you have 
> the HKL2000 scaling log, you should be able to retrieve this information. I 
> frankly wish we’d just deposit all data in the PDB rather than truncate based 
> on some criterion or another. 
>  
> Best, Doeke
>  
> From: Matt Mcleod  
> Sent: Wednesday, April 17, 2024 4:12 PM
> To: Hekstra, Doeke Romke 
> Cc: CCP4BB@JISCMAIL.AC.UK
> Subject: Re: [ccp4bb] Rescale merged data?
>  
> Sure thing.
>  
> A former student left somewhere between 30-50 datasets but they scaled the 
> data to the detector corners (or maybe edge) in HKL2000.  There are many of 
> the high-resolution bins with no reflections in them.  He then went forward 
> and merged this data, presumably in HKL2000 again and did his model 
> building/refinement.   We now need to re-refine the models against this data 
> for publication but we need a more suitable resolution cutoff for the data. 
>  
> Rather than go back and index/integrate all the data and then rescale the 
> data to a more appropriate place (then merge), I was wondering if there was a 
> way to take the merged reflections as either .sca or .mtz (from 
> scalepacktomtz output) and then rescale to a more appropriate resolution.  It 
> doesn't seem like the student left unmerged data.  
>  
> So, nothing fancy (aniostropy etc), there is just a lot of data that needs to 
> be adjusted and I am trying to avoid reprocessing all the frames again.
>  
> Matt
>  
> On Wed, 17 Apr 2024 at 15:59, Hekstra, Doeke Romke 
>  wrote:
> Hi Matt,
> 
> It would be helpful if you could describe your case in more detail. Do you 
> want to change the resolution cutoff after scaling? Do you want to keep more 
> data? Fewer? Or do you mean something different such as truncation to 
> generate amplitudes, application of anisotropic resolution cutoffs,  or 
> outlier rejection? Are you referring to data that were scaled in HKL2000?
> 
> Best, Doeke
> 
> -Original Message-
> From: CCP4 bulletin board  On Behalf Of Matt McLeod
> Sent: Wednesday, April 17, 2024 3:04 PM
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: [ccp4bb] Rescale merged data?
> 
> Hi all,
> 
> I am looking at a old students data and it looks like they didn't properly 
> cut off the data during scaling.  All of the files I have appear to be the 
> merged .sca (or mtz after converting with scalepacktomtz) - is there a way to 
> retruncate the data after merging or do I have to reprocess the data?
> 
> Thanks,
> 
> 
> 
> 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] i2 won't accept 90 degree angles in moniclinic

[Randy John Read]

Hi Huw,

When I was at the workshop, I have to admit that after ccp4i2 refused to do it 
I switched to Phenix and used phenix.reflection_file_converter! It would be 
better if i2’s reindex or change spacegroup task didn’t refuse, and I guess 
this might improve after the overly-zealous consistency checking is relaxed, or 
cad could be used following your suggestion.

As for how to report this in a publication, we actually ran into this when 
working on a Hyp-1 complex structure with Mariusz Jaskolski and Zbyszek Dauter, 
where the crystals were apparently perfectly tetartohedrally twinned. The 
apparent space group was some variant of P422 (which is how the data were 
processed and merged) but the true space group was identified as one choice of 
C2. Merging in the correct space group gave only 73% completeness, but the 
statistics for the data processed in higher symmetry were very similar so the 
expanded data set was used and deposited 
(https://doi.org/10.1107/s1399004713030319). Statistics for processing in both 
space groups were presented in Table 1.

Best wishes,

Randy

> On 23 Feb 2024, at 13:45, Huw Jenkins  wrote:
> 
> Hi Randy,
> 
>> On 23 Feb 2024, at 11:49, Randy John Read  wrote:
>> 
>> Why would we want to impose an arbitrary restriction on users for this 
>> relatively common scenario.
> 
> If the user has the unmerged data this can be imported into CCP4i2 via the 
> data reduction task and merged in P1. How would you expand the merged data to 
> P1 - using cad from a script? Perhaps this should be made possible in CCP4i2 
> after the merged data were imported. i2 already has a "Reindex or change 
> spacegroup" task (using POINTLESS)  but it won't do this:
> 
> "FATAL ERROR:
> Specified SPACEGROUP P1 must belong to same crystal system and point group
> as the input space group P 41 21 2"
> 
>> 
>> Note that this kind of confusion between twinning and true symmetry will 
>> mostly arise when the twin fractions are close to equal. Then: a) the 
>> twin-related intensities should really be measurements of the same thing, 
>> and you get more precise data by making them equal; b) the superimposed 
>> diffraction patterns will obey the higher symmetry, although the spots might 
>> start to split at higher resolution.
>> 
>> I would also argue that, if the twin fractions are experimentally 
>> indistinguishable from being equal, processing in higher symmetry and 
>> expanding the data to the correct lower symmetry is the correct approach to 
>> take for your final data set. 
> 
> That's a fair comment. In that case would you then report the merging 
> statistics for the higher symmetry data in "Table 1" and note that the merged 
> data were subsequently expanded to the correct lower symmetry?
> 
> 
> Huw


-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] i2 won't accept 90 degree angles in moniclinic

[Randy John Read]

In the recent crystallography workshop in Thailand, we had at least 3 cases 
brought by participants where we had reason to suspect that the data had been 
over-merged because of undetected twinning. Often there is more than one 
possible twinning operator and (as potentially in one of the cases brought up 
recently on this BB) the easiest way to sort out the true symmetry is just to 
expand the data to P1 and solve the structure by looking for the appropriate 
expanded number of copies. If the data are of reasonable quality 
(case-dependent) and the model is good (true more often than not, in these 
post-AlphaFold days), the multicopy MR solution can be pretty straightforward. 
Once you have a better idea of the true symmetry, then would be a good time to 
try reprocessing the data without assuming too high symmetry. Why would we want 
to impose an arbitrary restriction on users for this relatively common scenario.

Note that this kind of confusion between twinning and true symmetry will mostly 
arise when the twin fractions are close to equal. Then: a) the twin-related 
intensities should really be measurements of the same thing, and you get more 
precise data by making them equal; b) the superimposed diffraction patterns 
will obey the higher symmetry, although the spots might start to split at 
higher resolution.

I would also argue that, if the twin fractions are experimentally 
indistinguishable from being equal, processing in higher symmetry and expanding 
the data to the correct lower symmetry is the correct approach to take for your 
final data set. Obviously you still want to try to process the data in the 
correct space group to make this decision, looking at things like whether the 
observations related by true symmetry are more highly correlated than 
observations related by the twin operator(s).

Best wishes,

Randy

> On 23 Feb 2024, at 10:39, Huw Jenkins 
> <288da93ae744-dmarc-requ...@jiscmail.ac.uk> wrote:
>
>> On 23 Feb 2024, at 09:58, Winter, Graeme (DLSLtd,RAL,LSCI) 
>> <6a19cead4548-dmarc-requ...@jiscmail.ac.uk> wrote:
>>
>> so strictly it is possible and correct - if experimentally unlikely - to 
>> have the situation we are discussing here occur.
>
> I believe this is only technically possible because the MTZ format does not 
> store esds on the unit cell parameters? In the thaumatin example processing 
> the dataset here  - https://zenodo.org/records/4916649 - assuming monoclinic 
> symmetry results in unit cell:
>
> 58.176(2), 150.543(4), 58.2050(18), 90.0, 90.1058(9), 90.0
>
> The situation you describe would result in for example:
>
> 58.1087(18), 150.543(4), 58.1087(15), 90.0, 90.(1), 90.0
>
> and the test should really only fail for:
>
> 58.1087(18), 58.1087(18), 150.400(5), 90.0, 90.0, 90.0
>
> i.e. where a=b have same value and esd (as they were constrained to be 
> identical and esd on beta is 0 as it was constrained to be 90.
>
>
>
>> Telling users to “fiddle the parameters” so that the strict test is 
>> satisfied feels like a non-ideal answer: a warning when importing such data 
>> could be legitimate e.g. “hmm I note a = b and al=be=ga=90 _exactly_ this is 
>> unusual, I hope you know what you are doing” rather than a flat out error
>
> I think you have misunderstood here. I was suggesting telling users to 
> integrate/scale the data without imposing higher symmetry was the correct 
> thing to do? I don't see how "fiddling of parameters" is required.
>
> But I agree i2 should allow an override of this test.
>
>
> Huw
> 
>
> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Difficult Molecular replacement

[Randy John Read]

Hi,

This is also one of my favourite methods, although I hadn’t realised that you 
could just say “labin F=FC” instead of making fake intensities!

However, at a recent workshop I ran into a case where pointless misled me. This 
was a case with tNCS where one of the translation components was half a unit 
cell edge. As a result, the calculated diffraction data would have alternately 
strong and weak reflections along that axis regardless of whether there was a 
crystallographic 2(1) axis or an NCS translation of 1/2 parallel to that axis. 
Pointless assigned a 2(1) axis where there should have been a pure two-fold. On 
the other hand, Zanuda tried refinement in all the different possibilities and 
only one refined well.

Best wishes,

Randy

> On 21 Feb 2024, at 16:05, Kay Diederichs  
> wrote:
>
> Hi Vaheh,
>
> for this purpose, I use
>
> pointless hklin refmacXY.mtz < labin F=FC
> eof
>
> Thus, pointless determines the space group, including the crystallographic 
> screw axes, from the Fcalc.
>
> Best wishes,
> Kay
>
> On Wed, 21 Feb 2024 15:40:07 +, Oganesyan, Vaheh 
>  wrote:
> ...
>> This might be a silly question, but I do not know the answer: After 
>> refinement in P1 how do I distinguish which axis is crystallographic and 
>> which one in non-crystallographic?
>>
>> Vaheh
>
> 
>
> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Difficult Molecular replacement

[Randy John Read]

Hi,

It’s possible the true space group is P2(1) with the b-axis unique, and that 
subset of true symmetry is found repeatedly with different incorrect 
backgrounds of other copies. But I think the easiest way to resolve this 
unambiguously is to solve in P1, and let that uncover the true symmetry.

Best wishes,

Randy

> On 21 Feb 2024, at 13:56, Pedro Matias  wrote:
> 
> But curiously, all the 4 best solutions correspond to a SG with a 21 screw 
> along b.
> And amazingly none of the TF solutions is rejected due to clashes.
> On 21/02/2024 12:20, Eleanor Dodson wrote:
>> Lots of comments, but it would be easier to actually look at your integrated 
>> data!  
>> Some of the stats look a bit ropey - 
>> 621 reflections labelled as outliers by PHASER?
>> Very anisotropic 
>> Moments go mad at the highest resolution..
>> 
>> The good news - extremely strong signal from the rotation function means the 
>> model is probably a good one.
>> Bad news - translation function results do not select a definitive solution..
>> Possible reasons 
>> Unit Cell:   72.61   73.73  147.23   90.00   90.00   90.00
>> Most likely data problems - a axis ~ = b axis so twinning is possible 
>> 
>> I could add more comments if either you could share the unmerged data, or at 
>> least a pointless logfile..
>> Cheers Eleanor
>> 
>> 
>> 
>> 
>> 
>> On Wed, 21 Feb 2024 at 11:06, Randy John Read  wrote:
>> Hi Marco,
>> 
>> To add to what Kay has said:
>> 
>> The intensity moments from Phaser (between 1.5 and 2 for the second moments 
>> after correcting for anisotropy) are indicative of likely twinning. With the 
>> cell dimensions, it might be possible to have pseudomerohedral twinning in 
>> an orthorhombic space group, but given the lack of distinction among 
>> possible choices of orthorhombic spac group (noted by Kay), it seems much 
>> more likely that the true symmetry is lower and that you have pseudosymmetry 
>> combined with perfect twinning.
>> 
>> Judging from the strong and unambiguous rotation peak, your model is clearly 
>> very good, so I think it would be easy to ask Phaser to solve this by 
>> looking for 4 copies in space group P1. You can get P1 data either by 
>> expanding the orthorhombic data to P1 or by re-merging the data in P1. If 
>> the merging statistics were good, that would indicate that any twinning 
>> would be close to perfect, so just expanding the data would be a reasonable 
>> choice. Alternatively, you have reasonable redundancy so merging in P1 would 
>> be a plausible choice. I would probably go with expanding the data, figuring 
>> out from the MR solution what the real symmetry is, and then merging the 
>> data with that symmetry.
>> 
>> Unless there are some other pathologies, I think the MR in P1 is very likely 
>> to give a clear answer (or maybe 2 answers related by the twin operator). 
>> It’s formally possible that you could have a different number of copies 
>> (e.g. 6 in the unit cell) if the true symmetry were monoclinic, so keep an 
>> open mind on that question. You could just try finding the largest domain 
>> from splitting the AlphaFold model (presumably the domain for which you sent 
>> the log file), work out the symmetry from that solution, and then run a job 
>> to search for all the domains in the right space group. It’s generally a 
>> good idea, by the way, to ask Phaser to look for everything you expect to 
>> find in one job, because it has built-in logic to predict the best search 
>> order but then update it on the basis of preliminary results.
>> 
>> There are different ways to sort out the true symmetry from the MR solution, 
>> but within CCP4 the Zanuda procedure is a very effective choice.
>> 
>> Get in touch if you have any difficulties following this procedure, and it 
>> would be great to let the BB know the outcome!
>> 
>> Best wishes,
>> 
>> Randy Read
>> 
>> > On 21 Feb 2024, at 08:42, Kay Diederichs  
>> > wrote:
>> >
>> > Hi Marco:
>> >
>> > short comments (I sent you also a private mail):
>> > - the stats in CORRECT.LP look ok, but I'd like to know what ISa is, and 
>> > what the number of outliers is ("misfits"). Seeing the delta-CC1/2 stats 
>> > as a function of frame number is also useful for judging e.g. the 
>> > radiation damage - this is available from XDSGUI in the "statistics" tab. 
>> > Another remark: SPOT_RANGE=1 11 in COLSPOT will sample reciprocal space 
>> > poorly - I always use the

Re: [ccp4bb] Difficult Molecular replacement

[Randy John Read]

Hi Eleanor,

> On 21 Feb 2024, at 12:20, Eleanor Dodson  wrote:
> 
> Lots of comments, but it would be easier to actually look at your integrated 
> data! 
> Some of the stats look a bit ropey - 
> 621 reflections labelled as outliers by PHASER?
> Very anisotropic 
> Moments go mad at the highest resolution..

These 3 observations are actually all consequences of the strong anisotropy. 
Phaser flags reflections with very little information content as outliers, 
because they’re being ignored (as they would have almost no effect on the 
likelihood calculation), and these come from the data in the weak directions. 
The moments go mad, but in a way that is predictable from the sigmas on the 
intensities, as seen by the fact that the predicted and observed moments track 
each other as a function of resolution.

Best wishes,

Randy

> 
> The good news - extremely strong signal from the rotation function means the 
> model is probably a good one.
> Bad news - translation function results do not select a definitive solution..
> Possible reasons 
> Unit Cell:   72.61   73.73  147.23   90.00   90.00   90.00
> Most likely data problems - a axis ~ = b axis so twinning is possible 
> 
> I could add more comments if either you could share the unmerged data, or at 
> least a pointless logfile..
> Cheers Eleanor
> 
> 
> 
> 
> 
> On Wed, 21 Feb 2024 at 11:06, Randy John Read  wrote:
> Hi Marco,
> 
> To add to what Kay has said:
> 
> The intensity moments from Phaser (between 1.5 and 2 for the second moments 
> after correcting for anisotropy) are indicative of likely twinning. With the 
> cell dimensions, it might be possible to have pseudomerohedral twinning in an 
> orthorhombic space group, but given the lack of distinction among possible 
> choices of orthorhombic spac group (noted by Kay), it seems much more likely 
> that the true symmetry is lower and that you have pseudosymmetry combined 
> with perfect twinning.
> 
> Judging from the strong and unambiguous rotation peak, your model is clearly 
> very good, so I think it would be easy to ask Phaser to solve this by looking 
> for 4 copies in space group P1. You can get P1 data either by expanding the 
> orthorhombic data to P1 or by re-merging the data in P1. If the merging 
> statistics were good, that would indicate that any twinning would be close to 
> perfect, so just expanding the data would be a reasonable choice. 
> Alternatively, you have reasonable redundancy so merging in P1 would be a 
> plausible choice. I would probably go with expanding the data, figuring out 
> from the MR solution what the real symmetry is, and then merging the data 
> with that symmetry.
> 
> Unless there are some other pathologies, I think the MR in P1 is very likely 
> to give a clear answer (or maybe 2 answers related by the twin operator). 
> It’s formally possible that you could have a different number of copies (e.g. 
> 6 in the unit cell) if the true symmetry were monoclinic, so keep an open 
> mind on that question. You could just try finding the largest domain from 
> splitting the AlphaFold model (presumably the domain for which you sent the 
> log file), work out the symmetry from that solution, and then run a job to 
> search for all the domains in the right space group. It’s generally a good 
> idea, by the way, to ask Phaser to look for everything you expect to find in 
> one job, because it has built-in logic to predict the best search order but 
> then update it on the basis of preliminary results.
> 
> There are different ways to sort out the true symmetry from the MR solution, 
> but within CCP4 the Zanuda procedure is a very effective choice.
> 
> Get in touch if you have any difficulties following this procedure, and it 
> would be great to let the BB know the outcome!
> 
> Best wishes,
> 
> Randy Read
> 
> > On 21 Feb 2024, at 08:42, Kay Diederichs  
> > wrote:
> >
> > Hi Marco:
> >
> > short comments (I sent you also a private mail):
> > - the stats in CORRECT.LP look ok, but I'd like to know what ISa is, and 
> > what the number of outliers is ("misfits"). Seeing the delta-CC1/2 stats as 
> > a function of frame number is also useful for judging e.g. the radiation 
> > damage - this is available from XDSGUI in the "statistics" tab. Another 
> > remark: SPOT_RANGE=1 11 in COLSPOT will sample reciprocal space poorly - I 
> > always use the first half of the DATA_RANGE as SPOT_RANGE, at a negligible 
> > cost in CPU time.
> > - the Phaser logfile shows that the rotation function (table at line 1344) 
> > has only a single solution, but the translation function (table at line 
> > 7028)  does not even allow to determine the space gr

Re: [ccp4bb] Difficult Molecular replacement

[Randy John Read]

Hi Marco,

To add to what Kay has said:

The intensity moments from Phaser (between 1.5 and 2 for the second moments 
after correcting for anisotropy) are indicative of likely twinning. With the 
cell dimensions, it might be possible to have pseudomerohedral twinning in an 
orthorhombic space group, but given the lack of distinction among possible 
choices of orthorhombic spac group (noted by Kay), it seems much more likely 
that the true symmetry is lower and that you have pseudosymmetry combined with 
perfect twinning.

