Re: [ccp4bb] To Trim or Not to To Trim

[James Holton]

I have never met a computational chemist who did not "notice" when a 
side chain is modeled as more than one conformer.


-James Holton
MAD Scientist

On 3/31/2023 8:31 AM, Olga Moroz wrote:

Hi everyone,

I always thought it is better to truncate so that biologists looking 
at the structures are not misled?

Not sure it the best aproach though...

Olga



On 10 Mar 2023, at 02:45, Bernhard Lechtenberg 
<968307750321-dmarc-requ...@jiscmail.ac.uk> wrote:


Hi Rhys,
I am also all for leaving side chains and letting the B-factors deal 
with the weak/absent density.
I don’t think there is a consensus, but I kind of remember that 
somebody did a poll a few years ago and if I remember correctly the 
main approaches were the one described above, or trimming the 
side-chains.

Bernhard
*Bernhard C. Lechtenberg*PhD
NHMRC Emerging Leadership Fellow
Laboratory Head
Ubiquitin Signalling Division​​
elechtenber...@wehi.edu.au <mailto:lechtenber...@wehi.edu.au>
T +61 3 9345 2217

*From:*CCP4 bulletin board  on behalf of Rhys 
Grinter <22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>

*Date:*Friday, 10 March 2023 at 12:26 pm
*To:*CCP4BB@JISCMAIL.AC.UK 
*Subject:*[ccp4bb] To Trim or Not to To Trim

Hi All,
I'm trying to crowdsource an opinion on how people deal with 
modelling side chains with poorly resolved electron or cryoEM density.
My preference is to model the sidechain and allow the B-factors to go 
high in refinement to represent that the side chain is flexible. 
However, I'm aware that some people truncate sidechains if density is 
not present to justify modelling. I've also seen models where the 
sidechain is modelled but with zero occupancy if density isn't present.
Is there a consensus and justifying arguments for why one approach is 
better?

Cheers,
Rhys


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






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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Olga Moroz]

Hi everyone,

I always thought it is better to truncate so that biologists looking at the 
structures are not misled?
Not sure it the best aproach though...

Olga



> On 10 Mar 2023, at 02:45, Bernhard Lechtenberg 
> <968307750321-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi Rhys,
>  
> I am also all for leaving side chains and letting the B-factors deal with the 
> weak/absent density.
>  
> I don’t think there is a consensus, but I kind of remember that somebody did 
> a poll a few years ago and if I remember correctly the main approaches were 
> the one described above, or trimming the side-chains.
>  
> Bernhard
>  
> Bernhard C. Lechtenberg PhD
> NHMRC Emerging Leadership Fellow
> Laboratory Head
> Ubiquitin Signalling Division​​
> E lechtenber...@wehi.edu.au <mailto:lechtenber...@wehi.edu.au>
> T +61 3 9345 2217
>  
>  
> From: CCP4 bulletin board  on behalf of Rhys Grinter 
> <22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>
> Date: Friday, 10 March 2023 at 12:26 pm
> To: CCP4BB@JISCMAIL.AC.UK 
> Subject: [ccp4bb] To Trim or Not to To Trim
> 
> Hi All,
>  
> I'm trying to crowdsource an opinion on how people deal with modelling side 
> chains with poorly resolved electron or cryoEM density.
>  
> My preference is to model the sidechain and allow the B-factors to go high in 
> refinement to represent that the side chain is flexible. However, I'm aware 
> that some people truncate sidechains if density is not present to justify 
> modelling. I've also seen models where the sidechain is modelled but with 
> zero occupancy if density isn't present. 
>  
> Is there a consensus and justifying arguments for why one approach is better?
>  
> Cheers,
>  
> Rhys
>  
>  
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1 
> <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 
> <https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1>



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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Edwin Pozharski]

The most pertinent question is, of course, what is the average frequency of
the disordered chain controversy flareup.  Once we figure that out, some
profound mysteries of the Universe will reveal themselves.  I am betting
it's a simple combination of the solar cycle, inflation adjusted price of a
bag of roasted chestnuts at Heinzels Wintermärchen and certainly the time
elapsed since the most recent cicada emergence in Montgomery
County, Maryland.  I am still working on the fudge factors, but maybe I
should just ask a neural network.

Aha, here is the (non)-answer to "What should I do about disordered
sidechains"

Disordered sidechains can be a common issue in protein structures,
> particularly in regions that are flexible or have low electron density. One
> approach is to manually adjust the position of the sidechain using software
> tools such as Coot or PyMOL. However, if the resolution of the protein
> structure is low or the disorder is extensive, it may be difficult to
> accurately position the sidechain. In such cases, it may be necessary to
> consider alternative approaches such as mutagenesis, site-directed labeling
> or multiple conformer modeling. The specific approach would depend on the
> particular protein structure being studied and the goals of the research
> project.


Joking aside, here is what I learned from the vigorous dead horse beatings
in the past.

- Some people like keeping adding invisible atoms to side chains and
letting B factors sort it out, Caedite eos. Novit enim Dominus qui sunt
eius.
- Some people like to remove invisible atoms.  These weirdos claim that all
atoms should be treated equally, and that adding a bunch of arginine atoms
is not different from completing the his-tag or adding hydrogens at medium
resolution.  They point out that adding disordered side chains (marginally)
increases statistical error on the rest of the well defined structure (
https://tinyurl.com/yvz73ak2) and refer to the practice as orwellian.
Snobs, basically.
- A small minority assigns zero occupancy to disordered atoms.  I admire
them as they have gone beyond agnosticism.  Let me give you coordinates for
the atom, they say.  Let me then tell you that atoms are never there and
thus have zero occupancy at that position.  Meaningless, you say?  Well, so
is your existence and Santa Claus is your parents.

Now that my telomeres are a few hundred units shorter than they were last
time I whipped myself up into a frenzy over disordered side chains, I would
summarize my views this way.

*People are free to choose how to interpret experimental data that they
themselves obtained. *
*Published structures are reported to the PDB, and if someone doesn't like
what I did to my arginines - well, interpret it your own way, I am not
stopping you. *

Also, this (with full appreciation of the formal ccp4bb focus on
computational crystallography - but come on, you are not here for the
maximum likelihood formalism)

*I am yet to see an example of when different modeling of a disordered side
chain 30 angstroms away from the area where actual molecular biology
happens has changed conclusions derived from protein structure in question.*

Special thanks to Adrian Goldman for changing my attitude on this specific
point years ago.

Cheers.

---
I don't know why the sacrifice didn't work. The science seemed so solid.
Julien XIII, Lord of the Lemurs

On Thu, Mar 9, 2023 at 8:26 PM Rhys Grinter <
22087c81e8c6-dmarc-requ...@jiscmail.ac.uk> wrote:

> Hi All,
>
> I'm trying to crowdsource an opinion on how people deal with modelling
> side chains with poorly resolved electron or cryoEM density.
>
> My preference is to model the sidechain and allow the B-factors to go high
> in refinement to represent that the side chain is flexible. However, I'm
> aware that some people truncate sidechains if density is not present to
> justify modelling. I've also seen models where the sidechain is modelled
> but with zero occupancy if density isn't present.
>
> Is there a consensus and justifying arguments for why one approach is
> better?
>
> Cheers,
>
> Rhys
>
>
> --
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
>



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

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

Re: [ccp4bb] To Trim or Not to To Trim

[David A Case]

On Tue, Mar 28, 2023, James Holton wrote:


Members of an ensemble spread out over a supercell are equivalent to a 
conventional multi-conformer model, but only when it comes to the 
density derived from the coordinate atoms alone.  They are not the 
same when it comes to bulk solvent and also not the same when it comes 
to clashes at crystal packing interfaces.



...


You might think that with enough copies in the ensemble it 
should be possible to fit any density with geometrically reasonable 
molecules, but that is not what happens in practice. This is actually 
quite remarkable!  So many "free parameters" and yet the tug-o-war 
between R factors and geometry remains.  Now, of course, the real 
molecules in the real crystal are simultaneously obeying the laws of 
chemistry and generating the diffraction patterns that we see, but I 
have never found a way to make a macromolecular model that does both.  
Neither has anybody else, of course, but wouldn't it be cool if 
someone figured out how?