Judging from the strong and unambiguous rotation peak, your model is clearly 
very good, so I think it would be easy to ask Phaser to solve this by looking 
for 4 copies in space group P1. You can get P1 data either by expanding the 
orthorhombic data to P1 or by re-merging the data in P1. If the merging 
statistics were good, that would indicate that any twinning would be close to 
perfect, so just expanding the data would be a reasonable choice. 
Alternatively, you have reasonable redundancy so merging in P1 would be a 
plausible choice. I would probably go with expanding the data, figuring out 
from the MR solution what the real symmetry is, and then merging the data with 
that symmetry.

Unless there are some other pathologies, I think the MR in P1 is very likely to 
give a clear answer (or maybe 2 answers related by the twin operator). It’s 
formally possible that you could have a different number of copies (e.g. 6 in 
the unit cell) if the true symmetry were monoclinic, so keep an open mind on 
that question. You could just try finding the largest domain from splitting the 
AlphaFold model (presumably the domain for which you sent the log file), work 
out the symmetry from that solution, and then run a job to search for all the 
domains in the right space group. It’s generally a good idea, by the way, to 
ask Phaser to look for everything you expect to find in one job, because it has 
built-in logic to predict the best search order but then update it on the basis 
of preliminary results.

There are different ways to sort out the true symmetry from the MR solution, 
but within CCP4 the Zanuda procedure is a very effective choice.

Get in touch if you have any difficulties following this procedure, and it 
would be great to let the BB know the outcome!

Best wishes,

Randy Read

> On 21 Feb 2024, at 08:42, Kay Diederichs  
> wrote:
>
> Hi Marco:
>
> short comments (I sent you also a private mail):
> - the stats in CORRECT.LP look ok, but I'd like to know what ISa is, and what 
> the number of outliers is ("misfits"). Seeing the delta-CC1/2 stats as a 
> function of frame number is also useful for judging e.g. the radiation damage 
> - this is available from XDSGUI in the "statistics" tab. Another remark: 
> SPOT_RANGE=1 11 in COLSPOT will sample reciprocal space poorly - I always use 
> the first half of the DATA_RANGE as SPOT_RANGE, at a negligible cost in CPU 
> time.
> - the Phaser logfile shows that the rotation function (table at line 1344) 
> has only a single solution, but the translation function (table at line 7028) 
>  does not even allow to determine the space group - there is very little 
> contrast difference between potential "solutions".
>
> Best wishes,
> Kay
>
> On Tue, 20 Feb 2024 21:45:12 +, Marco Bravo  wrote:
>
>> https://www.dropbox.com/scl/fo/pmw9nisdmq53yir2jqcmu/h?rlkey=zaj6cnknkjgkowrzf0vrmqdyf=0
>>
>> Here is a link to my Molecular replacement and asymmetric unit contents 
>> logfiles.
>>
>> I ran Simple molecular replacement phaser MR through ccp4 cloud with my .mtz 
>> that was auto-processed at the ALS light source Beamline 831. I truncated my 
>> Alphaphold model into several domains as suggested and ran separate MR jobs. 
>> All of them are still running but one MR job for just one of the helicase 
>> domains finished and that is the molecular replacement job logfile i have 
>> posted.
>>
>> I also posted my Data collection output for the crystal in the dropbox file.
>>
>> I also attached my XDS.INP and XDSCONV.INP files for more troubleshooting 
>> help.
>>
>> Thank you all for helping me I hope this provides more information to help 
>> figure out what is going on. I can post more information if needed.
>>
>> 
>>
>> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
>> mailing list hosted by http://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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by 

Re: [ccp4bb] Difficult Molecular replacement

[Randy John Read]

Hi Marco,

I was just about to write with a different suggestion when Kay’s reply came in. 
What Kay suggests sounds like the best thing to do first!

However, if that isn’t the issue, then I’m wondering if it could be a problem 
with incorrectly diagnosed translational non-crystallographic symmetry (tNCS). 
The ambiguity about which axes are pure 2-folds or 2-fold screws is the kind of 
thing that occurs when there is an NCS translation with a component of 1/2 
along at least one of the axes.

I couldn’t find anything about how many copies you have in the asymmetric unit 
of the crystal, but is it an even number? Do any programs (including Phaser) 
find large native Patterson peaks away from the origin? Are the intensity 
distributions and/or moments unusual?

You could send me a copy of the log file from a Phaser run if you don’t know 
where to find this information.

Good luck!

Randy Read

> On 20 Feb 2024, at 09:11, Kay Diederichs  
> wrote:
>
> Hello Marco,
>
> this sounds to me like there is space group confusion in the sense that if 
> you use the output model from Phaser, it may say space group P22121 (with 
> a a MTZ file specifying space group P21212 (with a
> If this is indeed your problem, then the way to resolve this is
> a) in XDS.INP, specify UNIT_CELL_CONSTANTS= b c a 90 90 90 !(where a SPACE_GROUP_NUMBER=18,  JOB=CORRECT . Run XDS.
> b) remove UNIT_CELL_CONSTANTS and SPACEGROUP_NUMBER lines from XDSCONV.INP, 
> and run XDSCONV. Or use pointless/aimless
> c) run Phaser again with the resulting MTZ file (check that it has spacegroup 
> 18 i.e. P21212 and cell parameters with c shortest). Check that the resulting 
> PDB file has P21212 and b c a cell parameters in the header.
> d) run refinement
>
> Hope this helps,
> Kay
>
> On Tue, 20 Feb 2024 01:15:49 +, Marco Bravo  wrote:
>
>> Hello Todd, I get the best solution for p22121 space group after MR with an 
>> LLG score of 640 from phaser. and the Rfree is .4748. However after MR 
>> refinement does not lower the Rfree and it appears to make the Rfree worse. 
>> The XDS software indicates that my best solution is P21 21 2. Often Phaser 
>> MR places the solution in P 21 21 2. The helicase is a superfamily 2 
>> helicase and is only monomeric. Its a 543aa long protein with a MW of 62Kda. 
>> It should have two RecA like domains at the core but the protein I have 
>> crystallized has a previously uncharacterized n-terminal domain responsible 
>> for tight single stranded DNA binding.  I have tried different space groups 
>> manually but that resulted in clashing. I will be frank I do need to work on 
>> my crystallography background as crystal lattices, space group, and 
>> self-rotaion functions are limited. Thank you so far for your help , I will 
>> try further trimming down my search model into separate domain and trying 
>> that in the meantime.
>>
>> 
>>
>> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
>> mailing list hosted by http://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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Phaser 2.8.3: Hendrickson-Lattman coefficients generated from dataset lacking anomalous signal

[Randy John Read]

Hi,

Hendrickson-Lattman coefficients are just a way of storing phase probability 
information, and they can come from different sources including atomic models. 
Phaser puts in HL coefficients because they could be handy under some 
circumstances for combining the phase information from experimental phasing. 
You might notice that only A and B are non-zero for the molecular replacement 
HL coefficients. That’s because the phase probability distribution is unimodal 
for calculated phases, whereas it’s generally bimodal for experimental phases 
(thus requiring more coefficients).

Best wishes,

Randy Read

> On 8 Feb 2024, at 14:32, Nitin Kulhar 
> <9dfccc771c91-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Dear all
> 
> Is anomalous diffraction necessary for determining experimental phases and 
> the Hendrickson-Lattman coefficients (HLA, HLB, HLC, and HLD)?
> 
> MR solution from Phaser 2.8.3 (interfaced in ccp4 8.0.000 suite) seems to be 
> generating HLA/B/C/D coefficients from an x-ray diffraction dataset.
> 
> Wavelength: 
> 1.54179 Angstrom. 
> 
> Sample: 
> • 
> Protein-small molecule ligand complex crystal. 
> • No anomalous scatterer in the protein, the ligand, or the 
> crystallization condition.
> 
> Data reduction:
> Xtriage and aimless analyses did not indicate significant anomalous signal. 
> 
> I would appreciate any help in understanding the reasons for these 
> observations.
> 
> Thanks and regards
> Nitin Kulhar
> PhD student
> c/o Dr Rajakumara Eerappa
> Macromolecular Structural Biology Lab
> Department of Biotechnology
> Indian Institute of Technology Hyderabad
> Kandi, Sangareddy
> Telangana, India - 502284
> 
> Disclaimer:- This footer text is to convey that this email is sent by one of 
> the users of IITH. So, do not mark it as SPAM. 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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
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] Goniometer head free to a good home!

[Randy John Read]

Hi,

I’m happy to report that this goniometer head has now found a good new home!

Best wishes,

Randy

> On 27 Jan 2024, at 19:23, Randy John Read  wrote:
> 
> Hi,
> 
> Now that we don’t have a home lab source any more, there’s no need for the 
> goniometer head that we used to mount crystals on it. I’d be happy to give it 
> to someone who still has a home source and could make good use of it! 
> Preference would be given to people in the UK, mostly because that would 
> eliminate any potential import/export complications with sending it!
> 
> Best wishes,
> 
> Randy Read
> 
> <6D68-3347-4E88-B296-9BD350C14329_1_105_c.jpeg> 
> 
> ——
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical Research Tel: +44 1223 336500
> The Keith Peters Building
> 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

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Data reprocessing

[Randy John Read]

Hi,

The current policy of the PDB is that you can deposit a re-refinement of a 
structure deposited by someone else only if you publish a new paper about it, 
so that there is a citation associated with the structure. Of course, echoing 
Eleanor, if you find an issue with someone else’s structure the most collegiate 
thing to do is to inform the authors and attempt to work with them to fix up 
the problem.

Best wishes,

Randy

> On 27 Jan 2024, at 07:00, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Did you make the original deposition ? If so the pdb accepts re-depositions...
> 
> If it was done by someone else I guess the courteous thing is to notify them 
> and maybe talk to the pdb- redo team? 
> Cheers Eleanor 
> 
> On Sat, 27 Jan 2024 at 01:53, Lucas Souza  
> wrote:
> Dear all,
> 
> After auditing and reprocessing a deposited structure with clear processing 
> mistakes (missing/wrong residues and ligands with evident density) prior to 
> some analysis that are going to be published, what should be done? Re-deposit 
> the structure with a "reprocess" flag? Reopen the PDB deposition and update 
> the files? Or simply state in the publication that the data was reprocessed?
> 
> I'm curious if there's a consensus on handling situations of this nature.
> 
> Cheers,
> Lucas
> 
> 
> 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

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Goniometer head free to a good home!

[Randy John Read]

Hi,

Now that we don’t have a home lab source any more, there’s no need for the 
goniometer head that we used to mount crystals on it. I’d be happy to give it 
to someone who still has a home source and could make good use of it! 
Preference would be given to people in the UK, mostly because that would 
eliminate any potential import/export complications with sending it!

Best wishes,

Randy Read

[cid:266bd826-d9c3-418d-8c22-508c5960f976@GBRP265.PROD.OUTLOOK.COM]


——
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Resolution Discrepancy in Data Set

[Randy John Read]

Hi,

I think the point about an R-factor of 42% is a bit more subtle than it comes 
across in Martin’s reply. For random data without measurement errors (i.e. 
coming from a Wilson distribution of intensities), the expected R-factor for 
acentric data is something like 59%. The 42% in Evans & Murshudov comes about 
when the intensities have such large errors that the measurements don’t 
contribute any information to the amplitudes, so after running the 
French/Wilson algorithm in truncate, they all have the same value, i.e. the 
expected value of an amplitude before making a measurement. So 42% is the 
R-factor expected for data that are all equal to the average of the structure 
factor amplitudes they’re being compared to.

In this data set, with I/sig(I) dropping to about 0.1 in the high resolution 
shells, an R of 42% or so would be expected, because all the “measured” 
amplitudes (produced by truncate) will have almost the same value.

Best wishes,

Randy Read

> On 23 Jan 2024, at 14:49, Martin Malý  wrote:
>
> Dear all,
>
> I am sorry for a late reply. R-values should not exceed 0.42 which happened 
> in your case for shells 1.91-1.83 and 1.83-1.77. It is because theoretically 
> (under some assumptions), a perfect model gives an R value of 0.42 against 
> random data (Evans & Murshudov 2013 https://doi.org/10.1107/S090744491361 
> ). So as David wrote, your data are not better than 1.9 A. I would refine the 
> structure against data at resolution of 2.2 A (including modelling of water 
> molecules), then run paired refinement to add resolution shells 2.2-2.1, 
> 2.1-2.0 and 2.0-1.9A and choose an optimal high-resolution cutoff according 
> results. Feel free to ask.
>
> "the actual reflections used in calculations and reported statistics may be 
> different (in phenix.refine)"
> Thank for this comment, this was new for me, good to know! I can imagine it 
> can cause misleading situations/interpretations...
>
> Nowadays, I quite often hear an opinion "You do not have to cut your data at 
> high-resolution, refinement program will put low weights for noisy 
> high-resolution reflections so they will not harm your model." I just would 
> like to ask Pavel and other refinement experts - would you dis/agree? Such an 
> approach should be possible by principle but R-values etc. look quite ugly 
> then.
>
> Best wishes,
> Martin
>
> On Wed, 2024-01-17 at 18:15 -0800, Pavel Afonine wrote:
>>
>>
>> Hi Liliana,
>>
>> a few things to consider:
>>
>> 0) There is a Phenix mailing list for Phenix specific questions (phenixbb);
>>
>> 1) Bin completeness depends on (obviously) how binning is done (number of 
>> reflections per bin or number of bins or binning in d^2 or d^3 spacing or 
>> log-binning etc etc etc) -- all of this will affect the number of 
>> reflections in the "highest resolution bin" and corresponding statistics. 
>> And..."how binning is done" wildly differs by the program used or even 
>> within the same program!
>>
>> 2) Refinement (in Phenix) as well as many other Phenix tools apply temporary 
>> reflection omission according to the Read (1999) paper (Acta Cryst. (1999). 
>> D55, 1759-1764). This means while you have so many reflections in your input 
>> reflection file, the actual reflections used in calculations and reported 
>> statistics may be different. Most logs files keep a good record of this, 
>> meaning you can track this down by inspecting logs files carefully.
>>
>> Let me know if you need more assistance with this issue.
>>
>> Cheers,
>> Pavel
>>
>> On Wed, Jan 17, 2024 at 1:43 PM David Briggs  
>> wrote:
>>> Hi Liliana,
>>>
>>> Two things leap out at me when I look at your data summary.
>>>
>>> (1) Your data probably do not go to 1.77Å. The CC1/2 in your outer shell is 
>>> below any of the usual thresholds. There are discussions to be had about 
>>> what the threshold is, but normally CC1/2 values of 0.5 or sometimes 0.3 
>>> are used. You should also consider I/sigI.
>>>
>>> (2) I believe that by default Phenix.refine excludes weaker reflections 
>>> from refinement, which leads to the discrepancy in completeness statistics. 
>>> As your data do not extend as far as what is contained in your mtz file, 
>>> Phenix excludes those essentially "empty" reflections. Judging by the 
>>> Phenix refine output, I would estimate your data goes to somewhere around 
>>> 1.9Å
>>>
>>> The program Pairef can help inform your choice of high-resolution cutoff.
>>>
>>> This can be run from CCP4cloud, but is also available for Phenix, I believe.
>>>
>>> See https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8248825/
>>>
>>> Hope this helps,
>>>
>>> Dave
>>>
>>>
>>> Dr David C. Briggs CSci MRSB
>>> Principal Laboratory Research Scientist
>>> Signalling and Structural Biology Lab
>>> The Francis Crick Institute
>>> London, UK
>>> ==
>>> about.me/david_briggsFrom: CCP4 bulletin board  on 
>>> behalf of Liliana Margent 
>>> Sent: Wednesday, January 17, 2024 9:19:36 PM
>>> To: CCP4BB@JISCMAIL.AC.UK 
>>> 

Re: [ccp4bb] Phenix version 1.21 released

[Randy John Read]

For anyone who is confused:

Unfortunately, the link given here (https://phenix-online.org/download/) says 
that the latest release is 1.19.2-4158! That’s very odd because the same URL 
without the trailing slash (https://phenix-online.org/download) gets to a 
different page with the right information! Must be something relatively obscure 
about how URLs are interpreted. I’m using Safari in case other people are 
having a different experience...

Best wishes,

Randy Read

> On 17 Jan 2024, at 19:12, Paul Adams  wrote:
>
> The Phenix developers are pleased to announce that version 1.21 of Phenix is 
> now available (build 1.21-5207). Binary installers for Linux, Mac OSX, and 
> Windows (under Windows Subsystem for Linux), and the source installer, are 
> available at the download site:
>
> http://phenix-online.org/download/
>
> Highlights for the 1.21 version of Phenix include:
>
> - Likelihood-based docking of models into cryo-EM maps.
>• Full support for structure determination with AlphaFold models.
>• New joint X-ray and neutron refinement algorithm.
>• Quantum Mechanical Restraints (QMR) to calculate ligand restraints in 
> situ.
>
> Detailed changes are:
>
> - Full support for structure determination with AlphaFold models in Phenix GUI
>  - PredictAndBuild X-ray and Cryo-EM structure solution from data and 
> sequences
>  - Phenix AlphaFold server
>  - Video tutorials for prediction, X-ray structure solution and Cryo-EM map 
> interpretation
> - Cryo-EM tools support ChimeraX visualization
> - Cryo-EM density modification and anisotropic scaling display local 
> resolution
> - Tutorials available for automated structure determination with 
> PredictAndBuild
> - New em_placement and emplace_local tools
>  - likelihood-based docking of models into cryo-EM maps
> - MOPAC v22 is now distributed with Phenix
> - Quantum Mechanical Restraints (QMR) to calculate ligand restraints in situ
>  - Available in phenix.refine and separate command-line tool, 
> mmtbx.quantum_interface
>  - Defaults to MOPAC's PM6-D3H4
>  - Higher level QM available via 3rd-party Orca package
> - Restraints
>  - Approximately 27k of 37k restraints (QM calculated and validated) deployed 
> in the GeoStd
> - Automated tests for programs in the Phenix GUI
> - ERRASER no longer supported and removed from installation
> - Remove rebuild_predicted_model (duplicated by predict_and_build)
> - JSON output added to MolProbity structure validation scripts
>
> Please note that this publication should be used to cite the use of Phenix:
>
> Macromolecular structure determination using X-rays, neutrons and electrons: 
> recent developments in Phenix. Liebschner D, Afonine PV, Baker ML, Bunkóczi 
> G, Chen VB, Croll TI, Hintze B, Hung LW, Jain S, McCoy AJ, Moriarty NW, 
> Oeffner RD, Poon BK, Prisant MG, Read RJ, Richardson JS, Richardson DC, 
> Sammito MD, Sobolev OV, Stockwell DH, Terwilliger TC, Urzhumtsev AG, Videau 
> LL, Williams CJ, Adams PD: Acta Cryst. (2019). D75, 861-877. 
> https://doi.org/10.1107/S2059798319011471
>
> Full documentation is available here:
>
> http://www.phenix-online.org/documentation/
>
> There is a Phenix bulletin board:
>
> http://www.phenix-online.org/mailman/listinfo/phenixbb/
>
> Please consult the installer README file or online documentation for 
> installation instructions.
>
> Direct questions and problem reports to the bulletin board or:
>
> h...@phenix-online.org and b...@phenix-online.org
>
> Commercial users interested in obtaining access to Phenix should visit the 
> Phenix website for information about the Phenix Industrial Consortium.
>
> The development of Phenix is funded by the National Institute of General 
> Medical Sciences (NIH) under grant P01-GM063210. The maintenance and 
> distribution of Phenix is funded by the National Institute of General Medical 
> Sciences (NIH) under grant R24-GM141254. We also acknowledge the generous 
> support of the members of the Phenix Industrial Consortium.
>
> --
> Paul Adams (he/him/his)
> Associate Laboratory Director for Biosciences, LBL 
> (https://biosciences.lbl.gov/)
> Principal Investigator, Computational Crystallography Initiative, LBL 
> (http://cci.lbl.gov/)
> Vice President for Technology, the Joint BioEnergy Institute 
> (http://www.jbei.org/)
> Principal Investigator, ALS-ENABLE, Advanced Light Source 
> (http://als-enable.lbl.gov/)
> Laboratory Research Manager, ENIGMA Science Focus Area 
> (http://enigma.lbl.gov/)
> Adjunct Professor, Department of Bioengineering, UC Berkeley 
> (http://bioeng.berkeley.edu/)
> Member of the Graduate Group in Comparative Biochemistry, UC Berkeley 
> (http://compbiochem.berkeley.edu/)
>
> Building 91, Room 410
> Building 978, Room 4126
> Tel: 1-510-486-4225
> http://cci.lbl.gov/paul
> ORCID: -0001-9333-8219
>
> Lawrence Berkeley Laboratory
> 1 Cyclotron Road
> BLDG 91R0183
> Berkeley, CA 94720, USA.
>
> Executive Assistant: Michael Espinosa [ meespin...@lbl.gov ][ 1-510-333-6788 ]
> Phenix 

Re: [ccp4bb] what is isomorphous?

[Randy John Read]

I think we’ve strayed a bit from Doeke’s original question involving crystals 
A, B and C, where I think the consensus opinion would be that we would refer to 
crystal C as not being isomorphous to either A or B.

On the question of what “isomorphous” means in the context of related crystals, 
I’m not sure we have complete consensus. I would tend to say that any two 
crystals are isomorphous if they have related unit cells and similar fractional 
coordinates of the atoms, so that (operationally) their diffraction patterns 
are correlated. However, there might be differences of opinion on whether two 
crystals can be considered isomorphous if one has exact crystallographic 
symmetry and the other has pseudosymmetry. (I would probably be on the more 
permissive side here.)

In principle, I suppose being isomorphous (“same shape”) should be a binary 
decision, but in practice we’re interested in the implications of the degree to 
which perfect isomorphism is violated. So I would tend to use the term “poorly 
isomorphous” for a pair where the correlation between the diffraction patterns 
drops off well before the resolution limit. Crick was focused on percentage 
change in cell dimensions, but Bernhard is right that what matters is the ratio 
between the difference in cell lengths and the resolution of the data. It’s a 
bit counter-intuitive, but the effect of the difference between cell edges of 
20 and 25 is the same as for cell edges of 200 and 205! By the way, the first 
time I learned this was from K. Cowtan and I hadn’t realised it’s also in Jan 
Drenth’s book.

For isomorphous replacement (something some of us dimly remember from the days 
before AlphaFold), being poorly isomorphous is bad, but for cross-crystal 
averaging the more poorly isomorphous the better, because the molecular 
transform is being sampled in different places in reciprocal space.

Best wishes,

Randy Read

> On 21 Dec 2023, at 10:53, Jon Cooper 
> <488a26d62010-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hello Harry, 
> 
> I think this is the paper you mean:
> https://scripts.iucr.org/cgi-bin/paper?S0365110X56002552
> 
> They gave depressingly low estimates of how much the cell dimensions could 
> change in order for isomorphous replacement to still work. In reality, unit 
> cells can shrink and swell, but the fractional atomic coordinates remain 
> relatively unchanged (right?) so bigger unit cell differences still allow the 
> method to work. 
> 
> Best wishes, Jon Cooper. jon.b.coo...@protonmail.com
> 
> Sent from Proton Mail mobile
> 
> 
> 
>  Original Message 
> On 21 Dec 2023, 09:07, Harry Powell < 
> 193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
> Hi Didn’t Francis Crick have something to say about this in the early 1950s? 
> I’m sure it was published but off the top of my mind I can’t think where (one 
> of the more “established” members of this community will be able to give 
> chapter and verse)! If you want to read something a little more detailed than 
> people have mentioned here, there’s a “Methods in Enzymology” chapter by 
> Charlie Carter (?) et al from the early part of this century on the subject - 
> again, I can’t remember exactly who or when. Have a good break (which reminds 
> me to register for the CCP4 Study Weekend)! Harry > On 21 Dec 2023, at 08:04, 
> Tim Gruene wrote: > > Hi Doeke, > > you can take the coordinates of B and do 
> a rigid body refinement > against the data from A. If this map is sufficient 
> to reproduce model A > (including model building and more refinement cycles), 
> then B is > isomorphous to A. You can do this the other way round, and the 
> result > may not be the same - hence, the mathematical definition of 
> isomorphous > is not identical to the practical use of 'isomorphous' 
> structures when > it comes to phasing. You can repeat this for each side of 
> the triangle > (each in two directions) in order to label the semantic 
> triangle. > > Merry Christmas, more peace on earth and sanity for the 
> elections in > 2024! > > Tim > > On Wed, 20 Dec 2023 20:15:17 + "Hekstra, 
> Doeke Romke" > wrote: > >> Dear colleagues, >> >> Something to muse over 
> during the holidays: >> >> Let's say we have three crystal forms of the same 
> protein, for >> example crystallized with different ligands. Crystal forms A 
> and B >> have the same crystal packing, except that one unit cell dimension 
> >> differs by, for example, 3%. Crystal form C has a different crystal >> 
> packing arrangement altogether. What is the right nomenclature to >> describe 
> the relationship between these crystal forms? >> >> If A and B are 
> sufficiently different that their phases are >> essentially uncorrelated, 
> what do we call them? Near-isomorphous? >> Non-isomorphous? Do we need a 
> different term to distinguish them from >> C or do we call all three datasets 
> non-isomorphous? >> >> Thanks for helping us resolve our semantic tangle. >> 
> >> Happy holidays! >> Doeke >> 

Re: [ccp4bb] low resolution data refinement

[Randy John Read]

Hi,

Whether or not you’ll get anything useful from a 4.3 A MR solution depends on 
your question — maybe you’re interested in something large-scale like complex 
formation or domain movement, in which case it could tell you something.

In any case, it’s important to have a good starting model. Unless you have a 
higher-resolution crystal structure of the same protein, your best bet would be 
something like an AlphaFold model. Next, if you want to refine it, the 
structure will tend to get worse at that resolution unless you only allow it to 
change when required to agree with your data. There are tools for both Refmac 
and phenix.refine to use your good-quality starting model as a reference model, 
with local or torsion angle restraints, and you should be using these.

Best wishes,

Randy Read

> On 21 Nov 2023, at 12:02, Yahui Liu  wrote:
> 
> Dear all,
> I got a protein crystal dataset of 4.3 A and would like to some the structure 
> with MR.
> Now I am suffering with the refinement. I used both Refmac and Phenix.
> 
> Someone could give me a hand or  any suggestions?
> 
> All the best
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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
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] About model building

[Randy John Read]

Hi,

At 1.9A resolution there should be lots of possibilities, depending on the 
details.

You imply you have partial sequence information for chain B. Is there a genome 
for your plant or a relative of it? You could search for possible matches to 
your sequence, and then test all the AlphaFold models for the sequences that 
come up.

When you say chain A has known homologues, what was the sequence identity of 
the homologue you used as a model? If it wasn’t pretty high, you may well be 
better off using an AlphaFold model of the correct sequence for A.

Do you expect chain B to have helices or do you see helices in the map? Just 
adding some poly-Ala helices (which will tend to be locally very accurate if 
incomplete) will help to improve your phases. Arcimboldo would be a good tool 
for this, and you should also try it for completing from chain A.

Do you have any anomalous scattering signal in your data? Even if it’s not 
enough to solve the structure, any signal can be exploited, using MR-SAD, to 
improve your phases and in particular reduce model bias in your map.

Of course, other people have pointed out that you should make sure you can be 
confident in the MR solution. Also, you didn’t mentione whether there might be 
any issues in the data, like twinning. Presumably you would have mentioned if 
there was more than one copy of each protein in the a.u.

Best wishes,

Randy Read

> On 4 Nov 2023, at 14:04, Sam Tang  wrote:
> 
> Dear community,
> 
> I am solving the structure of a complex between proteins A and B, where A is 
> a protein with known homologs and B is a novel protein isolated from plant. 
> The diffraction data was at 1.9 Ang collected in-house, indexed to P321. 
> Using A as the search model, we have got a reasonable solution where, after 
> one round of refinement, the A chain fits the map pretty well. What's left 
> was to extend the termini and fit a few rotamers.
> 
> For protein B (B chain) I have tried the web version of ARP/wARP but the 
> outcome was not really good. The model was not successfully built as 
> indicated by low model completeness and score. The tricky thing may be that 
> we do not have the complete sequence information of this protein B in-hand. 
> (The other way round, we more or less wish to rely on the high resolution 
> data to confirm its sequence.) What approach would you then recommend to 
> build the B chain in this scenario? 
> 
> Thanks in advance and best regards,
> 
> Sam 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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
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] Obituaries for Michael James

[Randy John Read]

For anyone who knew Michael James (who was my PhD supervisor), you might be 
interested in looking at two recent obituaries. One is in the current issue of 
ACA Reflexions (https://www.amercrystalassn.org/aca-reflexions), and the other 
was published in the October issue of Acta Cryst D (Structural Biology): 
https://doi.org/10.1107/S2059798323006976.

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
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] anomalous map in Coot - Linux vs. Win difference?

[Randy John Read]

I’ve just been playing with a difference map to check something. If you fail to 
set the map as a difference map (either when opening the map initially, or 
using Calculate->Map Tools->Set map is a difference map), there’s nothing in 
the Display Manager->Properties window that tells you. I was wondering if that 
could be the issue, if the contour colour chosen when it isn’t set as a 
difference map doesn’t show up against the other map.

Randy

> On 2 Aug 2023, at 14:32, Andrea Smith  wrote:
> 
> Dear Eleanor,
> 
> yes, I made the printscreen with the setting that shows both maps contoured 
> at 0.08 = 3 rmsd.
> 
> I also attach a printscreen from a site where I believe a Zn ion is bound 
> (from my condition). Can see that Wincoot is drawing the map wrong...
> Andrea
> 
> 
> 
> On Wednesday, August 02, 2023 15:28 CEST, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
>  
>> 
>> Hmmm -are the two ano maps contoured at the same level? 
>> You can check this by setting the SCROLL option to the ano map.
>>  The 2mFo-DFc maps look pretty identical..which suggests the two input mtz 
>> files are similar.
>> You can get more information by viewing the inputs to make sure the values 
>> are the same..
>> Eleanor
>>   On Wed, 2 Aug 2023 at 13:53, Andrea Smith  
>> wrote:
>> Dear all,
>> 
>> I used refmac in CCP4Cloud and then I opened the generated .pdb and .mtz 
>> from the CCP4Cloud job both in Linux Coot 0.9.6. and WinCoot 0.9.6. I can 
>> see different anomalous maps in Coot and Wincoot - see attached printscreen 
>> where on the left there is a green electron density between the aspartates 
>> while on the right there is none. I tried to search if this is a bug but 
>> couldn't find info about this.
>> 
>> Am I doing something wrong? Do I have a local software issue?
>> 
>> Any advice would be appreciated, best,
>> Andrea
>>  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 

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Patenting ligand binding?

[Randy John Read]

I get the impression that these days you can patent just about anything, and 
instead of challenging your claims the patent examiners leave it to be settled 
later. A few years (probably one or two decades!) ago someone managed to patent 
the idea of a difference Fourier to detect binding, ignoring the many decades 
of prior art. As far as I am aware no-one ever tried to enforce that patent.

Randy

> On 28 Jul 2023, at 09:17, Winter, Graeme (DLSLtd,RAL,LSCI) 
> <6a19cead4548-dmarc-requ...@jiscmail.ac.uk> wrote:
>
> Discussion topics:
>
> - does this have any hope of being enforced since - as you point out - it is 
> a pretty obvious next step from X-rays
> - does anyone else who is _already doing this_ care
> - should we (individuals, not community) be patenting the obvious / ideas / 
> “inventions” just in case?
>
> (well aware that the authors of said patent will wake up to this in a few 
> hours)
>
> Personally, I am not comfortable with this, but maybe I am a dinosaur?
>
> Graeme
>
>> On 28 Jul 2023, at 09:13, Tim Gruene  wrote:
>>
>> Hi Graeme,
>>
>> do you want to discuss whether obvious steps can be patented?
>>
>> Cheers,
>> Tim
>>
>>
>> On Fri, 28 Jul 2023 07:45:38 + "Winter, Graeme (DLSLtd,RAL,LSCI)"
>> <6a19cead4548-dmarc-requ...@jiscmail.ac.uk> wrote:
>>
>>> Interesting
>>>
>>> https://www.freepatentsonline.com/20230228695.pdf
>>>
>>> Patent for use of electron diffraction to assess ligand binding
>>>
>>> Stumbled across this because the patent application cites my work -
>>> felt that this would be of interest to the community
>>>
>>> … discuss?
>>>
>>> Graeme
>>>
>>>
>>
>>
>>
>> --
>> --
>> Tim Gruene
>> Head of the Centre for X-ray Structure Analysis
>> Faculty of Chemistry
>> University of Vienna
>>
>> Phone: +43-1-4277-70202
>>
>> GPG Key ID = A46BEE1A
>>
>> 
>>
>> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
>> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
>> available at https://www.jiscmail.ac.uk/policyandsecurity/
>
>
> --
> 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/WA-JISC.exe?SUBED1=CCP4BB=1
>
> This message was issued to members of http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] AI and cis-peptides

[Randy John Read]

I’m not sure about other methods, but AlphaFold does predict peptides in both 
cis and trans configurations. In a recent paper, Osnat Herzberg and John Moult 
show that it was pretty successful in predicting proline cis-peptides, among 
the novel structures in the CASP15 set of targets 
(https://www.pnas.org/doi/10.1073/pnas.2221745120).

For non-proline cis-peptides, I’m not aware of published work but Tristan Croll 
has shown me examples of correctly-predicted non-proline cis-peptides, 
including cases where some of the related structures in the PDB have an 
incorrect trans configuration. This implies that AlphaFold is not slavishly 
reproducing what it has seen during training.

Best wishes,

Randy Read

> On 13 Jul 2023, at 13:56, Oliviero Carugo  
> wrote:
>
> Does anybody know if cis-peptides are predicted by the AI tools (AlphaFold2, 
> ColabFold, or ESM-2)?
>
> 
>
> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Mac M2 chip

[Randy John Read]

Also, we’re hoping that the next CCP4 update will include the changes to fix 
the same issue in CCP4, so please install it when it’s announced!

Randy Read

> On 5 Jul 2023, at 17:40, Luca Jovine 
> <9d27029f2b4b-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi Firdous, for Phaser the solution is simple - just install any nightly 
> version of PHENIX > 1.21rc1-5015 
> (fromhttps://phenix-online.org/download/nightly_builds.cgi?show_all=1), and 
> it should be fine.
> -Luca
>  From: Firdous Tarique 
> Date: Wednesday, 5 July 2023 at 18:33
> To: Luca Jovine 
> Cc: "CCP4BB@jiscmail.ac.uk" 
> Subject: Re: [ccp4bb] Mac M2 chip
>  Hi
>  I too have an issue with my latest M2 Mac mini where the Phaser in Phenix 
> runs extremely slow in comparision to my intel Mac laptop. I had also posted 
> it on the phenix bulletin board but so far did not receive any suggestions 
> how to fix it.
>  By any chance if you guys have the previous email threads where a similar 
> problem was fixed and now it is very fast, then can you please share that 
> with me?
>  Best
>  Firdous
>  On Wed, 5 Jul 2023, 10:54 Luca Jovine, 
> <9d27029f2b4b-dmarc-requ...@jiscmail.ac.uk> wrote:
>> 
>> 
>> Hi Dan,
>> Please see yesterday's posts on phenixbb (thread 
>> https://phenix-online.org/pipermail/phenixbb/2023-July/025527.html): there 
>> was an issue with Phaser, but this has now been fixed in and it runs 
>> superfast. There may be others, but whatever I've been throwing at my M2 in 
>> recent times seemed to run just fine (or, if there were issues, these were 
>> not M2-specific) - including other PHENIX programs, XDS, autoPROC/BUSTER and 
>> quite a few CCP4 programs of course.
>> HTH,
>> Luca
>> 
>> -Original Message-
>> From: CCP4 bulletin board > > on behalf of Daniel Bonsor 
>> >
>> Reply to: Daniel Bonsor > >
>> Date: Wednesday, 5 July 2023 at 17:22
>> To: "CCP4BB@JISCMAIL.AC.UK " 
>> mailto:CCP4BB@JISCMAIL.AC.UK>>
>> Subject: [ccp4bb] Mac M2 chip
>> 
>> 
>> [You don't often get email from 
>> a71eb7acd5aa-dmarc-requ...@jiscmail.ac.uk.
>>  Learn why this is important athttps://aka.ms/LearnAboutSenderIdentification 
>>  ]
>> 
>> 
>> Dear All,
>> 
>> 
>> Has anyone encountered problems with CCP4i/Phenix/XDS/etc on Macs with the 
>> M2 chip? I saw a post from 2020 on the M1 chip, but was curious about if 
>> there are any issues between the M2 chip and our crystallographic software.
>> 
>> 
>> Thanks for any information.
>> 
>> 
>> Dan
>> 
>> 
>> 
>> 
>> 
>> 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 
>> http://www.jiscmail.ac.uk/CCP4BB, a 
>> mailing list hosted by 
>> http://www.jiscmail.ac.uk/, terms & conditions 
>> are available athttps://www.jiscmail.ac.uk/policyandsecurity/ 
>> 
>> 
>> 
>> 
>> 
>> 
>> När du skickar e-post till Karolinska Institutet (KI) innebär detta att KI 
>> kommer att behandla dina personuppgifter. Här finns information om hur KI 
>> behandlar personuppgifter.
>> 
>> 
>> Sending email to Karolinska Institutet (KI) will result in KI processing 
>> your personal data. You can read more about KI’s processing of personal data 
>> here.
>> 
>> 
>> 
>> 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


-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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 

Re: [ccp4bb] British X-ray Crystallographers

[Randy John Read]

So have I! They’ll learn not to disrespect one of our favourite people.