I have not "figured out how", but do see what seem to be significant
improvements using supercell refinements combined with a force field to
represent the "laws of chemistry".  This is basically what James has
described, but also setting useAmber=True in phenix; I actually use my own
codes in order to get more speed and to gain better control over the
sampling.

Quite a number of high-resolution globular proteins refine to structures
with excellent geometries and Rwork in the 5-6% range (and Rfree 1-2%
higher).  Peptides can be even better: 2ol9 has 4.0/4.9 for Rwork/Rfree.
Please note that these are the best examples, illustrating what is possible.
They are not typical, and my current method seems to require data at 1 Ang.
or better.  But my results suggest that there is more backbone heterogeneity
than is typically accounted for in the atomic models deposited in the PDB.

The examples cited above use 12 copies of each protein chain; I don't know
much about the dependence on copy number.  Improving the model for solvent
contributions can improve things, but so far not by a lot.  And the
(current?) restriction to high-resolution data sets is a pretty severe
limitation.  But having a few people exploring the supercell idea seems
worthwhile, and I'm happy to share ideas (and code) with anyone who is
interested.

dave case



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] To Trim or Not to To Trim

[James Holton]

 another 
place to put the Lys, so why not build into that screaming, 6-sigma 
difference peak? Here is what happens when you do that:


CCtrue   Rwork%  Rfree% fo-fc(sigma)   description
0.8943 9.05 10.60  5.9         stump at CB
0.9580 9.95 11.60  6.4         stump at CG
0.9585 10.20 12.29  6.2         stump at CG, all 10 confs
0.9543 10.61 12.24  5.3         stump at CD, all 10 confs
0.9383 10.69 14.64 4.1 stump at CE, all 10 confs
0.9476 9.66 13.48 4.6 all atoms, all 10 confs
0.9214     7.09 11.8 5.6 three conformers (worst of 120 combos)
0.9718 6.53 8.55 4.3 three conformers (best of 120 combos)
0.9710     7.17 9.44 6.1 two conformers (best of 45 combos)
0.9471 10.35 15.04 5.1 single conformer (best of 10 choices)

If I add one CG, the other two chi1 positions light up. So, I tried 
building in all 10 true CG positions, and let the refinement decide 
what to do with them. The clear indication was that a CD should be 
added. After adding all the CDs, the difference peaks were weaker, 
but still indicating more atoms were needed.  Rwork and Rfree, 
however, tell the opposite story.  They get worse the more atoms you 
add.  CCtrue, on the other hand, was best when cutting everything 
after CG.  Why is that? Well, every time you add another atom you 
fill in the difference density, but then that atom pushes back the 
bulk solvent model that was filling in the density for the next 
atom.  The atom-to-solvent distance is roughly twice that of a 
covalent bond.  So again, square pegs and round holes.


Three conformers coming out as the winner may be because it is a 
selective process with a noisy score. In the ground truth there are 
10 conformers at equal occupancy, so no one triplet is really any 
better than any other. However, one has a density shape that fits 
better than other combos. My search over all possible quartets is 
still running.


But what if we got the solvent "right"?  Well, here is what that 
looks like:


CCtrue Rwork%  Rfree% fo-fc(sigma) description
0.9476 9.66 13.48 4.6 all atoms, all confs, refmac defaults
0.9696 6.15 8.88  3.7         all atoms, all confs, phenix.refine
0.9825     0.80    0.89  3.9     all atoms, all confs, true solvent
0.9824 0.92 1.26  7.3         true model, minus one H atom from 
ordered HIS side chain


You can see that the default solvent of phenix.refine fares better 
than refmac here, but since I generated the solvent with phenix 
refine it may have an unfair advantage. Nevertheless, providing the 
"true solvent" here is quite a striking drop in R factors.  This is 
not surprising since this was the last systematic error in this 
ground truth.  In all cases, I provided the true atomic positions at 
the start of refinement, so there was no confusion about 
strain-inducing local minima, such as which rotamer goes with which 
main chain shift. And yes, you can provide arbitrary bulk solvent 
maps to refmac5 using the "Fpart" feature.  I've had good luck with 
real data using bulk density derived form MD simulations.


What is more, once the R factors are this low I can remove just one 
hydrogen atom and it comes back as a 7.3-sigma difference peak. This 
corresponds to the protonation state of that His.  This kind of 
sensitivity is really attractive if you are looking for low-lying 
features, such as partially-occupied ligands.  Some may pooh-pooh R 
factors as "cosmetic" features of structures, but they are, in fact, 
nothing more or less than the % error between your model and your 
data.  This % error translates directly into the noise level of your 
map.  At 20% error there is no hope whatsoever of seeing 1-electron 
changes. This is because hydrogen is only 17% of a carbon.  But 3-5% 
error, which is a typical experimental error in crystallographic 
data, anything bigger than one electron is clear.


-James Holton
MAD Scientist



On 3/18/2023 2:10 PM, Nicholas Pearce wrote:
Not stupid, but essentially the same as modelling alt confs, though 
would probably give more overfitting. Alt confs can easily be 
converted to an ensemble (if done properly…).


Thanks,
Nick

———

Nicholas Pearce
Assistant Professor in Bioinformatics & DDLS Fellow
Linköping University
Sweden

--------
*From:* CCP4 bulletin board  on behalf of 
benjamin bax 

*Sent:* Saturday, March 18, 2023 10:07:26 PM
*To:* CCP4BB@JISCMAIL.AC.UK 
*Subject:* Re: [ccp4bb] To Trim or Not to To Trim
Hi,
Probably a stupid question.
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 
or 27 structures) and restrain well defined parts of structure to 
be ‘identical’ ? To give you a more NMR like chemically sensible 
ensemble of structures?

Ben


> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>
> Models for crystallography have two purposes: refinement and 
interpretation. Here these two purposes are in conflict. Neither 
case is handle

Re: [ccp4bb] To Trim or Not to To Trim

[James Holton]

ppen.


Very important: DO NOT LOOK AT THE MAPS that come out of this kind of 
refinement. Not directly. They will look all weird and distorted. You 
need to average over all the ASUs to recover interpretable 2Fo-Fc and 
Fo-Fc maps. The strictly correct way to do this is to cut out each ASU, 
re-orient and align these maps, and then average them all together. 
Fortunately, there is another way to do this that is much easier and faster:


7) Re-index the refinement output mtz file with "reindex h/2,k/2,l/2", 
and in the same run of "reindex" change your space group back to what it 
was in your normal refinement.  Do NOT run this mtz file through cad.  
Cad will throw out all but one ASU. In fact, don't open this re-indexed 
mtz file with any program other than coot or "fft". The reason is this 
mtz file still has P1 data. It is "overcomplete". Just like when cad 
performed the "outlim space 1" there is more than one ASU in the mtz.  
Perhaps it is a quirk in the fft algorithm, but this all-ASU averaging 
happens automatically if the input reflection data is "overcomplete".


8) finally, you might want to write a script for reversing the procedure 
for populating your supercell so that you can align all the members into 
a single ASU and look at them.



But wait!? Isn't this "over-fitting"?  No, it is not.  Over-fitting in 
general is when you drive your residual (aka Rwork) essentially to 
zero.  That doesn't happen here because the geometry term holds you 
back.  You might think that with enough copies in the ensemble it should 
be possible to fit any density with geometrically reasonable molecules, 
but that is not what happens in practice. This is actually quite 
remarkable!  So many "free parameters" and yet the tug-o-war between R 
factors and geometry remains.  Now, of course, the real molecules in the 
real crystal are simultaneously obeying the laws of chemistry and 
generating the diffraction patterns that we see, but I have never found 
a way to make a macromolecular model that does both.  Neither has 
anybody else, of course, but wouldn't it be cool if someone figured out how?


-James Holton
MAD Scientist

On 3/18/2023 6:35 PM, Oganesyan, Vaheh wrote:


Hi Ben,

All copies created by multiplying cell dimensions will act exactly 
same as the original one, mathematically exactly. Nick’s approach is 
better. Something similar to what Nick said was published around 
2002-2003. I was reviewing it. I did not understand then what the 
author was trying to achieve and kept thinking about it for few 
months. The author split model into 20 each with 5% occupancy. After 
refinement he got an ensemble that looked like NMR structures. I’m not 
sure, however, that adding that uncertainty will help answering any 
question.