Randy

> On 24 May 2023, at 10:44, Harry Powell 
> <193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
>
> Hi Gerard
>
> I’ve mentioned it to the organiser - we shall see how long the Garmen will be 
> there!
>
> Harry
>
>> On 24 May 2023, at 10:27, Gerard Bricogne  wrote:
>>
>> Dear Jon,
>>
>>Quite a line-up indeed.
>>
>>Might someone at the RSC correct the typo in Elspeth's surname (in two
>> places)? Or is it a plural form ;-) ?
>>
>>
>>With best wishes,
>>
>> Gerard
>>
>> --
>> On Tue, May 23, 2023 at 10:16:54PM +, Jon Cooper wrote:
>>> I am biased, but this looks like an interesting meeting:
>>>
>>> https://www.rsc.org/events/detail/76719/british-x-ray-crystallographers
>>>
>>> Best wishes, Jon Cooper.
>>> jon.b.coo...@protonmail.com
>>>
>>> Sent with [Proton Mail](https://proton.me/) secure email.
>>>
>>> 
>>>
>>> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
>>> mailing list hosted by http://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 http://www.jiscmail.ac.uk/CCP4BB, a 
>> mailing list hosted by http://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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] choosing an NMR structure from PDB

[Randy John Read]

Hi Harry,

My advice would be to use one of those new-fangled predicted models. You can 
find a model in the AlphaFold database at the EBI 
(https://alphafold.ebi.ac.uk/entry/P01132). If you look at it, there are parts 
(likely corresponding to the constructs that were crystallised) that look 
confidently predicted, connected by poorly-predicted loops. If you take the PDB 
file and the PAE matrix, you can run process_predicted_model either from Phenix 
or CCP4, which will give you individual files for the confident parts of the 
full prediction that are likely to have the correct relative orientations. (If 
you want to use the models for molecular replacement, you’ll find that the 
least-confident parts are downweighted by being assigned high B-factors, which 
is much better than having the best parts of the models, with pLDDT near 100, 
downweighted the most by interpreting pLDDT as a B-factor.) I would bet that 
these models will be more accurate than typical NMR models.

Best wishes,

Randy

> On 3 May 2023, at 11:45, Harry Powell 
> <193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
>
> Hi folks
>
> I was wondering.
>
> If there is a UniProt entry (for example, P01132, but there are plenty of 
> others) for which I want the “best” (whatever that might mean) representative 
> _experimental_ structure (i.e. not one of these new-fangled predicted models 
> that some folk say have removed the need for actually doing experiments), but 
> there are only NMR models - how do I choose?
>
> I don’t mean “which model from the ensemble do I choose” - that’s a different 
> question.
>
> For P01132, for example, I could choose (from the PDB) 1A3P, 1EGF, 1EPG, 
> 1EPH, 1EPI, 1EPJ, 1GK5 or 3EGF. Note that some of these are from the same 
> paper, so may be in different conditions (e.g. pH). All except the first 
> (1A3P) cover the same bit of sequence.
>
> Specifically, what should I look for in the downloadable files (mmCIF, for 
> example) from the PDB?
>
> Thoughts?
>
> Harry
> 
>
> 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 http://www.jiscmail.ac.uk/CCP4BB, a 
> mailing list hosted by http://www.jiscmail.ac.uk/, terms & conditions are 
> available at https://www.jiscmail.ac.uk/policyandsecurity/

-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: +44 1223 336500
The Keith Peters Building
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] Structure prediction - waiting to happen

[Randy John Read]

There’s also this preprint with Tom Terwilliger as lead author: 
https://www.biorxiv.org/content/10.1101/2022.11.21.517405v1. The title is 
“AlphaFold predictions: great hypotheses but no match for experiment”.

Best wishes,

Randy

> On 1 Apr 2023, at 18:18, Savvas Savvides 
> <9d24f7f13e09-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Dear Rams,
> 
> I salute you for sharing this.
> 
> Just a week ago, I also received a remark along these lines on a declined 
> grant application. The remark was the only unfavourable point, which 
> suggested that it must have weighed disproportionally towards the negative 
> outcome. This was a two-stage evaluation process and the grant was cut in 
> stage-1 where it was evaluated by a small group of evaluators, none of whom 
> was a structural biologist/biochemist. Stage-2 would have involved peer 
> review by international experts.
> 
> Despite my initial disbelief about what this remark might have caused and 
> upon reflection, I realized that it might be time to become proactive in 
> future applications in anticipation of the apparent growing trend towards 
> such remarks and perceptions.
> 
> I think that a generalized form of preemptive text might not serve the 
> purpose well, but perhaps well-articulated statements specific to the 
> proposed biological problem at hand (perhaps aided by illustrations 
> demonstrating the inability of structure prediction to address the problem at 
> hand) might be the better way to go. Even though many of us who teach courses 
> in experimental structural biology and structural bioinformatics at 
> undergraduate and graduate levels are already actively addressing many of 
> these issues, there is a much bigger and far more senior scientific 
> population out there that makes important decisions on science 
> policy/funding/infrastructures/evaluations/recruitment/etc that are not 
> getting such educational exposure.
> 
> The following resources provide good material and starting points to reflect 
> and elaborate upon.
> 
> The article by Perrakis and Sixma in EMBO Reports 
> https://www.embopress.org/doi/full/10.15252/embr.202154046
> 
> The recent comment paper in Nature Methods by Thomas Jane
> https://doi.org/10.1038/s41592-022-01760-4 
> 
> A correspondence in Science by Moore, Hendrickson, Henderson and Brunger
> https://www.science.org/doi/10.1126/science.abn9422?url_ver=Z39.88-2003_id=ori:rid:crossref.org_dat=cr_pub%20%200pubmed
> 
> 
> Best wishes
> Savvas
> 
> 
> 
> Savvas Savvides
> VIB Center for Inflammation Research
> Ghent University, Dept. of Biochemistry & Microbiology
> Technologiepark 71, 9052 Ghent, Belgium
> 
> Email: savvas.savvi...@ugent.be ; savvas.savvi...@irc.vib-ugent.be
> Phone: +32 (0)472 928 519 (mobile) ; +32 (0)9 331 36 60 (office)
> Web: https://savvideslab.sites.vib.be/en#/
> 
>> On 1 Apr 2023, at 16:57, Subramanian, Ramaswamy  wrote:
>> 
>> Ian,
>> 
>> Thank you.  This is not an April fools..  
>> Rams
>> subra...@purdue.edu
>> 
>> 
>> 
>>> On Apr 1, 2023, at 10:46 AM, Ian Tickle  wrote:
>>> 
>>>  External Email: Use caution with attachments, links, or sharing data 
>>> 
>>> 
>>> 
>>> Hi Ramaswamy
>>> 
>>> I assume this is an April Fool's but it's still a serious question because 
>>> many reviewers who are not crystallographers or electron microscopists may 
>>> not fully appreciate the difference currently between the precision of 
>>> structures obtained by experimental and predictive methods, though the 
>>> latter are certainly catching up.  The answer of course lies in the mean 
>>> co-ordinate precision, related to the map resolution.
>>> 
>>> Quoting https://people.cryst.bbk.ac.uk/~ubcg05m/precgrant.html :
>>> 
>>> "The accuracy and precision required of an experimentally determined model 
>>> of a macromolecule depends on the biological questions being asked of the 
>>> structure.  Questions involving the overall fold of a protein, or its 
>>> topological similarity to other proteins, can be answered by structures of 
>>> fairly low precision such as those obtained from very low resolution X-ray 
>>> crystal diffraction data [or AlphaFold].  Questions involving reaction 
>>> mechanisms require much greater accuracy and precision as obtained from 
>>> well-refined, high-resolution X-ray structures, including proper 
>>> statistical analyses of the standard uncertainties (s.u.'s) of atomic 
>>> positions and bond lengths.".
>>> 
>>> According to https://www.nature.com/articles/s41586-021-03819-2 :
>>> 
>>> The accuracy of AlphaFold structures at the time of writing (2021) was 
>>> around 1.0 Ang. RMSD for main-chain and 1.5 Ang. RMSD for side-chain atoms 
>>> and probably hasn't changed much since.  This is described as "highly 
>>> accurate"; however this only means that AlphaFold's accuracy is much higher 
>>> in comparison with other prediction methods, not in comparison with 
>>> experimental methods.  Also note that AlphaFold's accuracy is estimated by 
>>> comparison 

Re: [ccp4bb] anomalous data usage

[Randy John Read]

Hi Jon,

My understanding of the philosophy is that new users would prefer to think 
about crystallographic data objects, rather than worrying about the arcana of 
MTZ files and the many different flavours of columns. There are tradeoffs — it 
can indeed be more difficult to find the bits of information you need, but you 
should be thinking in terms of the stored objects from the imports at the 
beginning of the project, rather than the files that hold them.

Personally, I find multicolumn MTZ files easier to think about, but my brain 
probably rigidified a decade or two ago!

Best wishes,

Randy

> On 15 Mar 2023, at 13:09, Jon Cooper 
> <488a26d62010-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hello Ian, 
> 
> if I understand you and Eleanor correctly, this is the philosophy of the 
> mini-MTZ, i.e. if you are doing anything independent of i2, you have to dig 
> around a bit to find which output file contains the columns you need.
> 
> Best wishes, Jon Cooper. jon.b.coo...@protonmail.com
> 
> Sent from Proton Mail mobile
> 
> 
> 
>  Original Message 
> On 13 Mar 2023, 23:27, Ian Tickle < ianj...@gmail.com> wrote:
> 
> 
> Eleanor, which program is doing that and more to the point, why?
> 
> -- Ian
> 
> 
> On Mon, 13 Mar 2023 at 20:17, Eleanor Dodson  
> wrote:
> fIf you are using ccp4I2 for some forgotten reason the final output has one 
> reflection with I+ and I-, another with Imean, another with Fmean - aagghh
> 
> 
> On Mon, 13 Mar 2023 at 19:40, Ian Tickle  wrote:
> 
> Hi Gottfried
> 
> AIMLESS definitely outputs IMEAN (and SIGIMEAN) by default.
> 
> Cheers
> 
> -- Ian
> 
> 
> On Mon, 13 Mar 2023 at 18:53, Palm, Gottfried  wrote:
> Dear all, 
>   I have a few questions handling (non) anomalous data:
> By default aimless seems to produce Iplus and Iminus columns. Can I force it 
> to (also) create an Imean column?
> What does refmac do, when it gets Iplus and Iminus (and their sigmas) as 
> input. Does it take only one of them or does it calculate and use Imean?
> Greetings
>   Gottfried
> 
> 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
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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
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] problems in Rfree-flags when refining with REFMAC and processing with STARANISO

[Randy John Read]

Also, this might be a good time for the regular reminder *never* to use the MTZ 
output by Refmac as input for anything else, because the FOBS coming out of 
Refmac is not the same as the FOBS going in. It might well have had various 
corrections applied, including a correction for twinning. I think the newer 
ccp4i2 and CCP4Cloud interfaces might make it harder for you to do that than 
the old ccp4i did.

Best wishes,

Randy Read

> On 13 Jul 2022, at 10:35, Clemens Vonrhein  wrote:
> 
> Dear Vera,
> 
> On Wed, Jul 13, 2022 at 10:13:28AM +0200, Vera Pfanzagl wrote:
>> You write that I should make sure REFMAC reads in the test-set from
>> the staraniso.mtz and does not replace it with a new one.  This
>> would only be the case if it generates a new test set when I
>> initially import the reflection file, right? How can I check that
>> retrospectively with my history-riddled file?
> 
> The REFMAC program/binary itself will obviously read an MTZ file
> exactly as-is when run via
> 
>  refmac5 xyzin your.pdb hklin your.mtz xyzout refmac.pdb hklout refmac.mtz <  ... some commands ...
>  e
> 
> But "running REFMAC" can mean very different things when this is done
> via one of the available interfaces: there might be various
> preparation steps involved before the actual REFMAC binary is called.
> 
> Best thing: run e.g.
> 
>  mtzdmp input.mtz  -n 100 > tmp.1
>  mtzdmp output.mtz -n 100 > tmp.2
> 
> and compare those two text files (same number of reflections reported,
> same completeness for test-set flags etc).
> 
>> PS: the links in [2] are not accurate anymore, the interface of
>> CCP4i2 export alone already looks completely different and does not
>> seem to want external processing files.
> 
> Probably something for wwPDB/CCP4/REFMAC to look into and to provide
> an update.
> 
> Cheers
> 
> Clemens
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jiscmail.ac.uk%2Fcgi-bin%2FWA-JISC.exe%3FSUBED1%3DCCP4BB%26A%3D1data=05%7C01%7Crjr27%40universityofcambridgecloud.onmicrosoft.com%7Ce0cbc7b0db474dabc77408da64b30f07%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637933017705425094%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7Csdata=TLGsiJIln02X8xjtFGLfVmRfUSOkswK%2F4N8CseF9u9I%3Dreserved=0
> 
> This message was issued to members of 
> https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jiscmail.ac.uk%2FCCP4BBdata=05%7C01%7Crjr27%40universityofcambridgecloud.onmicrosoft.com%7Ce0cbc7b0db474dabc77408da64b30f07%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637933017705425094%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7Csdata=yyB6WLCqWnWwrZiKRgv%2Fbj%2FEPBNpSt8fARNORqlnmaY%3Dreserved=0,
>  a mailing list hosted by 
> https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jiscmail.ac.uk%2Fdata=05%7C01%7Crjr27%40universityofcambridgecloud.onmicrosoft.com%7Ce0cbc7b0db474dabc77408da64b30f07%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637933017705425094%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7Csdata=VxP6HX%2Fd05GYKAIYksG7xsfbPgYL0ix5broxj2RnHDg%3Dreserved=0,
>  terms & conditions are available at 
> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jiscmail.ac.uk%2Fpolicyandsecurity%2Fdata=05%7C01%7Crjr27%40universityofcambridgecloud.onmicrosoft.com%7Ce0cbc7b0db474dabc77408da64b30f07%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637933017705425094%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7Csdata=0X3Zdu2ZfbJNSZ54yjtRTaziKDgqpVQvyQ5Y3Qt3cQY%3Dreserved=0

-
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] CCP4i under M1Pro

[Randy John Read]

We don’t have any computers with Apple silicon in our group yet to test Phaser 
on ourselves. When preparing the Phenix distribution, Billy Poon has run tests 
and when Phaser is compiled for Intel and run through Rosetta 2, it’s generally 
faster than on a recent Intel Mac. However, there have been reports from Phenix 
users of severe slowdowns, and we don’t understand what the special 
circumstances are that cause it to slow down on specific jobs.

The best performance is when it’s compiled natively for the M1 processor, and I 
believe both Phenix and CCP4 are aiming to distribute natively-compiled 
software as soon as technical complications are worked out. So hopefully this 
is just temporary!

I hadn’t realised that the OpenMP wasn’t enabled on CCP4. In Phenix, there were 
issues preventing the use of OpenMP with Python2, but those should go away when 
the Python3 version is ready to deploy.

Best wishes,

Randy Read

> On 28 Jun 2022, at 16:45, Lau Kelvin 
> <5aaf8435dbef-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hello
> 
> I have noticed running normal programs like Phaser MR on a MacBook Pro 
> (M1Pro) that jobs are extremely slow compared to my older computers.
> 
> It doesn’t seem to be running over multiple CPUs anymore.
> 
> Has this also been seen by others?
> 
> Best
> 
> Kelvin
> 
> 
> -- 
> Kelvin Lau
> Protein production and structure core facility - PTPSP
> EPFL SV PTECH PTPSP 
> AI 2146 (Bâtiment AI) 
> Station 19 
> CH-1015 Lausanne
> Switzerland
> Email: kelvin@epfl.ch
> Phone: +41 21 69 34494
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] Validation of structure prediction

[Randy John Read]

Just to add one point that I don’t think I’ve seen yet. If what the referee 
wants is a data-free assessment of the expected quality of the model, I think 
that the best assessment at the moment is the one done by AlphaFold2 (or indeed 
RoseTTAFold if you’re using one of their models). The machine-learning 
algorithm is pretty good at assessing how good of a job it has done, either 
overall (predicted TM score) or locally (predicted lDDT score for AlphaFold2 or 
predicted RMSD for RoseTTAFold). There are cases of false positives (poor 
models that think they’re good) and false negatives (good models that think 
they’re bad), but these are in the minority from what I’ve seen so far.

Of the tools mentioned earlier, I think ProQ2 is the only one that is really 
trained to assess predicted model quality, but I suspect it’s not as good at 
assessing the quality of the latest generation of models as the tools 
generating those models are.

Best wishes,

Randy Read

> On 21 Dec 2021, at 11:12, Kay Diederichs  
> wrote:
> 
> Hi Reza,
> 
> the term "validation" as used by e.g. crystallographers - namely by checking 
> geometric parameters of a structure derived from experiment(s) - is 
> euphemistic since realistic geometry is a required but not sufficient 
> property of a model - it can be completely wrong even if it has good 
> geometry. This type of validation should rather be called "checked for 
> geometric consistency" or similar.
> 
> In general, the validation of a prediction should be performed by 
> experiment(s). Ideally, that would be X-ray structure solution or cryo-EM or 
> NMR. But it could also be e.g. a set of binding or cross-linking experiments, 
> with suitable positive and negative controls. 
> 
> That a computational prediction can be "validated" by another calculation may 
> be possible if the prediction is based on a set of facts that is disjoint 
> from that of the "validation calculation". But this is not the case e.g. for 
> models from AlphaFold or RoseTTaFold; MolProbility or similar programs could 
> at most indicate that the prediction has failed, but not that the prediction 
> is correct. Experiments are needed for this - and even these could be 
> inconclusive.
> 
> Best wishes,
> Kay
> 
> On Mon, 20 Dec 2021 16:10:02 +, Reza Khayat  wrote:
> 
>> ?Hi,
>> 
>> Can anyone suggest how to validate a predicted structure? Something similar 
>> to wwPDB validation without the need for refinement statistics. I realize 
>> this is a strange question given that the geometry of the model is 
>> anticipated to be fine if the structure was predicted by a server that 
>> minimizes the geometry to improve its statistics. Nonetheless, the journal 
>> has asked me for such a report. Thanks.
>> 
>> Best wishes,
>> 
>> Reza
>> 
>> 
>> Reza Khayat, PhD
>> Associate Professor
>> City College of New York
>> Department of Chemistry and Biochemistry
>> New York, NY 10031
>> 
>> 
>> 
>> 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/

-
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] [ccp4bb] question using a model from a multiple model PDB file using Phaser

[Randy John Read]

Hi Yong,

Sorry, this is a scripting feature I haven’t actually used before! It turns out 
there was an error in the documentation, and where you select one MODEL from a 
multimodel PDB file (including an ensemble from ensembler or even one model 
from an NMR ensemble), you have to put “MODEL SERIAL”. It takes the model that 
is named in the PDB file with the specified model number.