Vaheh

*From:* CCP4 bulletin board  *On Behalf Of 
*benjamin bax

*Sent:* Saturday, March 18, 2023 5:07 PM
*To:* CCP4BB@JISCMAIL.AC.UK
*Subject:* Re: [ccp4bb] To Trim or Not to To Trim

Hi,
Probably a stupid question.
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 
27 structures) and restrain well defined parts of structure to be 
‘identical’ ? To give you a more NMR like chemically sensible ensemble 
of structures?

Ben


> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>
> Models for crystallography have two purposes: refinement and 
interpretation. Here these two purposes are in conflict. Neither case 
is handled well by either trim or not trim scenario, but trimming 
results in a deficit for refinement and not-trimming results in a 
deficit for interpretation.

>
> Our computational tools are not “fixed” in the same way that the 
standard amino acids are “fixed” or your government’s bureaucracy 
pathways are “fixed”. They are open for debate and for adjustments. 
This is a fine example where it may be more productive to discuss the 
options for making changes to the model itself or its representation, 
to better account for awkward situations such as these. Otherwise we 
are left figuring out the best imperfect way to use an imperfect tool 
(as all tools are, to varying degrees!), which isn’t satisfying for 
enough people, enough of the time.

>
> I now appreciate the hypocrisy in the argument “do not trim, but 
also don’t model disordered regions”, even though I’d be keen to avoid 
trimming. This discussion has therefore softened my own viewpoint.

>
> My refinement models (as implemented in Vagabond) do away with the 
concept of B factors precisely for the anguish it causes here, and 
refines a distribution of protein conformations which is sampled to 
generate an ensemble. By describing the conformations through the 
torsion angles that comprise the protein, modelling flexibility of a 
disordered lysine is comparatively trivial, and indeed modelling all 
possible conformations of a disordered loop becomes feasible. Lysines 
end up looking like a frayed end of a rope.

Re: [ccp4bb] To Trim or Not to To Trim

[Harry Powell]

here is 
>> quite a striking drop in R factors.  This is not surprising since this was 
>> the last systematic error in this ground truth.  In all cases, I provided 
>> the true atomic positions at the start of refinement, so there was no 
>> confusion about strain-inducing local minima, such as which rotamer goes 
>> with which main chain shift.  And yes, you can provide arbitrary bulk 
>> solvent maps to refmac5 using the "Fpart" feature.  I've had good luck with 
>> real data using bulk density derived form MD simulations.
>> 
>> What is more, once the R factors are this low I can remove just one hydrogen 
>> atom and it comes back as a 7.3-sigma difference peak. This corresponds to 
>> the protonation state of that His.  This kind of sensitivity is really 
>> attractive if you are looking for low-lying features, such as 
>> partially-occupied ligands.  Some may pooh-pooh R factors as "cosmetic" 
>> features of structures, but they are, in fact, nothing more or less than the 
>> % error between your model and your data.  This % error translates directly 
>> into the noise level of your map.  At 20% error there is no hope whatsoever 
>> of seeing 1-electron changes. This is because hydrogen is only 17% of a 
>> carbon.  But 3-5% error, which is a typical experimental error in 
>> crystallographic data, anything bigger than one electron is clear.
>> 
>> -James Holton
>> MAD Scientist
>> 
>> 
>> 
>> On 3/18/2023 2:10 PM, Nicholas Pearce wrote:
>>> Not stupid, but essentially the same as modelling alt confs, though would 
>>> probably give more overfitting. Alt confs can easily be converted to an 
>>> ensemble (if done properly…). 
>>> 
>>> Thanks,
>>> Nick
>>> 
>>> ———
>>> 
>>> Nicholas Pearce
>>> Assistant Professor in Bioinformatics & DDLS Fellow
>>> Linköping University
>>> Sweden
>>> 
>>> From: CCP4 bulletin board  on behalf of benjamin bax 
>>> 
>>> Sent: Saturday, March 18, 2023 10:07:26 PM
>>> To: CCP4BB@JISCMAIL.AC.UK 
>>> Subject: Re: [ccp4bb] To Trim or Not to To Trim
>>> 
>>> Hi,
>>> Probably a stupid question. 
>>> Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27 
>>> structures) and restrain well defined parts of structure to be ‘identical’ 
>>> ? To give you a more NMR like chemically sensible ensemble of structures?
>>> Ben
>>> 
>>> 
>>>> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>>>> 
>>>> Models for crystallography have two purposes: refinement and 
>>>> interpretation. Here these two purposes are in conflict. Neither case is 
>>>> handled well by either trim or not trim scenario, but trimming results in 
>>>> a deficit for refinement and not-trimming results in a deficit for 
>>>> interpretation.
>>>> 
>>>> Our computational tools are not “fixed” in the same way that the standard 
>>>> amino acids are “fixed” or your government’s bureaucracy pathways are 
>>>> “fixed”. They are open for debate and for adjustments. This is a fine 
>>>> example where it may be more productive to discuss the options for making 
>>>> changes to the model itself or its representation, to better account for 
>>>> awkward situations such as these. Otherwise we are left figuring out the 
>>>> best imperfect way to use an imperfect tool (as all tools are, to varying 
>>>> degrees!), which isn’t satisfying for enough people, enough of the time.
>>>> 
>>>> I now appreciate the hypocrisy in the argument “do not trim, but also 
>>>> don’t model disordered regions”, even though I’d be keen to avoid 
>>>> trimming. This discussion has therefore softened my own viewpoint.
>>>> 
>>>> My refinement models (as implemented in Vagabond) do away with the concept 
>>>> of B factors precisely for the anguish it causes here, and refines a 
>>>> distribution of protein conformations which is sampled to generate an 
>>>> ensemble. By describing the conformations through the torsion angles that 
>>>> comprise the protein, modelling flexibility of a disordered lysine is 
>>>> comparatively trivial, and indeed modelling all possible conformations of 
>>>> a disordered loop becomes feasible. Lysines end up looking like a frayed 
>>>> end of a rope. Each conformation can produce its own solvent mask, which 
>>>> can 

Re: [ccp4bb] To Trim or Not to To Trim

[Harry Powell]

, such as 
> partially-occupied ligands.  Some may pooh-pooh R factors as "cosmetic" 
> features of structures, but they are, in fact, nothing more or less than the 
> % error between your model and your data.  This % error translates directly 
> into the noise level of your map.  At 20% error there is no hope whatsoever 
> of seeing 1-electron changes. This is because hydrogen is only 17% of a 
> carbon.  But 3-5% error, which is a typical experimental error in 
> crystallographic data, anything bigger than one electron is clear.
> 
> -James Holton
> MAD Scientist
> 
> 
> 
> On 3/18/2023 2:10 PM, Nicholas Pearce wrote:
>> Not stupid, but essentially the same as modelling alt confs, though would 
>> probably give more overfitting. Alt confs can easily be converted to an 
>> ensemble (if done properly…). 
>> 
>> Thanks,
>> Nick
>> 
>> ———
>> 
>> Nicholas Pearce
>> Assistant Professor in Bioinformatics & DDLS Fellow
>> Linköping University
>> Sweden
>> 
>> From: CCP4 bulletin board  on behalf of benjamin bax 
>> 
>> Sent: Saturday, March 18, 2023 10:07:26 PM
>> To: CCP4BB@JISCMAIL.AC.UK 
>> Subject: Re: [ccp4bb] To Trim or Not to To Trim
>>  
>> Hi,
>> Probably a stupid question. 
>> Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27 
>> structures) and restrain well defined parts of structure to be ‘identical’ ? 
>> To give you a more NMR like chemically sensible ensemble of structures?
>> Ben
>> 
>> 
>> > On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>> > 
>> > Models for crystallography have two purposes: refinement and 
>> > interpretation. Here these two purposes are in conflict. Neither case is 
>> > handled well by either trim or not trim scenario, but trimming results in 
>> > a deficit for refinement and not-trimming results in a deficit for 
>> > interpretation.
>> > 
>> > Our computational tools are not “fixed” in the same way that the standard 
>> > amino acids are “fixed” or your government’s bureaucracy pathways are 
>> > “fixed”. They are open for debate and for adjustments. This is a fine 
>> > example where it may be more productive to discuss the options for making 
>> > changes to the model itself or its representation, to better account for 
>> > awkward situations such as these. Otherwise we are left figuring out the 
>> > best imperfect way to use an imperfect tool (as all tools are, to varying 
>> > degrees!), which isn’t satisfying for enough people, enough of the time.
>> > 
>> > I now appreciate the hypocrisy in the argument “do not trim, but also 
>> > don’t model disordered regions”, even though I’d be keen to avoid 
>> > trimming. This discussion has therefore softened my own viewpoint.
>> > 
>> > My refinement models (as implemented in Vagabond) do away with the concept 
>> > of B factors precisely for the anguish it causes here, and refines a 
>> > distribution of protein conformations which is sampled to generate an 
>> > ensemble. By describing the conformations through the torsion angles that 
>> > comprise the protein, modelling flexibility of a disordered lysine is 
>> > comparatively trivial, and indeed modelling all possible conformations of 
>> > a disordered loop becomes feasible. Lysines end up looking like a frayed 
>> > end of a rope. Each conformation can produce its own solvent mask, which 
>> > can be summed together to produce a blurring of density that matches what 
>> > you would expect to see in the crystal.
>> > 
>> > In my experience this doesn’t drop the R factors as much as you’d assume, 
>> > because blurred out protein density does look very much like solvent, but 
>> > it vastly improves the interpretability of the model. This also better 
>> > models the boundary between the atoms you would trim and those you’d leave 
>> > untrimmed, by avoiding such a binary distinction. No fear of trimming and 
>> > pushing those errors unseen into the rest of the structure. No fear of 
>> > leaving atoms in with an inadequate B factor model that cannot capture the 
>> > nature of the disorder.
>> > 
>> > Vagabond is undergoing a heavy rewrite though, and is not yet ready for 
>> > human consumption. Its first iteration worked on 
>> > single-dataset-single-model refinement, which handled disordered side 
>> > chains well enough, with no need to decide to exclude atoms. The heart of 
>> > the 

Re: [ccp4bb] To Trim or Not to To Trim

[Eleanor Dodson]

 1-electron changes. This is because hydrogen is only
> 17% of a carbon.  But 3-5% error, which is a typical experimental error in
> crystallographic data, anything bigger than one electron is clear.
>
> -James Holton
> MAD Scientist
>
>
>
> On 3/18/2023 2:10 PM, Nicholas Pearce wrote:
>
> Not stupid, but essentially the same as modelling alt confs, though would
> probably give more overfitting. Alt confs can easily be converted to an
> ensemble (if done properly…).
>
> Thanks,
> Nick
>
> ———
>
> Nicholas Pearce
> Assistant Professor in Bioinformatics & DDLS Fellow
> Linköping University
> Sweden
>
> --
> *From:* CCP4 bulletin board 
>  on behalf of benjamin bax 
> 
> *Sent:* Saturday, March 18, 2023 10:07:26 PM
> *To:* CCP4BB@JISCMAIL.AC.UK 
> 
> *Subject:* Re: [ccp4bb] To Trim or Not to To Trim
>
> Hi,
> Probably a stupid question.
> Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27
> structures) and restrain well defined parts of structure to be ‘identical’
> ? To give you a more NMR like chemically sensible ensemble of structures?
> Ben
>
>
> > On 18 Mar 2023, at 12:04, Helen Ginn 
>  wrote:
> >
> > Models for crystallography have two purposes: refinement and
> interpretation. Here these two purposes are in conflict. Neither case is
> handled well by either trim or not trim scenario, but trimming results in a
> deficit for refinement and not-trimming results in a deficit for
> interpretation.
> >
> > Our computational tools are not “fixed” in the same way that the
> standard amino acids are “fixed” or your government’s bureaucracy pathways
> are “fixed”. They are open for debate and for adjustments. This is a fine
> example where it may be more productive to discuss the options for making
> changes to the model itself or its representation, to better account for
> awkward situations such as these. Otherwise we are left figuring out the
> best imperfect way to use an imperfect tool (as all tools are, to varying
> degrees!), which isn’t satisfying for enough people, enough of the time.
> >
> > I now appreciate the hypocrisy in the argument “do not trim, but also
> don’t model disordered regions”, even though I’d be keen to avoid trimming.
> This discussion has therefore softened my own viewpoint.
> >
> > My refinement models (as implemented in Vagabond) do away with the
> concept of B factors precisely for the anguish it causes here, and refines
> a distribution of protein conformations which is sampled to generate an
> ensemble. By describing the conformations through the torsion angles that
> comprise the protein, modelling flexibility of a disordered lysine is
> comparatively trivial, and indeed modelling all possible conformations of a
> disordered loop becomes feasible. Lysines end up looking like a frayed end
> of a rope. Each conformation can produce its own solvent mask, which can be
> summed together to produce a blurring of density that matches what you
> would expect to see in the crystal.
> >
> > In my experience this doesn’t drop the R factors as much as you’d
> assume, because blurred out protein density does look very much like
> solvent, but it vastly improves the interpretability of the model. This
> also better models the boundary between the atoms you would trim and those
> you’d leave untrimmed, by avoiding such a binary distinction. No fear of
> trimming and pushing those errors unseen into the rest of the structure. No
> fear of leaving atoms in with an inadequate B factor model that cannot
> capture the nature of the disorder.
> >
> > Vagabond is undergoing a heavy rewrite though, and is not yet ready for
> human consumption. Its first iteration worked on
> single-dataset-single-model refinement, which handled disordered side
> chains well enough, with no need to decide to exclude atoms. The heart of
> the issue lies in main chain flexibility, and this must be handled
> correctly, for reasons of interpretability and elucidating the biological
> impact. This model isn’t perfect either, and necessitates its own
> compromises - but will provide another tool in the structural biology
> arsenal.
> >
> > —-
> >
> > Dr Helen Ginn
> > Group leader, DESY
> > Hamburg Advanced Research Centre for Bioorganic Chemistry (HARBOR)
> > Luruper Chaussee 149
> > 22607 Hamburg
> >
> > 
> >
> > 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

Re: [ccp4bb] To Trim or Not to To Trim

[Guillaume Gaullier]

Yes, I meant maps from SPA. This is the great thing about cryoEM: most maps are 
discovery maps!

I agree that showing crystallographic maps requires more careful explanations 
for non-experts, but I still believe we, as a field trying to make our nerdy 
technical things more approachable to others, are better off showing maps than 
not.

Cheers,

Guillaume


On 17 Mar 2023, at 23:57, Huw Jenkins 
mailto:h.t.jenk...@me.com>> wrote:



On 17 Mar 2023, at 15:01, Guillaume Gaullier 
mailto:guillaume.gaull...@kemi.uu.se>> wrote:

CryoEM papers often show a map in a main figure, and as a reader I think it is 
very nice to show me the map that convinced you of some finding before showing 
me your interpretation of this map.

Surely a key difference is that cryoEM maps (assuming you mean SPA) are not 
calculated with phases derived from the model shown built into the map?

Most maps shown to convince the reader of the validity of the authors 
interpretation of the electron density in X-ray crystallography papers use some 
attempt at removing bias from 2Fo-Fc or Fo-Fc density when the phases have at 
some point in the refinement been derived from a model including the region of 
interest. Few are “discovery maps” to use Dale Tronrud’s term: 
https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg04276.html


Huw

VARNING: Klicka inte på länkar och öppna inte bilagor om du inte känner igen 
avsändaren och vet att innehållet är säkert.
CAUTION: Do not click on links or open attachments unless you recognise the 
sender and know the content is safe.









När du har kontakt med oss på Uppsala universitet med e-post så innebär det att 
vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du 
läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/

E-mailing Uppsala University means that we will process your personal data. For 
more information on how this is performed, please read here: 
http://www.uu.se/en/about-uu/data-protection-policy



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] To Trim or Not to To Trim

[Lijun Liu]

noise level of your
map.  At 20% error there is no hope whatsoever of seeing 1-electron
changes. This is because hydrogen is only 17% of a carbon.  But 3-5%
error, which is a typical experimental error in crystallographic
data, anything bigger than one electron is clear.

-James Holton
MAD Scientist



On 3/18/2023 2:10 PM, Nicholas Pearce
  wrote:


  
  

  Not stupid, but essentially the same as
modelling alt confs, though would probably give more
overfitting. Alt confs can easily be converted to an
ensemble (if done properly…). 