I’ve just fixed the documentation on our website.

Best wishes,

Randy Read

> On 9 Dec 2021, at 21:27, Yong Wang  wrote:
> 
> Hi Randy,
> 
> I did not use the ensemble as I am not doing the typical jobs: I don't want 
> an ensemble as a template, instead I would like to test each individual model 
> separately.  I re-formatted a multi-model pdb file (the original file did not 
> have the MODEL line) on my own.  Changing to "MODEL 0" did not change the 
> error message.  By the way, Phaser output suggested " SYNTAX ERROR: Use 
> SERIAL".  What is the usage of "SERIAL"?  I can't find that keyword in the 
> manual.
> 
> Thanks,
> 
> Yong
> 
> -----Original Message-
> From: CCP4 bulletin board  On Behalf Of Randy John Read
> Sent: Thursday, December 9, 2021 4:18 PM
> To: CCP4BB@JISCMAIL.AC.UK
> Subject: [EXTERNAL] Re: [ccp4bb] question using a model from a multiple model 
> PDB file using Phaser
> 
> EXTERNAL EMAIL: Use caution before replying, clicking links, and opening 
> attachments.
> 
> Hi Yong,
> 
> If you make an ensemble with ensembler, the models are numbered from 1 to n, 
> so there’s no requirement to start from 0.  Did you use ensembler to make 
> your ensemble model file?  If you did and Phaser is failing, could you send 
> me the file so we can figure out the problem?  Otherwise, the easiest 
> solution might be to use ensembler to generate the ensemble!
> 
> Best wishes,
> 
> Randy Read
> 
>> On 9 Dec 2021, at 21:00, Yong Wang 
>> <3c4fc05cc53b-dmarc-requ...@jiscmail.ac.uk> wrote:
>> 
>> Hi,
>> 
>> 
>> 
>> The first model number in the pdb file is 1.  In the script it is “MODEL 1”. 
>>  I will try “MODEL 0” in the script to see how it goes.
>> 
>> 
>> 
>> Thanks,
>> 
>> 
>> 
>> Yong
>> 
>> 
>> 
>> From: Boaz Shaanan  
>> Sent: Wednesday, December 8, 2021 7:37 AM
>> To: Yong Wang 
>> Cc: CCP4BB@JISCMAIL.AC.UK
>> Subject: [EXTERNAL] Re: [ccp4bb] question using a model from a multiple 
>> model PDB file using Phaser
>> 
>> 
>> 
>> EXTERNAL EMAIL: Use caution before replying, clicking links, and opening 
>> attachments.
>> 
>> 
>> Hi,
>> 
>> Did you try to edit 'Model 1' to 'Model 0'?
>> 
>> I seem to recall that it was reported on this bb as the solution to this 
>> problem.
>> 
>> Cheers,
>> 
>> Boaz 
>> 
>> Boaz Shaanan, Ph.D.
>> Dept. of Life Sciences
>> Ben Gurion University
>> Beer Sheva, Israel
>> 
>> 
>> 
>> On 7 Dec 2021 22:56, Yong Wang 
>> <3c4fc05cc53b-dmarc-requ...@jiscmail.ac.uk> wrote:
>> 
>> Hi All,
>> 
>> 
>> 
>> I am interested in using individual models from a PDB file containing 
>> multiple models with MODEL/ENDMDL records as a template for use in Phaser.   
>> The usage in the manual “ENSEMBLE  [PDB|CIF]  [RMS 1 | 
>> ID 2 | CARD ON3] CHAIN ""4 MODEL 5 ” seems to support this.  
>> However, I am getting an error shown below:
>> 
>> 
>> 
>> ENSEmble mol PDB out.pdb RMS 1.5 MODEL 1
>> 
>> 
>> 
>> $$
>> 
>> 
>> 
>> $TEXT:Warning: $$ Baubles Markup $$
>> 
>> --
>> 
>> SYNTAX ERROR: Use SERIAL
>> 
>> --
>> 
>> $$
>> 
>> 
>> 
>> I am assuming that the error is from using the “MODEL 1” and not due to 
>> something else.  Does anyone have any suggestion on this problem?
>> 
>> 
>> 
>> Thanks,
>> 
>> 
>> 
>> Yong
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> Yong Wang, Ph.D.
>> 
>> 
>> 
>> Senior Research Advisor
>> 
>> Global Structural Biology
>> 
>> Lilly Research Laboratories
>> 
>> Lilly Corporate Center
>> 
>> Indianapolis, IN 46285
>> 
>> U.S.A.
>> 
>> 
>> 
>> Tel: +1 (317) 655 9145
>> 
>> 
>> 
>> 
>> 
>> 
>

Re: [ccp4bb] question using a model from a multiple model PDB file using Phaser

[Randy John Read]

Hi Yong,

If you make an ensemble with ensembler, the models are numbered from 1 to n, so 
there’s no requirement to start from 0.  Did you use ensembler to make your 
ensemble model file?  If you did and Phaser is failing, could you send me the 
file so we can figure out the problem?  Otherwise, the easiest solution might 
be to use ensembler to generate the ensemble!

Best wishes,

Randy Read

> On 9 Dec 2021, at 21:00, Yong Wang 
> <3c4fc05cc53b-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi,
> 
>  
> 
> The first model number in the pdb file is 1.  In the script it is “MODEL 1”.  
> I will try “MODEL 0” in the script to see how it goes.
> 
>  
> 
> Thanks,
> 
>  
> 
> Yong
> 
>  
> 
> From: Boaz Shaanan  
> Sent: Wednesday, December 8, 2021 7:37 AM
> To: Yong Wang 
> Cc: CCP4BB@JISCMAIL.AC.UK
> Subject: [EXTERNAL] Re: [ccp4bb] question using a model from a multiple model 
> PDB file using Phaser
> 
>  
> 
> EXTERNAL EMAIL: Use caution before replying, clicking links, and opening 
> attachments.
>  
> 
> Hi,
> 
> Did you try to edit 'Model 1' to 'Model 0'?
> 
> I seem to recall that it was reported on this bb as the solution to this 
> problem.
> 
> Cheers,
> 
> Boaz 
> 
> Boaz Shaanan, Ph.D.
> Dept. of Life Sciences
> Ben Gurion University
> Beer Sheva, Israel
> 
>  
> 
> On 7 Dec 2021 22:56, Yong Wang 
> <3c4fc05cc53b-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi All,
> 
>  
> 
> I am interested in using individual models from a PDB file containing 
> multiple models with MODEL/ENDMDL records as a template for use in Phaser.   
> The usage in the manual “ENSEMBLE  [PDB|CIF]  [RMS 1 | 
> ID 2 | CARD ON3] CHAIN ""4 MODEL 5 ” seems to support this.  
> However, I am getting an error shown below:
> 
>  
> 
> ENSEmble mol PDB out.pdb RMS 1.5 MODEL 1
> 
>  
> 
> $$
> 
> 
> 
> $TEXT:Warning: $$ Baubles Markup $$
> 
> --
> 
> SYNTAX ERROR: Use SERIAL
> 
> --
> 
> $$
> 
>  
> 
> I am assuming that the error is from using the “MODEL 1” and not due to 
> something else.  Does anyone have any suggestion on this problem?
> 
>  
> 
> Thanks,
> 
>  
> 
> Yong
> 
>  
> 
>  
> 
>  
> 
> Yong Wang, Ph.D.
> 
>  
> 
> Senior Research Advisor
> 
> Global Structural Biology
> 
> Lilly Research Laboratories
> 
> Lilly Corporate Center
> 
> Indianapolis, IN 46285
> 
> U.S.A.
> 
>  
> 
> Tel: +1 (317) 655 9145
> 
>  
> 
>  
> 
>  
> 
>  
> 
> 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
> 

-
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] am I doing this right?

[Randy John Read]

Hi Kay,

No, I still think the answer should come out the same if you have good reason 
to believe that all the 100 pixels are equally likely to receive a photon (for 
instance because your knowledge of the geometry of the source and the detector 
says the difference in their positions is insignificant, i.e. part of your 
prior expectation). Unless the exact position of the spot where you detect the 
photon is relevant, detecting 1 photon on a big pixel and detecting the same 
photon on 1 of 100 smaller pixels covering the same area are equivalent events. 
What should be different in the analysis, if you're thinking about individual 
pixels, is that the expected value for a photon landing on any of the pixels 
will be 100 times lower for each of the smaller pixels than the single big 
pixel, so that the expected value of their sum is the same. You won't get to 
that conclusion without having a different prior probability for the two cases 
that reflects the 100-fold lower flux through the smaller area, regardless of 
the total power of the source.

Best wishes,
Randy

On 21 Oct 2021, at 13:03, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de>> wrote:

Randy,

I must admit that I am not certain about my answer, but I lean toward thinking 
that the result (of the two thought experiments that you describe) is not the 
same. I do agree that it makes sense that the expectation value is the same, 
and the math that I sketched in 
https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=CCP4BB;bdd31b04.2110 actually 
shows this. But the variance? To me, a 100-pixel patch with all zeroes is no 
different from sequentially observing 100 pixels, one after the other. For the 
first of these pixels, I have no idea what the count is, until I observe it. 
For the second, I am less surprised that it is 0 because I observed 0 for the 
first. And so on, until the 100th. For the last one, my belief that I will 
observe a zero before I read out the pixel is much higher than for the first 
pixel. The variance is just the inverse of the amount of error (squared) that 
we assign to our belief in the expectation value. And that amount of belief is 
very different. I find it satisfactory that the sigma goes down with the sqrt() 
of the number of pixels.

Also, I don't find an error in the math of my posting of Mon, 18 Oct 2021 
15:00:42 +0100 . I do think that a uniform prior is not realistic, but this 
does not seem to make much difference for the 100-pixel thought experiment.

We could change the thought experiment in the following way - you observe 99 
pixels with zero counts, and 1 with 1 count. Would you still say that both the 
big-pixel-single-observation and the 100-pixel experiment should give 
expectation value of 2 and variance of 2? I wouldn't.

Best wishes,
Kay

On Thu, 21 Oct 2021 09:00:23 +, Randy John Read 
mailto:rj...@cam.ac.uk>> wrote:

Just to be a bit clearer, I mean that the calculation of the expected value and 
its variance should give the same answer if you're comparing one pixel for a 
particular length of exposure with the sum obtained from either a larger number 
of smaller pixels covering the same area for the same length of exposure, or 
the sum from the same pixel measured for smaller time slices adding up to the 
same total exposure.

On 21 Oct 2021, at 09:54, Randy John Read 
mailto:rj...@cam.ac.uk><mailto:rj...@cam.ac.uk>> wrote:

I would think that if this problem is being approached correctly, with the 
right prior, it shouldn't matter whether you collect the same signal 
distributed over 100 smaller pixels or the same pixel measured for the same 
length of exposure but with 100 time slices; you should get the same answer. So 
I would want to formulate the problem in a way where this invariance is 
satisfied. I thought it was, from some of the earlier descriptions of the 
problem, but this sounds worrying.

I think you're trying to say the same thing here, Kay. Is that right?

Best wishes,

Randy

On 21 Oct 2021, at 08:51, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de><mailto:kay.diederi...@uni-konstanz.de>>
 wrote:

Hi Ian,

it is Iobs=0.01 and sigIobs=0.01 for one pixel, but adding 100 pixels each with 
variance=sigIobs^2=0.0001 gives  0.01 , yielding a 100-pixel-sigIobs of 0.1 - 
different from the 1 you get. As if repeatedly observing the same count of 0 
lowers the estimated error by sqrt(n), where n is the number of observations 
(100 in this case).

best wishes,
Kay

On Wed, 20 Oct 2021 13:08:33 +0100, Ian Tickle 
mailto:ianj...@gmail.com><mailto:ianj...@gmail.com>> wrote:

Hi Kay

Can I just confirm that your result Iobs=0.01 sigIobs=0.01 is the estimate
of the true average intensity *per pixel* for a patch of 100 pixels?  So
then the total count for all 100 pixels is 1 with variance also 1, or in
general for k observed counts in the patch, expectation = variance = k+1
for the total count, irrespective of the number of pixels

Re: [ccp4bb] am I doing this right?

[Randy John Read]

Just to be a bit clearer, I mean that the calculation of the expected value and 
its variance should give the same answer if you're comparing one pixel for a 
particular length of exposure with the sum obtained from either a larger number 
of smaller pixels covering the same area for the same length of exposure, or 
the sum from the same pixel measured for smaller time slices adding up to the 
same total exposure.

On 21 Oct 2021, at 09:54, Randy John Read 
mailto:rj...@cam.ac.uk>> wrote:

I would think that if this problem is being approached correctly, with the 
right prior, it shouldn't matter whether you collect the same signal 
distributed over 100 smaller pixels or the same pixel measured for the same 
length of exposure but with 100 time slices; you should get the same answer. So 
I would want to formulate the problem in a way where this invariance is 
satisfied. I thought it was, from some of the earlier descriptions of the 
problem, but this sounds worrying.

I think you're trying to say the same thing here, Kay. Is that right?

Best wishes,

Randy

On 21 Oct 2021, at 08:51, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de>> wrote:

Hi Ian,

it is Iobs=0.01 and sigIobs=0.01 for one pixel, but adding 100 pixels each with 
variance=sigIobs^2=0.0001 gives  0.01 , yielding a 100-pixel-sigIobs of 0.1 - 
different from the 1 you get. As if repeatedly observing the same count of 0 
lowers the estimated error by sqrt(n), where n is the number of observations 
(100 in this case).

best wishes,
Kay

On Wed, 20 Oct 2021 13:08:33 +0100, Ian Tickle 
mailto:ianj...@gmail.com>> wrote:

Hi Kay

Can I just confirm that your result Iobs=0.01 sigIobs=0.01 is the estimate
of the true average intensity *per pixel* for a patch of 100 pixels?  So
then the total count for all 100 pixels is 1 with variance also 1, or in
general for k observed counts in the patch, expectation = variance = k+1
for the total count, irrespective of the number of pixels?  If so then that
agrees with my own conclusion.  It makes sense because Iobs=0.01
sigIobs=0.01 cannot come from a Poisson process (which obviously requires
expectation = variance = an integer), whereas the total count does come
from a Poisson process.

The difference from my approach is that you seem to have come at it via the
individual pixel counts whereas I came straight from the Agostini result
applied to the whole patch.  The number of pixels seems to me to be
irrelevant for the whole patch since the design of the detector, assuming
it's an ideal detector with DQE = 1 surely cannot change the photon flux
coming from the source: all ideal detectors whatever their pixel layout
must give the same result.  The number of pixels is then only relevant if
one needs to know the average intensity per pixel, i.e. the total and s.d.
divided by the number of pixels.  Note the pixels here need not even
correspond to the hardware pixels, they can be any arbitrary subdivision of
the detector surface.

Best wishes

-- Ian


On Tue, 19 Oct 2021 at 12:39, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de>>
wrote:

James,

I am saying that my answer to "what is the expectation and variance if I
observe a 10x10 patch of pixels with zero
counts?" is Iobs=0.01 sigIobs=0.01 (and Iobs=sigIobs=1 if there is only
one pixel) IF the uniform prior applies. I agree with Gergely and others
that this prior (with its high expectation value and variance) appears
unrealistic.

In your posting of Sat, 16 Oct 2021 12:00:30 -0700 you make a calculation
of Ppix that appears like a more suitable expectation value of a prior to
me. A suitable prior might then be 1/Ppix * e^(-l/Ppix) (Agostini §7.7.1).
The Bayesian argument is IIUC that the prior plays a minor role if you do
repeated measurements of the same value, because you use the posterior of
the first measurement as the prior for the second, and so on. What this
means is that your Ppix must play the role of a scale factor if you
consider the 100-pixel experiment.
However, for the 1-pixel experiment, having a more suitable prior should
be more important.

best,
Kay




On Mon, 18 Oct 2021 12:40:45 -0700, James Holton 
mailto:jmhol...@lbl.gov>> wrote:

Thank you very much for this Kay!

So, to summarize, you are saying the answer to my question "what is the
expectation and variance if I observe a 10x10 patch of pixels with zero
counts?" is:
Iobs = 0.01
sigIobs = 0.01 (defining sigIobs = sqrt(variance(Iobs)))

And for the one-pixel case:
Iobs = 1
sigIobs = 1

but in both cases the distribution is NOT Gaussian, but rather
exponential. And that means adding variances may not be the way to
propagate error.

Is that right?

-James Holton
MAD Scientist



On 10/18/2021 7:00 AM, Kay Diederichs wrote:
Hi James,

I'm a bit behind ...

My answer about the basic question ("a patch of 100 pixels each with
zero counts - what is the variance?") you ask is the following:

1) we all know the Poisson P

Re: [ccp4bb] am I doing this right?

[Randy John Read]

I would think that if this problem is being approached correctly, with the 
right prior, it shouldn't matter whether you collect the same signal 
distributed over 100 smaller pixels or the same pixel measured for the same 
length of exposure but with 100 time slices; you should get the same answer. So 
I would want to formulate the problem in a way where this invariance is 
satisfied. I thought it was, from some of the earlier descriptions of the 
problem, but this sounds worrying.

I think you're trying to say the same thing here, Kay. Is that right?

Best wishes,

Randy

On 21 Oct 2021, at 08:51, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de>> wrote:

Hi Ian,

it is Iobs=0.01 and sigIobs=0.01 for one pixel, but adding 100 pixels each with 
variance=sigIobs^2=0.0001 gives  0.01 , yielding a 100-pixel-sigIobs of 0.1 - 
different from the 1 you get. As if repeatedly observing the same count of 0 
lowers the estimated error by sqrt(n), where n is the number of observations 
(100 in this case).

best wishes,
Kay

On Wed, 20 Oct 2021 13:08:33 +0100, Ian Tickle 
mailto:ianj...@gmail.com>> wrote:

Hi Kay

Can I just confirm that your result Iobs=0.01 sigIobs=0.01 is the estimate
of the true average intensity *per pixel* for a patch of 100 pixels?  So
then the total count for all 100 pixels is 1 with variance also 1, or in
general for k observed counts in the patch, expectation = variance = k+1
for the total count, irrespective of the number of pixels?  If so then that
agrees with my own conclusion.  It makes sense because Iobs=0.01
sigIobs=0.01 cannot come from a Poisson process (which obviously requires
expectation = variance = an integer), whereas the total count does come
from a Poisson process.