  
  
  
Thanks,
Nick


———


Nicholas Pearce
Assistant Professor in Bioinformatics &
  DDLS Fellow
Linköping University
Sweden


  

  
  
      From: CCP4
  bulletin board  on behalf of
  benjamin bax 
  Sent: Saturday, March 18, 2023 10:07:26 PM
  To: CCP4BB@JISCMAIL.AC.UK 
  Subject: Re: [ccp4bb] To Trim or Not to To Trim
 
  
  
Hi,
  Probably a stupid question. 
  Could you multiply a, b and c cell dimensions by 2 or 3
  (to give 8 or 27 structures) and restrain well defined
  parts of structure to be ‘identical’ ? To give you a more
  NMR like chemically sensible ensemble of structures?
  Ben
  
  
  > On 18 Mar 2023, at 12:04, Helen Ginn
   wrote:
  > 
  > Models for crystallography have two purposes:
  refinement and interpretation. Here these two purposes are
  in conflict. Neither case is handled well by either trim
  or not trim scenario, but trimming results in a deficit
  for refinement and not-trimming results in a deficit for
  interpretation.
  > 
  > Our computational tools are not “fixed” in the same
  way that the standard amino acids are “fixed” or your
  government’s bureaucracy pathways are “fixed”. They are
  open for debate and for adjustments. This is a fine
  example where it may be more productive to discuss the
  options for making changes to the model itself or its
  representation, to better account for awkward situations
  such as these. Otherwise we are left figuring out the best
  imperfect way to use an imperfect tool (as all tools are,
  to varying degrees!), which isn’t satisfying for enough
  people, enough of the time.
  > 
  > I now appreciate the hypocrisy in the argument “do
  not trim, but also don’t model disordered regions”, even
  though I’d be keen to avoid trimming. This discussion has
  therefore softened my own viewpoint.
  > 
  > My refinement models (as implemented in Vagabond) do
  away with the concept of B factors precisely for the
  anguish it causes here, and refines a distribution of
  protein conformations which is sampled to generate an
  ensemble. By describing the conformations through the
  torsion angles that comprise the protein, modelling
  flexibility of a disordered lysine is comparatively
  trivial, and indeed modelling all possible conformations
  of a disordered loop becomes feasible. Lysines end up
  looking like a frayed end of a rope. Each conformation can
  produce its own solvent mask, which can be summed together
  to produce a blurring of density that matches what you
  would expect to see in the crystal.
  > 
  > In my experience this doesn’t drop the R factors as
  much as you’d assume, because blurred out protein density
  does look very much like solvent, but it vastly improves
  the interpretability of the model. This also better models
  the boundary between the atoms you would trim and those
  you’d leave untrimmed, by avoiding such a binary
  distinction. No fear of trimming and pushing those errors
  unseen into the rest of the structure. No fear of leaving
  atoms in with an inadequate B factor model that cannot
  capture the nature of the disorder.
  > 
  > Vagabond is undergoing a heavy rewrite though, and is
  not yet ready for human consumption. Its first iteration
  worked on single-dataset-single-mod

Re: [ccp4bb] To Trim or Not to To Trim

[Lijun Liu]

  CCtrue  
  Rwork%  Rfree%
   
  fo-fc(sigma)  
  description
  0.9476
  9.66  
  13.48 
  4.6
  all atoms, all
  confs, refmac
  defaults
  0.9696    
  6.15   
  8.88  3.7
          all
  atoms, all
  confs,
  phenix.refine
  0.9825     0.80    0.89  3.9         all
  atoms, all confs, true solvent   
0.9824 0.92    1.26  7.3
          true
  model, minus
  one H atom
  from ordered
  HIS side chain
  
You can see that the default solvent of phenix.refine fares better
than refmac here, but since I generated the solvent with phenix
refine it may have an unfair advantage. Nevertheless, providing the
"true solvent" here is quite a striking drop in R factors.  This is
not surprising since this was the last systematic error in this
ground truth.  In all cases, I provided the true atomic positions at
the start of refinement, so there was no confusion about
strain-inducing local minima, such as which rotamer goes with which
main chain shift.  And yes, you can provide arbitrary bulk solvent
maps to refmac5 using the "Fpart" feature.  I've had good luck with
real data using bulk density derived form MD simulations.

What is more, once the R factors are this low I can remove just one
hydrogen atom and it comes back as a 7.3-sigma difference peak. This
corresponds to the protonation state of that His.  This kind of
sensitivity is really attractive if you are looking for low-lying
features, such as partially-occupied ligands.  Some may pooh-pooh R
factors as "cosmetic" features of structures, but they are, in fact,
nothing more or less than the % error between your model and your
data.  This % error translates directly into the noise level of your
map.  At 20% error there is no hope whatsoever of seeing 1-electron
changes. This is because hydrogen is only 17% of a carbon.  But 3-5%
error, which is a typical experimental error in crystallographic
data, anything bigger than one electron is clear.

-James Holton
MAD Scientist



On 3/18/2023 2:10 PM, Nicholas Pearce
  wrote:


  
  

  Not stupid, but essentially the same as
modelling alt confs, though would probably give more
overfitting. Alt confs can easily be converted to an
ensemble (if done properly…). 


  
  
  
Thanks,
Nick


———


Nicholas Pearce
Assistant Professor in Bioinformatics &
  DDLS Fellow
Linköping University
Sweden


  

  
  
      From: CCP4
  bulletin board  on behalf of
  benjamin bax 
  Sent: Saturday, March 18, 2023 10:07:26 PM
  To: CCP4BB@JISCMAIL.AC.UK 
  Subject: Re: [ccp4bb] To Trim or Not to To Trim
 
  
  
Hi,
  Probably a stupid question. 
  Could you multiply a, b and c cell dimensions by 2 or 3
  (to give 8 or 27 structures) and restrain well defined
  parts of structure to be ‘identical’ ? To give you a more
  NMR like chemically sensible ensemble of structures?
  Ben
  
  
  > On 18 Mar 2023, at 12:04, Helen Ginn
   wrote:
  > 
  > Models for crystallography have two purposes:
  refinement 

Re: [ccp4bb] To Trim or Not to To Trim

[James Holton]

4 6.1 two conformers (best of 45 combos)
0.9471 10.35 15.04 5.1 single conformer (best of 10 choices)

If I add one CG, the other two chi1 positions light up.  So, I tried 
building in all 10 true CG positions, and let the refinement decide what 
to do with them. The clear indication was that a CD should be added. 
After adding all the CDs, the difference peaks were weaker, but still 
indicating more atoms were needed.  Rwork and Rfree, however, tell the 
opposite story.  They get worse the more atoms you add.  CCtrue, on the 
other hand, was best when cutting everything after CG.  Why is that?  
Well, every time you add another atom you fill in the difference 
density, but then that atom pushes back the bulk solvent model that was 
filling in the density for the next atom.  The atom-to-solvent distance 
is roughly twice that of a covalent bond.  So again, square pegs and 
round holes.


Three conformers coming out as the winner may be because it is a 
selective process with a noisy score. In the ground truth there are 10 
conformers at equal occupancy, so no one triplet is really any better 
than any other. However, one has a density shape that fits better than 
other combos. My search over all possible quartets is still running.


But what if we got the solvent "right"?  Well, here is what that looks like:

CCtrue Rwork%  Rfree% fo-fc(sigma) description
0.9476 9.66 13.48 4.6 all atoms, all confs, refmac defaults
0.9696 6.15 8.88  3.7         all atoms, all confs, phenix.refine
0.9825     0.80    0.89  3.9         all atoms, all confs, true solvent
0.9824 0.92    1.26  7.3         true model, minus one H atom 
from ordered HIS side chain


You can see that the default solvent of phenix.refine fares better than 
refmac here, but since I generated the solvent with phenix refine it may 
have an unfair advantage. Nevertheless, providing the "true solvent" 
here is quite a striking drop in R factors.  This is not surprising 
since this was the last systematic error in this ground truth.  In all 
cases, I provided the true atomic positions at the start of refinement, 
so there was no confusion about strain-inducing local minima, such as 
which rotamer goes with which main chain shift.  And yes, you can 
provide arbitrary bulk solvent maps to refmac5 using the "Fpart" 
feature.  I've had good luck with real data using bulk density derived 
form MD simulations.


What is more, once the R factors are this low I can remove just one 
hydrogen atom and it comes back as a 7.3-sigma difference peak. This 
corresponds to the protonation state of that His.  This kind of 
sensitivity is really attractive if you are looking for low-lying 
features, such as partially-occupied ligands.  Some may pooh-pooh R 
factors as "cosmetic" features of structures, but they are, in fact, 
nothing more or less than the % error between your model and your data.  
This % error translates directly into the noise level of your map.  At 
20% error there is no hope whatsoever of seeing 1-electron changes. This 
is because hydrogen is only 17% of a carbon.  But 3-5% error, which is a 
typical experimental error in crystallographic data, anything bigger 
than one electron is clear.


-James Holton
MAD Scientist



On 3/18/2023 2:10 PM, Nicholas Pearce wrote:
Not stupid, but essentially the same as modelling alt confs, though 
would probably give more overfitting. Alt confs can easily be 
converted to an ensemble (if done properly…).


Thanks,
Nick

———

Nicholas Pearce
Assistant Professor in Bioinformatics & DDLS Fellow
Linköping University
Sweden

----
*From:* CCP4 bulletin board  on behalf of 
benjamin bax 

*Sent:* Saturday, March 18, 2023 10:07:26 PM
*To:* CCP4BB@JISCMAIL.AC.UK 
*Subject:* Re: [ccp4bb] To Trim or Not to To Trim
Hi,
Probably a stupid question.
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 
27 structures) and restrain well defined parts of structure to be 
‘identical’ ? To give you a more NMR like chemically sensible ensemble 
of structures?

Ben


> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>
> Models for crystallography have two purposes: refinement and 
interpretation. Here these two purposes are in conflict. Neither case 
is handled well by either trim or not trim scenario, but trimming 
results in a deficit for refinement and not-trimming results in a 
deficit for interpretation.

>
> Our computational tools are not “fixed” in the same way that the 
standard amino acids are “fixed” or your government’s bureaucracy 
pathways are “fixed”. They are open for debate and for adjustments. 
This is a fine example where it may be more productive to discuss the 
options for making changes to the model itself or its representation, 
to better account for awkward situations such as these. Otherwise we 
are left figuring out the best imperfect way to use an imperfect to

Re: [ccp4bb] To Trim or Not to To Trim

[Pietro Roversi]

s 
time to run another poll?


Personally, I call any other approach R-factor cosmetics. The goal in 
model building is not to achieve the lowest possible R-factors, it’s 
to build the most physically meaningful, most likely to be correct, 
model. So if you know that the side chain is part of the protein, you 
should model it the best way you can. If it’s there, just disordered, 
then the most correct way to model it is to let it have high 
B-factors. Most molecular graphics programs don’t flag zero-occupancy 
atoms, so the user might never notice. Truncation of a side chain, 
unless there is evidence that it really physically isn’t there, is 
also misleading, in my opinion. I don’t believe that it is more 
helpful to the non-expert user than high B-factors either.


If people who are not structural biologists themselves don’t know how 
to use a structure, then we need to educate them better. It is very 
straightforward these days to look at electron density in the PDB 
viewer. It used to be difficult, but nowadays there’s no excuse for 
not checking the electron density. The PDB validation flags RSRZ 
outliers. You can easily colour a structure by B-factors. It doesn’t 
take that much effort to teach students how to validate structures. 
The main point you need to get across is that it is necessary to do 
so. And this needs to be done not only in courses aimed at prospective 
experimental structural biologists, of course, but whenever students 
use structures in any way.


This is just the opinion of someone who feels very strongly about 
teaching structure validation and rejoices when students’ reply to the 
question “What was the most important thing you learned today?” is: 
“Don’t blindly trust anything.”


Cheers

/Julia

-- Dr. Julia Griese

Associate Professor (Docent)

Principal Investigator

Department of Cell and Molecular Biology

Uppsala University

BMC, Box 596

SE-75124 Uppsala

Sweden

email: julia.gri...@icm.uu.se

phone: +46-(0)18-471 4982

http://www.icm.uu.se/structural-biology/griese-lab/ 
<http://www.icm.uu.se/structural-biology/griese-lab/>


*From: *CCP4 bulletin board  on behalf of 
Bernhard Lechtenberg <968307750321-dmarc-requ...@jiscmail.ac.uk>

*Reply-To: *Bernhard Lechtenberg 
*Date: *Friday, March 10, 2023 at 05:07
*To: *"CCP4BB@JISCMAIL.AC.UK" 
*Subject: *Re: [ccp4bb] To Trim or Not to To Trim

I found the poll I wrote about earlier. This actually is way older 
than I had expected (2011). You can see the poll results (which was 
run by Ed Pozharski) and discussion at the time here in the CCP4BB 
archive:


https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html 
<https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html>


In brief, the results of 240 respondents were:

Delete the atoms 43%

Let refinement take care of it by inflating B-factors    41%

Set occupancy to zero    12%

Other 4%

Bernhard

*From: *CCP4 bulletin board  on behalf of 
Debanu Das 

*Date: *Friday, 10 March 2023 at 2:56 pm
*To: *CCP4BB@JISCMAIL.AC.UK 
*Subject: *Re: [ccp4bb] To Trim or Not to To Trim

We dealt with this in-depth during structural genomics days when we 
deposited over 1500 novel, high-quality, experimentally-phased 
structures into the PDB. Think it’s prudent to trim/truncate side 
chains without reliable density.


Non-structural biologists using PDB structures without expert help can 
err in any of these scenarios: misinterpreting most common/random 
rotamer, zero occupancy atoms, B-factors, etc.


What is the value of populating the PDB, which is a structural model 
repository, with such information that is not there, i.e., reliable 
structural model?


Any trained crystallographer/structural biologist can easily add in 
side chain information if needed for modeling/computational chemistry 
reasons.


Best regards,

Debanu

On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch <mailto:jxb...@case.edu>> wrote:


I’d say no trimming to side chains for the following reason: There
are non-structural biologists using PDB files and if atoms are
missing they don’t know what to do. A better approach is where no
side chain density allows support of placement, pick the most 
common

rotamer and set the occupancy to zero for those atoms lacking
density support. More work for you but more accurate in my 
opinion.


Jürgen

___

Jürgen Bosch, PhD, MBA

Center for Global Health & Diseases

Case Western Reserve University

Cleveland, OH 44106

https://www.linkedin.com/in/jubosch/
<https://www.linkedin.com/in/jubosch/>

CEO & Co-Founder at InterRayBio, LLC

On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg
<968307750321-dmarc-requ...@jiscmail.ac.uk
<mailto:968307750321-dmarc-requ...@ji

Re: [ccp4bb] To Trim or Not to To Trim

[Oganesyan, Vaheh]

Hi Ben,

All copies created by multiplying cell dimensions will act exactly same as the 
original one, mathematically exactly. Nick’s approach is better. Something 
similar to what Nick said was published around 2002-2003. I was reviewing it. I 
did not understand then what the author was trying to achieve and kept thinking 
about it for few months. The author split model into 20 each with 5% occupancy. 
After refinement he got an ensemble that looked like NMR structures. I’m not 
sure, however, that adding that uncertainty will help answering any question.