The difference from my approach is that you seem to have come at it via the
individual pixel counts whereas I came straight from the Agostini result
applied to the whole patch.  The number of pixels seems to me to be
irrelevant for the whole patch since the design of the detector, assuming
it's an ideal detector with DQE = 1 surely cannot change the photon flux
coming from the source: all ideal detectors whatever their pixel layout
must give the same result.  The number of pixels is then only relevant if
one needs to know the average intensity per pixel, i.e. the total and s.d.
divided by the number of pixels.  Note the pixels here need not even
correspond to the hardware pixels, they can be any arbitrary subdivision of
the detector surface.

Best wishes

-- Ian


On Tue, 19 Oct 2021 at 12:39, Kay Diederichs 
mailto:kay.diederi...@uni-konstanz.de>>
wrote:

James,

I am saying that my answer to "what is the expectation and variance if I
observe a 10x10 patch of pixels with zero
counts?" is Iobs=0.01 sigIobs=0.01 (and Iobs=sigIobs=1 if there is only
one pixel) IF the uniform prior applies. I agree with Gergely and others
that this prior (with its high expectation value and variance) appears
unrealistic.

In your posting of Sat, 16 Oct 2021 12:00:30 -0700 you make a calculation
of Ppix that appears like a more suitable expectation value of a prior to
me. A suitable prior might then be 1/Ppix * e^(-l/Ppix) (Agostini §7.7.1).
The Bayesian argument is IIUC that the prior plays a minor role if you do
repeated measurements of the same value, because you use the posterior of
the first measurement as the prior for the second, and so on. What this
means is that your Ppix must play the role of a scale factor if you
consider the 100-pixel experiment.
However, for the 1-pixel experiment, having a more suitable prior should
be more important.

best,
Kay




On Mon, 18 Oct 2021 12:40:45 -0700, James Holton 
mailto:jmhol...@lbl.gov>> wrote:

Thank you very much for this Kay!

So, to summarize, you are saying the answer to my question "what is the
expectation and variance if I observe a 10x10 patch of pixels with zero
counts?" is:
Iobs = 0.01
sigIobs = 0.01 (defining sigIobs = sqrt(variance(Iobs)))

And for the one-pixel case:
Iobs = 1
sigIobs = 1

but in both cases the distribution is NOT Gaussian, but rather
exponential. And that means adding variances may not be the way to
propagate error.

Is that right?

-James Holton
MAD Scientist



On 10/18/2021 7:00 AM, Kay Diederichs wrote:
Hi James,

I'm a bit behind ...

My answer about the basic question ("a patch of 100 pixels each with
zero counts - what is the variance?") you ask is the following:

1) we all know the Poisson PDF (Probability Distribution Function)
P(k|l) = l^k*e^(-l)/k!  (where k stands for for an integer >=0 and l is
lambda) which tells us the probability of observing k counts if we know l.
The PDF is normalized: SUM_over_k (P(k|l)) is 1 when k=0...infinity is 1.
2) you don't know before the experiment what l is, and you assume it is
some number x with 0<=x<=xmax (the xmax limit can be calculated by looking
at the physics of the experiment; it is finite and less than the overload
value of the pixel, otherwise you should do a different experiment). Since

Re: [ccp4bb] [phenixbb] NORA workshop on AlphaFold2 and RoseTTAFold, 31 August and 1 September

[Randy John Read]

Hi,

I’ve just double-checked with the organisers. We’ve been asked to agree to 
having our talks recorded. Some speakers might opt out, but the ones that are 
recorded will be available for participants to view at a more convenient time.

Randy

> On 26 Aug 2021, at 22:22, Joel Tyndall  wrote:
> 
> Hi,
> Any chance this will be recorded for those of us in a different dimension?
> 
> J
> 
> -Original Message-
> From: phenixbb-boun...@phenix-online.org  
> On Behalf Of Randy John Read
> Sent: Friday, 27 August 2021 6:27 am
> To: CCP4BB@jiscmail.ac.uk; PHENIX user mailing list 
> 
> Subject: [phenixbb] NORA workshop on AlphaFold2 and RoseTTAFold, 31 August 
> and 1 September
> 
> Hi,
> 
> I've just learned that this online workshop is generally open, not just to 
> members of the Norwegian Artificial Intelligence Consortium (NORA), and it's 
> free!  Information, including the current programme and a link to 
> registration, can be found here: 
> 
> https://apc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.nora.ai%2Fevents%2FAlphaFoldv2.0-and-RoseTTAFold-workshop%2520.htmldata=04%7C01%7Cjoel.tyndall%40otago.ac.nz%7Cd376abedf6994ecdf93e08d968bfc1af%7C0225efc578fe4928b1579ef24809e9ba%7C1%7C0%7C637655995083551647%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000sdata=Bn4T%2FOgLlA1Qher%2FUwOK5HTkJMLS264KWKR%2FtB7c71Q%3Dreserved=0
> 
> I'll be talking a bit about experiences with models from AlphaFold2 and 
> RoseTTAFold, but the real draw should be talks by people who really 
> understand these two deep learning algorithms: Minkyung Baek (first author on 
> the RoseTTAFold paper, from David Baker's group) and a representative of the 
> DeepMind team that developed the two versions of AlphaFold. Sameer Velankar 
> will also be there to talk about the exciting new AlphaFold database at the 
> EBI.
> 
> 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
> 
> 
> ___
> phenixbb mailing list
> pheni...@phenix-online.org
> https://apc01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fphenix-online.org%2Fmailman%2Flistinfo%2Fphenixbbdata=04%7C01%7Cjoel.tyndall%40otago.ac.nz%7Cd376abedf6994ecdf93e08d968bfc1af%7C0225efc578fe4928b1579ef24809e9ba%7C1%7C0%7C637655995083551647%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000sdata=CByF3Vkz5C9DbG3UqNGU0H7Uywt68vVGYOexnwPTS3w%3Dreserved=0
> Unsubscribe: phenixbb-le...@phenix-online.org

-
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] [phenixbb] NORA workshop on AlphaFold2 and RoseTTAFold, 31 August and 1 September

[Randy John Read]

9 am CEST (Norway). Pretty early for our North American colleagues, especially 
those on the west coast!

> On 27 Aug 2021, at 05:18, Robert Rose  wrote:
> 
> Does the workshop start on August 31 at 9 am GMT?
> 
> On Thu, Aug 26, 2021 at 2:27 PM Randy John Read  wrote:
> Hi,
> 
> I’ve just learned that this online workshop is generally open, not just to 
> members of the Norwegian Artificial Intelligence Consortium (NORA), and it’s 
> free!  Information, including the current programme and a link to 
> registration, can be found here: 
> 
> https://www.nora.ai/events/AlphaFoldv2.0-and-RoseTTAFold-workshop%20.html
> 
> I’ll be talking a bit about experiences with models from AlphaFold2 and 
> RoseTTAFold, but the real draw should be talks by people who really 
> understand these two deep learning algorithms: Minkyung Baek (first author on 
> the RoseTTAFold paper, from David Baker’s group) and a representative of the 
> DeepMind team that developed the two versions of AlphaFold. Sameer Velankar 
> will also be there to talk about the exciting new AlphaFold database at the 
> EBI.
> 
> 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
> 
> 
> ___
> phenixbb mailing list
> pheni...@phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
> Unsubscribe: phenixbb-le...@phenix-online.org

-
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] NORA workshop on AlphaFold2 and RoseTTAFold, 31 August and 1 September

[Randy John Read]

Hi,

I’ve just learned that this online workshop is generally open, not just to 
members of the Norwegian Artificial Intelligence Consortium (NORA), and it’s 
free!  Information, including the current programme and a link to registration, 
can be found here: 

https://www.nora.ai/events/AlphaFoldv2.0-and-RoseTTAFold-workshop%20.html

I’ll be talking a bit about experiences with models from AlphaFold2 and 
RoseTTAFold, but the real draw should be talks by people who really understand 
these two deep learning algorithms: Minkyung Baek (first author on the 
RoseTTAFold paper, from David Baker’s group) and a representative of the 
DeepMind team that developed the two versions of AlphaFold. Sameer Velankar 
will also be there to talk about the exciting new AlphaFold database at the EBI.

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




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] biomolecular NMR for IDPs

[Randy John Read]

What the DeepMind people say about this is that when part of a predicted 
structure has a low predicted lDDT (a CASP measure of model quality), this is a 
good indicator that it may be intrinsically disordered. 

Best wishes,

Randy Read

> On 17 Aug 2021, at 17:13, Sorin Draga  wrote:
> 
> Dear all, 
> 
> Regarding the AlphaFold2: while its results are impressive for properly 
> folded proteins, I would caution against using it for IDPs: they do not have 
> a stable 3D structure that can be predicted,  instead they occupy an ensemble 
> of structures, with a high degree of heterogeneity. While you will find 
> models of IDPs predicted by AlphaFold, they only represent (at best) one of 
> many local minima that these proteins occupy.
> 
> Again, many thanks for all your excellent contributions! 
> 
> Kind regards, 
> 
> Sorin 
> 
> On Tue, Aug 17, 2021 at 11:59 AM George Sheldrick  
> wrote:
> 
> 
> 
> 
> 
> As Joel has suggested before, alphafold on an IDP would be interesting and 
> would seem like a zero-cost starting point - perhaps one you have tried 
> already.
> 
> 
> 
> Sent from ProtonMail mobile
> 
> 
> 
>  Original Message 
> On 15 Aug 2021, 15:53, Scott Horowitz < scott.horow...@du.edu> wrote:
> 
> Hi Sorin,
> 
> I hate to say it, but this is a really tough and expensive one. Solving a 
> true conformational ensemble of one IDP of decent size (~>70 residues) at 
> something like decent resolution is hard, and not that many labs actually do 
> it (it's usually a different set of NMR techniques than solving folding 
> proteins, and that knowledge is even somewhat specialized even within the NMR 
> community). Solving a co-structural ensemble of two IDPs that bind is even 
> harder, and I'm hard pressed to remember a single case right now where it's 
> been done (probably has, but very rarely). Assuming they express really well 
> and produce decent spectra, it is in theory doable, but I'd assume multiple 
> years of work by a very good student or postdoc from a lab that specializes 
> in this and many thousands of dollars (I'd very roughly assume ~$10k in 
> materials costs alone) would be required for that co-structure.
> 
> The SAXS route is certainly less expensive and faster if it works and gets 
> you the info you need, but it certainly will be low-res. I'm not as familiar 
> with it, but if you can differentially label the proteins, the neutron 
> equivalent of SAXS might also help with the co-structural ensemble to 
> differentiate which protein is where in the resulting blob.
> 
> Scott
> 
> Scott Horowitz, Ph.D.
> Assistant Professor
> Department of Chemistry & Biochemistry
> Knoebel Institute for Healthy Aging
> University of Denver
>  
> ECS Building
> 2155 E. Wesley Ave
> Denver, CO 80208
> Phone: 303-871-4326
> Fax: 303-871-7915
> Zoom Room: https://udenver.zoom.us/my/scotthorowitz
> Email: scott.horow...@du.edu
> Office: Room 561   Lab: Room 505
> 
> From: CCP4 bulletin board  on behalf of Roopa Thapar 
> <070a21fba45f-dmarc-requ...@jiscmail.ac.uk>
> Sent: Sunday, August 15, 2021 8:20 AM
> To: CCP4BB@JISCMAIL.AC.UK 
> Subject: [EXTERNAL] Re: [ccp4bb] biomolecular NMR for IDPs
>  
> [External Email From]: owner-ccp...@jiscmail.ac.uk
> 
> 
> Hello Sorin,
> 
> 
> 1. The cost of getting NMR data on the IDPs you propose depends upon the 
> expression levels of the protein/s as you will need to label with 15N and 13C 
> - and depending upon your overall yields per liter of E.coli culture, this 
> can add up.  In addition you will need to run triple resonance experiments - 
> so you should look into the hourly charge to access the NMR spectrometers 
> where you are located.  Moreover, you need to account for time required for 
> optimization of solution conditions to collect the NMR data as the sample 
> needs to be homogenous (as in no aggregation) at millimolar or hundreds of 
> micromolar concentration.   As Ethan Merritt suggested, it would be a good 
> idea to use SAXS first as it requires very little sample, no isotope 
> labeling, and you can try to narrow down the solution conditions that would 
> be best suited for NMR.  The Kratky plots, Rg values under different solution 
> conditions can give very useful information about conformational states and 
> ensembles populated by IDPs.  However, although NMR tends to be more 
> expensive than other techniques but is perfect for IDPs as you point out you 
> can get residue specific information.  A combined NMR/SAXS approach has 
> proven to be very useful to validate computational models.
> 
> 2. In general, CROs are much more expensive particularly for generating 
> isotopically labeled samples - it is cost-prohibitive for academic labs.  
> Genscript is one CRO that will express proteins, but I am not sure if they 
> will make isotopically labeled proteins for NMR.
> 
> 3. The amount of protein needed depends upon the size of the molecule.  You 
> will need at least 2-3 samples that are differentially labeled 

Re: [ccp4bb] Arcimboldo

[Randy John Read]

Yes, I should have mentioned the truncation by estimated RMS error, which was 
part of the protocol we used in the paper with David Baker’s group.  The error 
estimates are calibrated, I believe, for just the C-alpha atoms, so they are 
probably generally underestimates of the errors in the rest of the atoms in a 
residue, particularly the side chains; otherwise, the atom weighting obtained 
by converting the error estimate into an equivalent B-factor would be enough to 
account for errors.

Randy

> On 3 Jul 2021, at 17:36, Adam Simpkin  wrote:
> 
> I agree that RoseTTAFold would be a good idea! We've been looking at the 
> models for MR recently and in many cases they're excellent. We've found it 
> can be useful to truncate the models based on the predicted rms errors with 
> AMPLE (clearer solutions or even the difference between success & failure) so 
> this might be a useful strategy to consider if the models don't work in MR by 
> themselves.  
> 
> 
> 
> 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/

-
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] Arcimboldo

[Randy John Read]

If there isn’t already a suitable model, this looks like a good opportunity to 
try out the powerful new RoseTTAFold deep learning algorithm by Minkyung Baek 
in David Baker’s lab.  It can be used through the Robetta server 
(http://robetta.bakerlab.org, choose Structure Prediction->Submit and then 
choose the RoseTTAFold option).  It’s already been used to solve a number of 
unsolved structures, including the ones discussed in a paper on the algorithm 
(https://www.biorxiv.org/content/10.1101/2021.06.14.448402v1) and others I’ve 
heard of through the grapevine.

Best wishes,

Randy

> On 2 Jul 2021, at 12:52, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Yes - nothing very significant in the self rotation list of peaks..
> Is the resolution really 2.6A though - if so it will be very difficult to 
> find a single helix which would be quite a small fraction of 70x2 residues 
> Iwould think..
> Is there no suitable model?
> Eleabor
> 
> On Fri, 2 Jul 2021 at 12:45, leo john  wrote:
> Thank you all.
> 
> I have tried both space groups. I am attaching the required file. 
> 
> Cheers
> John
> 
> On Fri, Jul 2, 2021 at 12:40 PM Eleanor Dodson  
> wrote:
> I presume you have tested both P41 and P43 space groups?
> It is a bit hard to check peak heights in the self rotation. Could you attach 
> the molrep output - I think it is called xxxmolrep.doc.txt in the 
> job directory..
> Eleanor
> 
> 
> On Fri, 2 Jul 2021 at 12:33, leo john  wrote:
> Hi Group;
> 
> I have collected data at approximately 2 Ang and automated softwares at 
> Diamond have processed it in spacegroup P41.
> My peptide is 70 residues long with 3 helices of minimum 17 residues in it. 
> According to Mat. Coeff. there are 2 molecules. I am also attaching the 
> self-rotation function output map (please give some opinion on it also i.e 
> number of molecules, symmetry).
> 
> Since the molecule has mainly alpha-helices, I am trying to use  Arcimboldo 
> but with no luck. 
> I have used both the coiled-coil mode and normal mode for it and maximum 
> value of CC after 5 cycles I get is closed to 20 that is not great for a 
> solution. Even upon looking the structure solution by Arcimboldo, I found a 
> lot of loops and I do not get any density for sidechains. A few of the parts 
> of the helices and loops are overlapping. 
> 
> I have used CCPi2 with the following options:
> The assymmetric unit contains 1 molecule of mol wt 7300
> search for 3 helices of length 14 residues
> 
> Any help will be appreciated.
> 
> 
> Thank you
> John
> 
> 
> 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
> 

-
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] TWIN?

[Randy John Read]

I agree with Kay that, with a good model, solving in P1 is likely to be the 
easiest comprehensive solution.  If that doesn’t work, Phaser starts every 
MR_AUTO job by making a list of all the subgroups, including the different 
potential indexings (represented by Hall symbols), so you could also work 
through those systematically.  You provide a “SPACEGROUP HALL” command with one 
of the Hall symbols for each possibility, and Phaser will expand the data from 
the higher symmetry and reindex as required.

Best wishes,

Randy Read

> On 9 Jun 2021, at 09:37, Kay Diederichs  
> wrote:
> 
> Hi Almudena,
> 
> if it is a packing problem, you need to find the correct subgroup of P6522 
> (179).
> Take a look at 
> https://strucbio.biologie.uni-konstanz.de/xdswiki/index.php/Space_group_determination#Subgroup_and_supergroup_relations_of_these_space_groups
>  .
> Subgroups of 179 are 20, 153, 154, 170, and recursively each of these has 
> subgroups:
> 20 has  4 and 5 as subgroups, 153 and 154 both have 5 and 145 as subgroups, 
> 170 has  4 and 145 as subgroups - taken together, 4, 5 and 145.
> These have 1 as subgroup.
> So the true space group could be P1, P21, C2, P32, C2221, P3212, P3221, P65.
> You could run molecular replacement with all of these. Unfortunately, there 
> may be several ways to index the data in some of these space groups, and they 
> may not be equivalent. For example, I think there are 3 non-equivalent ways 
> to index in C2 or P21 if the Laue class is 6/m .
> 
> If there is twinning, the intensity statistics should tell about that - but 
> they may be set off by tNCS. 
> 
> One way to overcome the mess of possible space groups and settings plus the 
> twinning possibility is to index and solve the structure in P1. That should 
> allow a packing without clashes, and one could identify the correct space 
> group by running POINTLESS on the Fcalc, and/or Zanuda. Since you have a good 
> model, I'd try that.
> 
> Hope this helps,
> Kay
> 
> On Tue, 8 Jun 2021 17:14:30 +0200, Almudena Ponce Salvatierra 
>  wrote:
> 
>> Hello everyone,
>> 
>> I am working with an RNA-only structure, data are at 3 Angstroms, and at
>> first, I thought was in the C2 space group (with 6 molecules in the AU).
>> 
>> I can't finish building! it, because the structure seems to get in the way
>> of its neighbor symmetrically! See the attached picture, please! The only 4
>> residues that the structure is missing "have to go there", where one
>> structure meets the other one. The R factors for this spacegroup are around
>> 0.3.
>> 
>> However, Phenix Xtriage suggests the symmetry may be higher. So I reindex
>> in P622, do MR with Phaser (trying all possible space groups in that point
>> group), and it finds a unique solution in space group P 65 2 2 with TFZ of
>> 50 and LLG >3000. Of course, the problem persists (one molecule sort of
>> interfering with the neighbor), not only that but also the refinement
>> R-factors are substantially higher, 0.37.
>> 
>> I have to say that the refinement maps look better when I am working with
>> the C2 spacegroup. I can't understand, though, what is happening at that
>> "supposedly interface" between the two molecules. Has anybody experienced
>> anything like this in the past? Can it be a twin? Is it something else?
>> and... how to fix it?
>> 
>> Thank you very much in advance.
>> 
>> Best wishes,
>> 
>> Almudena
>> 
>> 
>> 
>> 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/

-
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 

Re: [ccp4bb] [ccp4bb] AW: [ccp4bb] (R)MS

[Randy John Read]

It’s also important to keep in mind that, in this equation, u is the component 
of the displacement *in the direction of the diffraction vector*.  If you 
assume isotropic displacements and you know the mean-squared value of the 
overall xyz vector displacement, you have to divide that mean-squared value by 
3 to get the variance in any particular direction.  This is a source of 
considerable confusion in crystallography textbooks!