Vaheh

From: CCP4 bulletin board  On Behalf Of benjamin bax
Sent: Saturday, March 18, 2023 5:07 PM
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] To Trim or Not to To Trim
Hi,
Probably a stupid question.
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27 
structures) and restrain well defined parts of structure to be ‘identical’ ? To 
give you a more NMR like chemically sensible ensemble of structures?
Ben


> On 18 Mar 2023, at 12:04, Helen Ginn 
> mailto:ccp...@hginn.co.uk>> wrote:
>
> Models for crystallography have two purposes: refinement and interpretation. 
> Here these two purposes are in conflict. Neither case is handled well by 
> either trim or not trim scenario, but trimming results in a deficit for 
> refinement and not-trimming results in a deficit for interpretation.
>
> Our computational tools are not “fixed” in the same way that the standard 
> amino acids are “fixed” or your government’s bureaucracy pathways are 
> “fixed”. They are open for debate and for adjustments. This is a fine example 
> where it may be more productive to discuss the options for making changes to 
> the model itself or its representation, to better account for awkward 
> situations such as these. Otherwise we are left figuring out the best 
> imperfect way to use an imperfect tool (as all tools are, to varying 
> degrees!), which isn’t satisfying for enough people, enough of the time.
>
> I now appreciate the hypocrisy in the argument “do not trim, but also don’t 
> model disordered regions”, even though I’d be keen to avoid trimming. This 
> discussion has therefore softened my own viewpoint.
>
> My refinement models (as implemented in Vagabond) do away with the concept of 
> B factors precisely for the anguish it causes here, and refines a 
> distribution of protein conformations which is sampled to generate an 
> ensemble. By describing the conformations through the torsion angles that 
> comprise the protein, modelling flexibility of a disordered lysine is 
> comparatively trivial, and indeed modelling all possible conformations of a 
> disordered loop becomes feasible. Lysines end up looking like a frayed end of 
> a rope. Each conformation can produce its own solvent mask, which can be 
> summed together to produce a blurring of density that matches what you would 
> expect to see in the crystal.
>
> In my experience this doesn’t drop the R factors as much as you’d assume, 
> because blurred out protein density does look very much like solvent, but it 
> vastly improves the interpretability of the model. This also better models 
> the boundary between the atoms you would trim and those you’d leave 
> untrimmed, by avoiding such a binary distinction. No fear of trimming and 
> pushing those errors unseen into the rest of the structure. No fear of 
> leaving atoms in with an inadequate B factor model that cannot capture the 
> nature of the disorder.
>
> Vagabond is undergoing a heavy rewrite though, and is not yet ready for human 
> consumption. Its first iteration worked on single-dataset-single-model 
> refinement, which handled disordered side chains well enough, with no need to 
> decide to exclude atoms. The heart of the issue lies in main chain 
> flexibility, and this must be handled correctly, for reasons of 
> interpretability and elucidating the biological impact. This model isn’t 
> perfect either, and necessitates its own compromises - but will provide 
> another tool in the structural biology arsenal.
>
> —-
>
> Dr Helen Ginn
> Group leader, DESY
> Hamburg Advanced Research Centre for Bioorganic Chemistry (HARBOR)
> Luruper Chaussee 149
> 22607 Hamburg
>
> 
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1<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<http://www.jiscmail.ac.uk/CCP4BB>, a mailing list 
> hosted by www.jiscmail.ac.uk<http://www.jiscmail.ac.uk>, terms & conditions 
> are available at 
> https://www.jiscmail.ac.uk/policyandsecurity/<ht

Re: [ccp4bb] To Trim or Not to To Trim

[Nicholas Pearce]

Not stupid, but essentially the same as modelling alt confs, though would 
probably give more overfitting. Alt confs can easily be converted to an 
ensemble (if done properly…).

Thanks,
Nick

———

Nicholas Pearce
Assistant Professor in Bioinformatics & DDLS Fellow
Linköping University
Sweden


From: CCP4 bulletin board  on behalf of benjamin bax 

Sent: Saturday, March 18, 2023 10:07:26 PM
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] To Trim or Not to To Trim

Hi,
Probably a stupid question.
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27 
structures) and restrain well defined parts of structure to be ‘identical’ ? To 
give you a more NMR like chemically sensible ensemble of structures?
Ben


> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
>
> Models for crystallography have two purposes: refinement and interpretation. 
> Here these two purposes are in conflict. Neither case is handled well by 
> either trim or not trim scenario, but trimming results in a deficit for 
> refinement and not-trimming results in a deficit for interpretation.
>
> Our computational tools are not “fixed” in the same way that the standard 
> amino acids are “fixed” or your government’s bureaucracy pathways are 
> “fixed”. They are open for debate and for adjustments. This is a fine example 
> where it may be more productive to discuss the options for making changes to 
> the model itself or its representation, to better account for awkward 
> situations such as these. Otherwise we are left figuring out the best 
> imperfect way to use an imperfect tool (as all tools are, to varying 
> degrees!), which isn’t satisfying for enough people, enough of the time.
>
> I now appreciate the hypocrisy in the argument “do not trim, but also don’t 
> model disordered regions”, even though I’d be keen to avoid trimming. This 
> discussion has therefore softened my own viewpoint.
>
> My refinement models (as implemented in Vagabond) do away with the concept of 
> B factors precisely for the anguish it causes here, and refines a 
> distribution of protein conformations which is sampled to generate an 
> ensemble. By describing the conformations through the torsion angles that 
> comprise the protein, modelling flexibility of a disordered lysine is 
> comparatively trivial, and indeed modelling all possible conformations of a 
> disordered loop becomes feasible. Lysines end up looking like a frayed end of 
> a rope. Each conformation can produce its own solvent mask, which can be 
> summed together to produce a blurring of density that matches what you would 
> expect to see in the crystal.
>
> In my experience this doesn’t drop the R factors as much as you’d assume, 
> because blurred out protein density does look very much like solvent, but it 
> vastly improves the interpretability of the model. This also better models 
> the boundary between the atoms you would trim and those you’d leave 
> untrimmed, by avoiding such a binary distinction. No fear of trimming and 
> pushing those errors unseen into the rest of the structure. No fear of 
> leaving atoms in with an inadequate B factor model that cannot capture the 
> nature of the disorder.
>
> Vagabond is undergoing a heavy rewrite though, and is not yet ready for human 
> consumption. Its first iteration worked on single-dataset-single-model 
> refinement, which handled disordered side chains well enough, with no need to 
> decide to exclude atoms. The heart of the issue lies in main chain 
> flexibility, and this must be handled correctly, for reasons of 
> interpretability and elucidating the biological impact. This model isn’t 
> perfect either, and necessitates its own compromises - but will provide 
> another tool in the structural biology arsenal.
>
> —-
>
> Dr Helen Ginn
> Group leader, DESY
> Hamburg Advanced Research Centre for Bioorganic Chemistry (HARBOR)
> Luruper Chaussee 149
> 22607 Hamburg
>
> 
>
> 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%3D1=05%7C01%7Cnicholas.pearce%40LIU.SE%7Cb01b3fe2d210435bd43108db27f4d9fc%7C913f18ec7f264c5fa816784fe9a58edd%7C0%7C0%7C638147704962813881%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C=qb0fv349eLSwriyNUYdYjYw7FvjshVdZcJ%2FfUO0L2UI%3D=0
>
> This message was issued to members of 
> https://eur01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jiscmail.ac.uk%2FCCP4BB=05%7C01%7Cnicholas.pearce%40LIU.SE%7Cb01b3fe2d210435bd43108db27f4d9fc%7C913f18ec7f264c5fa816784fe9a58edd%7C0%7C0%7C638147704962813881

Re: [ccp4bb] To Trim or Not to To Trim

[benjamin bax]

Hi,
Probably a stupid question. 
Could you multiply a, b and c cell dimensions by 2 or 3 (to give 8 or 27 
structures) and restrain well defined parts of structure to be ‘identical’ ? To 
give you a more NMR like chemically sensible ensemble of structures?
Ben


> On 18 Mar 2023, at 12:04, Helen Ginn  wrote:
> 
> Models for crystallography have two purposes: refinement and interpretation. 
> Here these two purposes are in conflict. Neither case is handled well by 
> either trim or not trim scenario, but trimming results in a deficit for 
> refinement and not-trimming results in a deficit for interpretation.
> 
> Our computational tools are not “fixed” in the same way that the standard 
> amino acids are “fixed” or your government’s bureaucracy pathways are 
> “fixed”. They are open for debate and for adjustments. This is a fine example 
> where it may be more productive to discuss the options for making changes to 
> the model itself or its representation, to better account for awkward 
> situations such as these. Otherwise we are left figuring out the best 
> imperfect way to use an imperfect tool (as all tools are, to varying 
> degrees!), which isn’t satisfying for enough people, enough of the time.
> 
> I now appreciate the hypocrisy in the argument “do not trim, but also don’t 
> model disordered regions”, even though I’d be keen to avoid trimming. This 
> discussion has therefore softened my own viewpoint.
> 
> My refinement models (as implemented in Vagabond) do away with the concept of 
> B factors precisely for the anguish it causes here, and refines a 
> distribution of protein conformations which is sampled to generate an 
> ensemble. By describing the conformations through the torsion angles that 
> comprise the protein, modelling flexibility of a disordered lysine is 
> comparatively trivial, and indeed modelling all possible conformations of a 
> disordered loop becomes feasible. Lysines end up looking like a frayed end of 
> a rope. Each conformation can produce its own solvent mask, which can be 
> summed together to produce a blurring of density that matches what you would 
> expect to see in the crystal.
> 
> In my experience this doesn’t drop the R factors as much as you’d assume, 
> because blurred out protein density does look very much like solvent, but it 
> vastly improves the interpretability of the model. This also better models 
> the boundary between the atoms you would trim and those you’d leave 
> untrimmed, by avoiding such a binary distinction. No fear of trimming and 
> pushing those errors unseen into the rest of the structure. No fear of 
> leaving atoms in with an inadequate B factor model that cannot capture the 
> nature of the disorder.
> 
> Vagabond is undergoing a heavy rewrite though, and is not yet ready for human 
> consumption. Its first iteration worked on single-dataset-single-model 
> refinement, which handled disordered side chains well enough, with no need to 
> decide to exclude atoms. The heart of the issue lies in main chain 
> flexibility, and this must be handled correctly, for reasons of 
> interpretability and elucidating the biological impact. This model isn’t 
> perfect either, and necessitates its own compromises - but will provide 
> another tool in the structural biology arsenal.
> 
> —-
> 
> Dr Helen Ginn
> Group leader, DESY
> Hamburg Advanced Research Centre for Bioorganic Chemistry (HARBOR)
> Luruper Chaussee 149
> 22607 Hamburg
> 
> 
> 
> 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/