Best wishes,

Randy Read

> On 28 May 2021, at 11:09, Ian Tickle  wrote:
> 
> 
> Hi Jonathan
> 
> On Thu, 27 May 2021 at 18:34, Hughes, Jonathan 
>  wrote:
> 
>  "B = 8π2  where u is the r.m.s. displacement of a scattering center, and 
> <...> denotes time averaging"
> 
> Neither of those statements is necessarily correct: u is the _instantaneous_ 
> displacement which of course is constantly changing (on a timescale of the 
> order of femtoseconds) and cannot be measured.  So u2 is the squared 
> instantaneous displacement,   is the mean-squared displacement, and so 
> the root-mean-squared displacement (which of course is amenable to 
> measurement) is sqrt(), not the same thing at all as u.
> 
> Incidentally, the 8π2 constant factor comes from Fourier-transforming the 
> Debye-Waller factor expression I mentioned earlier.
> 
> Also for crystals at least, the averaging is not only over time, it's over 
> all unit cells, i.e. the displacements are not only thermal in origin but 
> also due to spatial static disorder (instantaneous differences between unit 
> cells).
> 
> it would seem to me that we would be able to interpret things MUCH more 
> easily with u rather than anything derived from u².
> 
> So then I think what you mean is sqrt() rather than , which seems not 
> unreasonable.
> 
> Cheers
> 
> -- Ian
> 
> 
> 
> 
>  
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] MR problem in muti-domain structure

[Randy John Read]

Brief followup:

1. Typo: Phaser struggled with the second copy of domain *2*, not domain 1.

2. I thought the single job searching for 1+1+2+3+3+2 would work, but it failed 
at the last step.  The easiest way to solve this with Phaser might be to do one 
job with 1+1+2+3+3, and then do a second job adding the second copy of domain 
2.  The reason for this, I think, is that Phaser refines all of the placed 
models simultaneously, including their relative B factors and estimated RMS 
errors, at the end of the whole job, not after placing every model.  After each 
intermediate placement, Phaser only refines the parameters of the last 
component placed and not the estimated RMS error.  The improvement the full 
refinement gives in the 5-domain solution is just enough that the second copy 
of domain 2 can finally be found.

3. If you want to find this and some other tutorials, go to 
http://legacy.ccp4.ac.uk/tutorials/targets/standard/, expand the section on 
“Basic phasing tutorials” under “ccp4i tutorials”, and download phasing.tgz, 
which contains what you need to run 9 tutorials in this section (of which this 
is #7).

Best wishes,

Randy Read

> On 11 May 2021, at 14:57, Randy John Read  wrote:
> 
> Hi,
> 
> I had a hard time tracking down the data in the legacy CCP4 site but 
> eventually found it!  Then I tried solving the MR tutorial case with Phaser.
> 
> Instead of using the tutorial domain 1, 2 and 3 files, I prepared them from 
> the deposited PDB file using a sequence alignment followed by using sculptor 
> to trim off non-conserved ends of side chains and any missing loops.  Then I 
> did a default run of Phaser, searching for 2 copies of each domain. Phaser 
> automatically set the search order to domain 1, then domain 2 then domain 3 
> (they have similar sequence identities and that sorts them from biggest to 
> smallest).  It was clear from the statistics (increase in log-likelihood gain 
> or LLG and TFZ score) that both copies of domain 1 were correct, the first 
> copy of domain 2, and both copies of domain 3, even though an incorrect 
> placement of domain 2 occurred before looking for domain 3.  Unfortunately, 
> the second copy of domain 3 was rejected because of packing problems with the 
> badly-placed domain 2 copy.
> 
> So I went back to the partial solution with 2 copies of domain 1 and one of 
> domain 2, then added both copies of domain 3 followed by the second copy of 
> domain 2.  This all worked well, giving clear solutions.  No intermediate 
> refinement necessary, though if it hadn’t worked it would have been possible 
> to refine the intermediate solution and grab the improved models for any 
> missing domains.  Fifty cycles of jelly-body refinement of the final MR 
> solution in Refmac gave an R-free of 43.7%.
> 
> Note that in these jobs Phaser struggled with the second copy of domain 1, 
> only pulling it out in the refinement step, probably because the B-factor had 
> to be increased substantially to agree with the data.  Another search order 
> might have been more efficient (maybe 1, 2, 3, 1 2, 3), but this worked.  
> Note that you can over-ride the automatic choice of search order, but Phaser 
> will by default choose its own search order (which is usually but not always 
> better).
> 
> Hope that helps on the Phaser front!
> 
> Best wishes
> 
> Randy Read
> 
>> On 11 May 2021, at 02:53, Lande  wrote:
>> 
>> Hi everyone, I started my work on data processing in protein crystallography 
>> about half year ago and have done a fair number of “easy” structures. 
>> Recently, I processed a tutorial structure containing 3 different domains 
>> and two chains of them (in particular, it’s “ mth685, high resolution data” 
>> from ccp4 legacy tutorial), but it was failed during MR process.
>> 
>> According to this tutorial, I did MR with MOLREP by adding each domain in 
>> order. The first and second added domains, which are domain 1 from chain A 
>> and domain 2 from chain B, are correct; however, when I tried to add the 
>> next domain, R work/R free did not drop and clearly MOLREP did not find the 
>> correct solution according to the density map. I also tried Phaser but the 
>> result was the same.
>> 
>> screen shot for domain 2 in chain A: https://ibb.co/YB2JnRR
>> 
>> https://ibb.co/YB2JnRR;>> src="https://i.ibb.co/0yq14DD/screenshot-on-wrong-domain.png; 
>> alt="screenshot-on-wrong-domain" border="0">> href='https://imgbb.com/'>pio pio 3 menu
>> 
>> In addition, I found an article on that data 
>> (https://journals.iucr.org/d/issues/2008/01/00/ba5117/index.html) but I 
>> failed to get the same result expect the first two domains.
>> 
>> I wonder if anyone 

Re: [ccp4bb] MR problem in muti-domain structure

[Randy John Read]

Hi,

I had a hard time tracking down the data in the legacy CCP4 site but eventually 
found it!  Then I tried solving the MR tutorial case with Phaser.

Instead of using the tutorial domain 1, 2 and 3 files, I prepared them from the 
deposited PDB file using a sequence alignment followed by using sculptor to 
trim off non-conserved ends of side chains and any missing loops.  Then I did a 
default run of Phaser, searching for 2 copies of each domain. Phaser 
automatically set the search order to domain 1, then domain 2 then domain 3 
(they have similar sequence identities and that sorts them from biggest to 
smallest).  It was clear from the statistics (increase in log-likelihood gain 
or LLG and TFZ score) that both copies of domain 1 were correct, the first copy 
of domain 2, and both copies of domain 3, even though an incorrect placement of 
domain 2 occurred before looking for domain 3.  Unfortunately, the second copy 
of domain 3 was rejected because of packing problems with the badly-placed 
domain 2 copy.

So I went back to the partial solution with 2 copies of domain 1 and one of 
domain 2, then added both copies of domain 3 followed by the second copy of 
domain 2.  This all worked well, giving clear solutions.  No intermediate 
refinement necessary, though if it hadn’t worked it would have been possible to 
refine the intermediate solution and grab the improved models for any missing 
domains.  Fifty cycles of jelly-body refinement of the final MR solution in 
Refmac gave an R-free of 43.7%.

Note that in these jobs Phaser struggled with the second copy of domain 1, only 
pulling it out in the refinement step, probably because the B-factor had to be 
increased substantially to agree with the data.  Another search order might 
have been more efficient (maybe 1, 2, 3, 1 2, 3), but this worked.  Note that 
you can over-ride the automatic choice of search order, but Phaser will by 
default choose its own search order (which is usually but not always better).

Hope that helps on the Phaser front!

Best wishes

Randy Read

> On 11 May 2021, at 02:53, Lande  wrote:
> 
> Hi everyone, I started my work on data processing in protein crystallography 
> about half year ago and have done a fair number of “easy” structures. 
> Recently, I processed a tutorial structure containing 3 different domains and 
> two chains of them (in particular, it’s “ mth685, high resolution data” from 
> ccp4 legacy tutorial), but it was failed during MR process.
> 
> According to this tutorial, I did MR with MOLREP by adding each domain in 
> order. The first and second added domains, which are domain 1 from chain A 
> and domain 2 from chain B, are correct; however, when I tried to add the next 
> domain, R work/R free did not drop and clearly MOLREP did not find the 
> correct solution according to the density map. I also tried Phaser but the 
> result was the same.
> 
> screen shot for domain 2 in chain A: https://ibb.co/YB2JnRR
> 
> https://ibb.co/YB2JnRR;> src="https://i.ibb.co/0yq14DD/screenshot-on-wrong-domain.png; 
> alt="screenshot-on-wrong-domain" border="0"> href='https://imgbb.com/'>pio pio 3 menu
> 
> In addition, I found an article on that data 
> (https://journals.iucr.org/d/issues/2008/01/00/ba5117/index.html) but I 
> failed to get the same result expect the first two domains.
> 
> I wonder if anyone has done this tutorial before or has faced similar 
> problems in his own work. I have no idea on how to deal with it. Any 
> suggestions or comments will be helpful. Thanks in ahead.
> 
> Link for that tutorial: 
> ftp://ftp.ccp4.ac.uk/tutorials/ccp4_phasing/7_domains/mth685.pdf
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] What the best strategy for density fitting of antibody

[Randy John Read]

Hi Jessica,

The suggestion by Eleanor and Thierry to break the model into two parts for 
molecular replacement should work.  However, if it doesn’t there’s another 
option, which can work well in marginal cases.  Fix the domain that’s 
well-placed.  Make a note of the Euler angles from the MR solution that placed 
this domain correctly.  Now run a second Phaser job, but this time choose the 
brute-force rotation function option, and instead of searching all possible 
orientations just take orientations within, say, 10 degrees of the same set of 
Euler angles (which will give you similar relative orientations).  There won’t 
be that many orientations, so to avoid missing the best choice turn off 
rotation clustering and select all of those angles for use in the next step, 
which is to run the translation searches.

If you were doing this with keyword scripts, the relevant keywords would be

   TARGET ROT BRUTE
   ROTATE VOLUME AROUND EULER 10. 20. 30 RANGE 10.
   PEAKS ROT CLUSTER OFF
   PEAKS ROT SELECT ALL

(at least if the Euler angles for the first domain were 10, 20, 30).  However, 
it shouldn’t be that hard to figure out how to set the same options in the 
ccp4i or ccp4i2 GUIs.

Best wishes,

Randy Read

> On 24 Mar 2021, at 21:31, Jessica Besaw  wrote:
> 
> Hi friends!
> 
> I need advice about the best (or easiest) strategy to refine a piece of an 
> antibody. 
> 
> As seen in the figure below, the right part of the protein fits the density 
> very well and the left part does not (likely the result of the molecular 
> replacement model). I can clearly see that the entirety of left domain needs 
> to be rotated/translated quite a significant amount at a proline residue 
> (shown as sticks in the middle of the figure). 
> 
> What is the best strategy to do this in phenix.refine / coot? Tutorial 
> suggestions/links would be very helpful. 
> 
> 
> 
> 
> Cheers!
> 
> Jessica Besaw
> 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] Challenging Molecular Replacement

[Randy John Read]

Yes, I agree with Eleanor.  But if there’s translational NCS, either the 
off-origin Patterson peak is below 20% of the origin peak so that Phaser is 
missing it, or perhaps you’re turning off the tNCS correction?  If Phaser 
detects tNCS it will place two copies at once.

We’ve had cases where, as Eleanor said, it’s hard to tell the difference 
between a crystallographic 2(1) with a non-crystallographic 2-fold or the 
reverse, so you have to try all possible choices of space group.

Best wishes,

Randy Read

> On 16 Feb 2021, at 18:31, Eleanor Dodson 
> <176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Well - your LLG etc looks good but those two solutions have symmetry 
> equivalent rotation angles so must be related by a translation vector. 
> 
>SOLU 6DIM ENSE ense_1 EULER  269.0   80.4  177.3 FRAC  0.24 -0.22 -0.07 
> BFAC -1.51 #TFZ==6.5
>SOLU 6DIM ENSE ense_1 EULER   89.0   80.4  177.3 FRAC  0.26 -1.01 -0.07 
> BFAC  2.26 #TFZ==41.8
> Look at your data processing - if you use I2 there will be warnings.
> And with a translation vector you often have to consider alternate 
> spacegroups.
> 
> If you want to send more information I might be able to be of more help.
> Eleanor
> 
> On Tue, 16 Feb 2021 at 18:21, Muhammad Bashir Khan  wrote:
> Dear All;
> 
> I have data set of about 3.0A. It's a multidomain protein. I tried several 
> options with MR but it's was not working. I trimmed one of the search models, 
> it does not give any solution using CCP4 molrep. 
> 
> I used Phenix phaser it gives a solution after several hours with the 
> following output.
> 
> ** SINGLE solution
> 
> ** Solution written to PDB file:  MgtA_phaser.1.pdb
> ** Solution written to MTZ file:  MgtA_phaser.1.mtz
>Solution annotation (history):
>SOLU SET  RFZ=2.7 TFZ=6.3 PAK=6 LLG=62 TFZ==6.5 RFZ=2.2 TFZ=28.0 PAK=9 
> LLG=1207 TFZ==42.0 LLG=1370 TFZ==41.8 PAK=9
> LLG=1370 TFZ==41.8
>SOLU SPAC P 21 2 21
>SOLU 6DIM ENSE ense_1 EULER  269.0   80.4  177.3 FRAC  0.24 -0.22 -0.07 
> BFAC -1.51 #TFZ==6.5
>SOLU 6DIM ENSE ense_1 EULER   89.0   80.4  177.3 FRAC  0.26 -1.01 -0.07 
> BFAC  2.26 #TFZ==41.8
>SOLU ENSEMBLE ense_1 VRMS DELTA -0.7291 #RMSD  1.50 #VRMS  1.23
> 
> Running Phenix refine on the solution gives a breakup map and the R factors 
> are not decreasing below 47.61 and 50.94.
> 
> Thank you for any suggestions in advance
> 
> Regards;
> 
> Bashir
> 
> 
> -- 
> --
> Muhammad Bashir Khan, Ph.D.
> Research Associate
> Department of Biochemistry
> Medical Science Bldg.
> Lab 3-27
> University of Alberta
> Edmonton AB, T6G 2H7
> 
> Phone: 780-492-4577-
> e-mail: m...@ualberta.ca
> 
> 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
> 

-
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] TNCS and oligomeric state

[Randy John Read]

Dear John,

It’s hard to be absolutely certain from the reproduction, but it looks like you 
have equally high 2-fold axes all around the xy plane in the self-rotation 
function.  Do you have an explanation for that?

It would be helpful to know the heights of the Patterson peaks relative to the 
origin peak.  However, assuming that the ones you have listed are all 
relatively high, this can be interpreted in terms of a commensurate modulation 
or, in alternative terminology, a multiple tNCS vector.  All of the peaks are 
approximate multiples of the same translation vector, i.e. 1/2, 1/2, 1/6.

1/2, 1/2, 1/6
2/2, 2/2, 2/6, equivalent to 0, 0, 1/3
3/2, 3/2, 3/6, equivalent to 1/2, 1/2, 1/2

The next three in this series would be the same vectors in the opposite 
direction, i.e. related by the inversion operator in the Patterson.  If the 
Matthews coefficient suggests 2 copies, you could probably have 3 in the 
asymmetric unit related by multiples of the same vector, but not 6.

So in Phaser you would express this as
TNCS NMOL 3
TNCS TRA VECTOR 0.5 0.5 0.169

Best wishes,

Randy Read

> On 3 Feb 2021, at 11:34, leo john  wrote:
> 
> Sticking to the same doubt:
> 
> Just to check the correct symmetry and space group, I have now processed my 
> data in P1 and I4 followed by running MOLREP for self-rotation function 
> (Figure attached).
> 
> Can we conclude now on the correct space group and oligomeric state. 
> 
> Thank You all
> John
> 
> On Wed, Feb 3, 2021 at 8:34 AM leo john  wrote:
> Dear All:
> 
> I have a peptide that is crystallized in the space group I4122 with cell 
> dimensions 40,40, 200 Ang. 
> Mathews Coefficient suggests 2 molecules/ ASU. 
> Since I do not have a starting model for MR, I went for experimental phasing. 
> However, unsuccessful so far. 
> 
> Details about my dataset:
> 
> 1) It has been collected at 2 Ang and has an Anomalous signal till 2.7
> 2) It has got Translational NCS.
> 3) I checked the self-rotation function using MOLREP (figure attached), but 
> not able to make anything out of it. Please suggest.
> 4) I got the translational and rotational peaks from the log file of MOLREP
> 
> Peak 1: trans.vector /ort/ :19.27019.27098.550
>trans.vector /frac/: 0.500 0.500 0.500
>Peak 2: trans.vector /ort/ :19.27019.27033.326
>trans.vector /frac/: 0.500 0.500 0.169
>Peak 3: trans.vector /ort/ :-0.000-0.00065.224
>trans.vector /frac/: 0.000 0.000 0.331
> 
> Sol_Rf   1 0.000.000.000.000.000.000.4448E+06  
> 4.53
> Sol_Rf   289.99  112.50  180.000.00  180.00  -45.000.3561E+06  
> 3.63
> Sol_Rf   3   109.566.31  180.006.31 -140.88  173.690.4912E+05  
> 0.50
> 
> These are unique peaks.
> 
> My questions are:
> 
> 1) How to read these MOLREP records and the plots to conclude something. How 
> can I guess the oligomeric state using the plot. Any link or suggestion will 
> be really helpful.
> 2) How to specify the TNCS during phaser run. 
> 
> Thank You all
> John
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] possible solution with Phaser

[Randy John Read]

Dear Luca,

I think all of those orientations are symmetry-related copies of the same 
orientation.  This would imply that there is tNCS, but it doesn’t look like 
Phaser is applying tNCS.  The second and fourth molecules are in the same 
orientation, but differing by 1/6 of the a-axis cell translation.  Is there a 
large off-origin native Patterson peak, perhaps at 1/6,0,0?

Also, when there is tNCS it can be difficult to disentangle crystallographic 
and non-crystallographic symmetry.  Did you try all the space groups in the 
same point group?

Marc makes a good point about the number of copies.  The Matthews volume only 
gives you a rough estimate of the true number of copies, and the signal in 
placing the fourth molecule is very unconvincing compared to the first three.

Best wishes,

Randy Read

> On 6 Jan 2021, at 07:38, Luca Mazzei  wrote:
> 
> Hi all,
> 
> I am struggling with the MR of a homo-dimer using Phaser. Matt_coeff strongly 
> suggests the presence of 4 mol per asym unit (space group P6222). The results 
> after a search of 4 mol per asymmetric unit of my monomer are the following:
> 
> ** SINGLE solution
> 
> ** Solution written to SOL file:  phaser_3f6v_A1_MOLREP.sol
> 
> ** Solution written to PDB file:  phaser_3f6v_A1_MOLREP.1.pdb
> ** Solution written to MTZ file:  phaser_3f6v_A1_MOLREP.1.mtz
>Solution annotation (history):
>SOLU SET  RFZ=3.9 TFZ=8.5 PAK=3 LLG=65 TFZ==10.0 RFZ=2.6 TFZ=17.0 PAK=3 
> LLG=326 TFZ==29.8 (& TFZ==22.5 & TFZ==19.7)
> LLG+=(326 & 529 & 593) LLG=795 TFZ==5.4 PAK=5 LLG=795 TFZ==5.4 PAK=5 
> LLG=795 TFZ==5.4
>SOLU SPAC P 62 2 2
>SOLU 6DIM ENSE autoMR EULER  125.3   60.8  300.2 FRAC  0.17 -0.13  0.08 
> BFAC -9.34 #TFZ==10.0
>SOLU 6DIM ENSE autoMR EULER  305.4   60.8  300.2 FRAC  0.27 -0.16  0.08 
> BFAC -5.29 #TFZ==29.8
>SOLU 6DIM ENSE autoMR EULER  294.7  119.1  120.3 FRAC  0.45  0.01  0.26 
> BFAC  0.34 #TFZ==22.5
>SOLU 6DIM ENSE autoMR EULER  305.5   61.6  300.1 FRAC  0.11 -0.16  0.08 
> BFAC 29.19 #TFZ==5.4
>SOLU ENSEMBLE autoMR VRMS DELTA -0.2463 #RMSD  0.93 #VRMS  0.79
> 
> It seems (also looking at the maps) that it correctly places three monomers 
> out of four. How can I use this information to improve the search of the 
> fourth monomer using the same template model?
> 
> Thanks in advance for your help,
> 
> Best regards
> 
> Luca Mazzei
> 
> Luca Mazzei - PhD
> Laboratory of Bioinorganic Chemistry
> Department of Pharmacy and Biotechnology (FaBiT)
> Alma Mater Studiorum - University of Bologna
> Viale Giuseppe Fanin, 40 - 40127, Bologna - Italy
> Tel: +39 0512096235
> 
> 
> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?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/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] Finding partial occupancy monomer by MR ?

[Randy John Read]

Hi Phil,

It’s always a difficult balance between looking hard in a difficult case and 
possibly finding an answer with a weak signal, and stopping when the search is 
really unlikely to succeed.  Phaser perhaps errs too much toward trying really 
hard.

There’s one really blunt tool in Phaser, the KILL TIME command that tells it to 
stop after a certain number of minutes have passed.  I don’t particularly like 
or recommend that option.

A more nuanced approach is to use the PURGE commands to control the amount of 
branching that takes place in the searches — this is where the CPU time can 
really build up.  What I like is to set something like PURGE RNP NUM 20, in 
which case only 20 possible partial solutions after refinement are preserved as 
starting points to look for the next copy.  You can choose the number depending 
on your case and how long each rotation search and each translation search 
takes.  There are PURGE commands to control things at the rotation and 
translation stages as well, but I usually find that the RNP one gives enough 
control.

Best wishes,

Randy

> On 10 Dec 2020, at 16:00, Phil Jeffrey  wrote:
> 
> Thanks for the suggestions.
> 
> The idea that it's related to a trigonal space group and twinning or pseudo 
> space group is an interesting one, but this is C2221 and the intensity stats 
> don't show twinning.  Twinned P21 -> C2221 doesn't solve the non-unit 
> occupancy in this case.  Since the other monomers are full-occupancy it can't 
> be 3 overlapping dimers so the phenomenon is rather unusual in my finite 
> experience.  (Also only one set of Se peaks for this 4th monomer).
> 
> I used Herman's suggestion of finding 3 monomers first (with very large 
> RFZ/TFZ/LLG since the monomers had been refined against the data) since 
> that's very fast.  And then Phaser took a long while to not find the 4th 
> monomer.  Once I figure out how to make modern versions of phaser to "fail 
> quickly" like the older versions I'll scan a range of homology% and see if 
> that changes anything.
> 
> Phil
> 
> 
> On 12/10/20 9:46 AM, Schreuder, Herman /DE wrote:
>> Dear Phil,
>> 0.32 is awfully close to 1/3, which brings a nice mathematical puzzle to my 
>> mind to see if the 1/3 occupancy is somehow related to the 3 fully occupied 
>> monomers... It may also be related to a (trigonal??) space group...
>> You probably have already tried it, but phaser has the option to give it 
>> already solved molecules and ask it to search for additional molecules. Here 
>> I would indeed lower the expected % homology significantly, to crudely 
>> compensate for the low occupancy. In contrast to the advice of Dale, I would 
>> play around with the % homology to find the value which works best.
>> My 2 cents,
>> Herman
>> -Ursprüngliche Nachricht-
>> Von: CCP4 bulletin board  Im Auftrag von Phil Jeffrey
>> Gesendet: Donnerstag, 10. Dezember 2020 14:49
>> An: CCP4BB@JISCMAIL.AC.UK
>> Betreff: [ccp4bb] Finding partial occupancy monomer by MR ?
>> Preamble:
>> I have an interesting crystal form with 3 monomers (~400aa) at full 
>> occupancy and apparently one at much reduced occupancy.  It was built 
>> recently from Se-SAD and was in moderately good condition: Rfree=32% for 
>> trimer, 2.6 Å.  In recent refinement cycles it became obvious that there was 
>> a 4th monomer in a region of weaker/choppy 2Fo-Fc and Fo-Fc density that 
>> corresponded to a "confusing" set of low-occupancy SeMet sites found by 
>> SHELXD and Phaser-EP.  The experimental map was bad in that region and was 
>> probably flattened during density modification anyway, in retrospect.
>> Question:
>> Phaser failed to find the 4th monomer after trivially finding the other
>> 3 with a recent version of the monomer.  I'm wondering if there's a way to 
>> indicate "this one is partial occupancy" to Phaser, or if there's a way to 
>> improve the odds of success beyond just lowering the expected % homology.  
>> Or if anyone has had success with other programs.  This is perhaps a rare 
>> edge case but I naively expected Phaser to work.
>> In the end I used the weak SeMet sites to locate the monomer and the 
>> occupancy appears to be around 0.32 in refinement.
>> Cheers,
>> Phil Jeffrey
>> Princeton
>> 
>> To unsubscribe from the CCP4BB list, click the following link:
>> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jiscmail.ac.uk%2Fcgi-bin%2FWA-JISC.exe%3FSUBED1%3DCCP4BB%26A%3D1data=04%7C01%7CHerman.Schreuder%40SANOFI.COM%7C5ee2ee2e87f04d853f0408d89d126615%7Caca3c8d6aa714e1aa10e03572fc58c0b%7C0%7C1%7C63743205007789%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000sdata=G28AUNQrAgQYblmaYBVnXESTXiekmWzfTLPMMX%2B%2BOgw%3Dreserved=0
>> This message was issued to members of 
>> 

Re: [ccp4bb] [ccp4bb] External: Re: [ccp4bb] AlphaFold: more thinking and less pipetting (?)

[Randy John Read]

Several people have mentioned lack of peer review as a reason to doubt the 
significance of the AlphaFold2 results.  There are different routes to peer 
review and, while the results have not been published in a peer review journal, 
I would have to say (as someone who has been an assessor for two CASPs, as well 
as having editorial responsibilities for a peer-reviewed journal), the peer 
review at CASP is much more rigorous than the peer review that most papers 
undergo.  The targets are selected from structures that have recently been 
solved but not published or disseminated, and even just tweeting a C-alpha 
trace is probably enough to get a target cancelled.  In some cases (as we’ve 
heard here) the people determining the structure are overly optimistic about 
when their structure solution will be finished, so even they may not know the 
structure at the time it is predicted.  The assessors are blinded to the 
identities of the predictors, and they carry out months of calculations and 
inspections of the models, computing ranking scores before they find out who 
made the predictions.  Most assessors try to bring something new to the 
assessment, because the criteria should get more stringent as the predictions 
get better, and they have new ideas of what to look for, but there’s always 
some overlap with “traditional” measures such as GDT-TS, GDT-HA (more stringent 
high-accuracy version of GDT) and lDDT.

Of course we’d all like to know the details of how AlphaFold2 works, and the 
DeepMind people could have been (and should be) much more forthcoming, but 
their results are real.  They didn’t have any way of cheating, being selective 
about what they reported, or gaming the system in any other way that the other 
groups couldn’t do.  (And yes, when we learned that DeepMind was behind the 
exceptionally good results two years ago at CASP13, we made the same half-jokes 
about whether Gmail had been in the database they were mining!)

Best wishes,

Randy Read

> On 9 Dec 2020, at 10:36, Hughes, Jonathan 
>  wrote:
> 
> i think the answer to all these doubts and questions is quite simple: the 
> AlphaFold2 people must make all details of their methods public (source code) 
> and, as would probably be necessary, open their system for inspection and use 
> by independent experts. isn't that what peer review and reproducibility are 
> all about? those rules date from the time before every tom, dick and 
> henriette could publicize anything they like inside their own zuckerberg 
> bubble. my opinion is that this is a virtual infectious disease that will 
> cause humanity far bigger problems than corona ever will – i just hope i'm 
> wrong!
> best
> jon
>  
> Von: CCP4 bulletin board  Im Auftrag von Mark J van 
> Raaij
> Gesendet: Mittwoch, 9. Dezember 2020 11:14
> An: CCP4BB@JISCMAIL.AC.UK
> Betreff: Re: [ccp4bb] External: Re: [ccp4bb] AlphaFold: more thinking and 
> less pipetting (?)
>  
> on the day the news came out, I did wonder if the AlphaFold2 team somehow had 
> access to all the preliminary PDB files sent around via Gmail (which belongs 
> to the same company), but more as a joke/conspirational thought.
> "our" target T1052, was also predicted very well by domains and as a monomer. 
> It will be interesting to see how well future iterations of the method can 
> assemble the complete protein chain and the complete protein chains into the 
> correct heteromer.
>  
> 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
> Section Editor Acta Crystallographica F
> https://journals.iucr.org/f/
> 
>  
> On 9 Dec 2020, at 10:37, Cedric Govaerts  wrote:
>  
> Dear All
>  
> After about 10 (!) years of (very) hard work we solved the structures of our 
> dearest membrane transporter.  Dataset at 2.9 And resolution, fairly 
> anisotropic, experimental phasing, and many long nights with Coot and 
> Buster to achieve model refinement. 
>  
> The experimental structure had a well defined ligand nicely coordinated but 
> also a lipid embedded inside the binding cavity (a complete surprise but 
> biologically relevant) and two detergent molecules well defined 
> (experimental/crystallisation artefact).
>  
> As our paper was accepted basically when CASP organisers were calling for 
> targets I offered my baby to the computing Gods. However we only provided the 
> sequence to CASP, no info regarding any ligand or lipid.
>  
> Less than a month after, the CASP team contacted us and send us the best 
> model.  In fact it was 2 half models as the transporter is a pseudo dimer, 
> with the N-lobe and C-lobe moving relative to each other during transport 
> cycle, thus divided as two domains in CASP.
>  
> The results were breathtaking. 0.7 And RSMD on one half, 0.6 on the other. 
> And yes, group 427 was the superpower (did not know at the time that it was 
> AlphaFold).
>  
> We had long discussions with 

Re: [ccp4bb] AlphaFold: more thinking and less pipetting (?)

[Randy John Read]

Hi Frank,

Yes, until CASP7 (back in 2006), I used to like saying that there are many more 
ways to make a homology model worse than the starting template than to make it 
better, and that homology modelling programs were very good at finding them!  
After seeing that at least some models (e.g. from Rosetta) were actually better 
in CASP7, I had to stop saying that!

It’s not just anecdotal.  Even in CASP7, most models were still worse for MR 
than the best template someone could have found.

Randy

> On 4 Dec 2020, at 12:22, Frank von Delft  wrote:
> 
> I guess that also means that AlphaFold has learnt the crystal-structure-ness 
> that older homology methods never achieved - which is why (anecdotally?) a 
> "better" homology model tended to give worse MR performance than the "worse" 
> template?
> 
> (Or something like that, I'm parrotting what I remember people (maybe Randy?) 
> saying long ago about the problems with homology models in MR.)
> 
> 
> On 04/12/2020 11:57, Adam Simpkin wrote:
>> I thought I might be able to add a little to this conversation as I 
>> performed some MR runs as part of the CASP14 High Accuracy analysis. There 
>> were 30 targets with reflection data. Of these, AlphaFold2 models could be 
>> used to directly solve 24 structures after converting
>> RMS error predictions to simulated B-factors to aid the MR 
>> (10.1002/prot.25800).
>> 
>> Some of the models did contain sufficient local errors to impede MR. 
>> However, we were able to obtain a further 3 solutions by using AMPLE to 
>> truncate the models based on the per-residue RMS error predictions provided. 
>> In fact, a moderate truncation in AMPLE improved the quality of the MR 
>> solution in ~78% that succeeded by removing the few incorrectly models loops 
>> (typically at lattice interfaces).
>> 
>> A final thing to note was that the 3 structures that didn’t work still 
>> provided high quality model predictions (GDT_TS of 69, 84 & 83). These 
>> targets all contained multiple chains in the ASU and one was fairly low 
>> resolution (>3 Angstroms). Overall though I think the take home is clear, 
>> these models are really good and when the method or something similar is 
>> more publicly available I think it will definitely simplify MR for 
>> troublesome targets.
>> 
>> Best wishes,
>> 
>> Adam
>> 
>> 
>> 
>> 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/

-
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] AlphaFold: more thinking and less pipetting (?)

[Randy John Read]

Hi James,

One really interesting (and to me surprising) aspect of how well AlphaFold2 
does is that it does really well without actually understanding chemistry and 
physics.  (John Jumper from DeepMind talked about choices of deep learning 
model types and how they affect the “inductive bias” that allows things like 
chemistry to be learned indirectly, but it’s not programmed in.)  The best 
example is the fantastic models they made of monomers from trimeric proteins.  
The monomers can be assembled into trimers that look very much like the real 
thing, but they really modelled just monomers — somehow the machine learning 
algorithm implicitly knows about the trimers from the existence of distant 
homologues in the PDB.  As Joana said, AlphaFold2 did very well even on targets 
with no identifiable homologues, but I suspect that targets like these trimers 
will still require the presence of homologues.  Anyway, the modelled monomer 
makes no sense as a monomer, and any sensible force field would much prefer 
something else that buries more surface area!

Following up on some other comments, AlphaFold2 is a pretty complete 
reinvention compared to the original AlphaFold from 2 years ago.  AlphaFold 
followed a two-step process, where probability distributions for distances were 
learned in the first step (similar to the co-evolution constraints inferred by 
algorithms like the ones from Marks & Sander), and then those distance 
distributions were used in a minimisation step to fold the protein.  If I 
recall, the first step used a convolutional deep neural network.  In 
AlphaFold2, it’s all done in one end-to-end process going from sequence (and 
multiple sequence alignments) to xyz coordinates.  The model type has changed 
to something called an attention module, which John Jumper said acts to 
implicitly and iteratively learn a graph representing atoms and their 
interactions.

Once this algorithm or others like it are available to the community, it is 
indeed going to change the focus of what we do as structural biologists, but 
importantly it’s going to allow us to do more and to focus more on the 
biological questions than the technology.  (How and when it will become 
available is not entirely clear: John Jumper mentioned “internal discussions” 
in DeepMind about how to share with the community, and said there would be more 
news on that next year.)

Best wishes,

Randy Read

> On 4 Dec 2020, at 01:34, James Holton  wrote:
> 
> It is a major leap forward for structure prediction for sure.  A hearty 
> congratulations to all those teams over all those years.
> 
> The part I don't understand is the accuracy.  If we understand what holds 
> molecules together so well, then why is it that when I refine an X-ray 
> structure and turn the X-ray weight term down to zero ... the molecule blows 
> up in my face?
> 
> -James Holton
> MAD Scientist
> 
> 
> On 12/3/2020 3:17 AM, Isabel Garcia-Saez wrote:
>> Dear all,
>> 
>> Just commenting that after the stunning performance of AlphaFold that uses 
>> AI from Google maybe some of us we could dedicate ourselves to the noble art 
>> of gardening, baking, doing Chinese Calligraphy, enjoying the clouds pass or 
>> everything together (just in case I have already prepared my subscription to 
>> Netflix).
>> 
>> https://www.nature.com/articles/d41586-020-03348-4
>> 
>> Well, I suppose that we still have the structures of complexes (at the 
>> moment). I am wondering how the labs will have access to this technology in 
>> the future (would it be for free coming from the company DeepMind - 
>> Google?). It seems that they have already published some code. Well, 
>> exciting times. 
>> 
>> Cheers,
>> 
>> Isabel
>> 
>> 
>> Isabel Garcia-Saez   PhD
>> Institut de Biologie Structurale
>> Viral Infection and Cancer Group (VIC)-Cell Division Team
>> 71, Avenue des Martyrs
>> CS 10090
>> 38044 Grenoble Cedex 9
>> France
>> Tel.: 00 33 (0) 457 42 86 15
>> e-mail: isabel.gar...@ibs.fr
>> FAX: 00 33 (0) 476 50 18 90
>> http://www.ibs.fr/
>> 
>> 
>> 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
> 

-
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