Re: [ccp4bb] To Trim or Not to To Trim

[Helen Ginn]

Models for crystallography have two purposes: refinement and 
interpretation. Here these two purposes are in conflict. Neither case is 
handled well by either trim or not trim scenario, but trimming results 
in a deficit for refinement and not-trimming results in a deficit for 
interpretation.


Our computational tools are not “fixed” in the same way that the 
standard amino acids are “fixed” or your government’s bureaucracy 
pathways are “fixed”. They are open for debate and for adjustments. This 
is a fine example where it may be more productive to discuss the options 
for making changes to the model itself or its representation, to better 
account for awkward situations such as these. Otherwise we are left 
figuring out the best imperfect way to use an imperfect tool (as all 
tools are, to varying degrees!), which isn’t satisfying for enough 
people, enough of the time.


I now appreciate the hypocrisy in the argument “do not trim, but also 
don’t model disordered regions”, even though I’d be keen to avoid 
trimming. This discussion has therefore softened my own viewpoint.


My refinement models (as implemented in Vagabond) do away with the 
concept of B factors precisely for the anguish it causes here, and 
refines a distribution of protein conformations which is sampled to 
generate an ensemble. By describing the conformations through the 
torsion angles that comprise the protein, modelling flexibility of a 
disordered lysine is comparatively trivial, and indeed modelling all 
possible conformations of a disordered loop becomes feasible. Lysines 
end up looking like a frayed end of a rope. Each conformation can 
produce its own solvent mask, which can be summed together to produce a 
blurring of density that matches what you would expect to see in the 
crystal.


In my experience this doesn’t drop the R factors as much as you’d 
assume, because blurred out protein density does look very much like 
solvent, but it vastly improves the interpretability of the model. This 
also better models the boundary between the atoms you would trim and 
those you’d leave untrimmed, by avoiding such a binary distinction. No 
fear of trimming and pushing those errors unseen into the rest of the 
structure. No fear of leaving atoms in with an inadequate B factor model 
that cannot capture the nature of the disorder.


Vagabond is undergoing a heavy rewrite though, and is not yet ready for 
human consumption. Its first iteration worked on 
single-dataset-single-model refinement, which handled disordered side 
chains well enough, with no need to decide to exclude atoms. The heart 
of the issue lies in main chain flexibility, and this must be handled 
correctly, for reasons of interpretability and elucidating the 
biological impact. This model isn’t perfect either, and necessitates its 
own compromises - but will provide another tool in the structural 
biology arsenal.


—-

Dr Helen Ginn
Group leader, DESY
Hamburg Advanced Research Centre for Bioorganic Chemistry (HARBOR)
Luruper Chaussee 149
22607 Hamburg



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] To Trim or Not to To Trim

[Dale Tronrud]

 is necessary to do so. And this needs 
to be done not only in courses aimed at prospective experimental 
structural biologists, of course, but whenever students use structures 
in any way.


This is just the opinion of someone who feels very strongly about 
teaching structure validation and rejoices when students’ reply to the 
question “What was the most important thing you learned today?” is: 
“Don’t blindly trust anything.”


Cheers

/Julia

--

Dr. Julia Griese

Associate Professor (Docent)

Principal Investigator

Department of Cell and Molecular Biology

Uppsala University

BMC, Box 596

SE-75124 Uppsala

Sweden

email: julia.gri...@icm.uu.se

phone: +46-(0)18-471 4982

http://www.icm.uu.se/structural-biology/griese-lab/ 
<http://www.icm.uu.se/structural-biology/griese-lab/>


*From: *CCP4 bulletin board  on behalf of 
Bernhard Lechtenberg <968307750321-dmarc-requ...@jiscmail.ac.uk>

*Reply-To: *Bernhard Lechtenberg 
*Date: *Friday, March 10, 2023 at 05:07
*To: *"CCP4BB@JISCMAIL.AC.UK" 
*Subject: *Re: [ccp4bb] To Trim or Not to To Trim

I found the poll I wrote about earlier. This actually is way older than 
I had expected (2011). You can see the poll results (which was run by Ed 
Pozharski) and discussion at the time here in the CCP4BB archive:


https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html 
<https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html>


In brief, the results of 240 respondents were:

Delete the atoms 43%

Let refinement take care of it by inflating B-factors    41%

Set occupancy to zero    12%

Other 4%

Bernhard

*From: *CCP4 bulletin board  on behalf of Debanu 
Das 

*Date: *Friday, 10 March 2023 at 2:56 pm
*To: *CCP4BB@JISCMAIL.AC.UK 
*Subject: *Re: [ccp4bb] To Trim or Not to To Trim

We dealt with this in-depth during structural genomics days when we 
deposited over 1500 novel, high-quality, experimentally-phased 
structures into the PDB. Think it’s prudent to trim/truncate side chains 
without reliable density.


Non-structural biologists using PDB structures without expert help can 
err in any of these scenarios: misinterpreting most common/random 
rotamer, zero occupancy atoms, B-factors, etc.


What is the value of populating the PDB, which is a structural model 
repository, with such information that is not there, i.e., reliable 
structural model?


Any trained crystallographer/structural biologist can easily add in side 
chain information if needed for modeling/computational chemistry reasons.


Best regards,

Debanu

On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch <mailto:jxb...@case.edu>> wrote:


I’d say no trimming to side chains for the following reason: There
are non-structural biologists using PDB files and if atoms are
missing they don’t know what to do. A better approach is where no
side chain density allows support of placement, pick the most common
rotamer and set the occupancy to zero for those atoms lacking
density support. More work for you but more accurate in my opinion.

Jürgen

___

Jürgen Bosch, PhD, MBA

Center for Global Health & Diseases

Case Western Reserve University

Cleveland, OH 44106

https://www.linkedin.com/in/jubosch/
<https://www.linkedin.com/in/jubosch/>

CEO & Co-Founder at InterRayBio, LLC

On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg
<968307750321-dmarc-requ...@jiscmail.ac.uk
<mailto:968307750321-dmarc-requ...@jiscmail.ac.uk>> wrote:

Hi Rhys,

I am also all for leaving side chains and letting the B-factors
deal with the weak/absent density.

I don’t think there is a consensus, but I kind of remember that
somebody did a poll a few years ago and if I remember correctly
the main approaches were the one described above, or trimming
the side-chains.

Bernhard

*Bernhard C. Lechtenberg* PhD
NHMRC Emerging Leadership Fellow
Laboratory Head
Ubiquitin Signalling Division​​

E lechtenber...@wehi.edu.au <mailto:lechtenber...@wehi.edu.au>
T +61 3 9345 2217

*From: *CCP4 bulletin board mailto:CCP4BB@JISCMAIL.AC.UK>> on behalf of Rhys Grinter
<22087c81e8c6-dmarc-requ...@jiscmail.ac.uk
<mailto:22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>>
*Date: *Friday, 10 March 2023 at 12:26 pm
*To: *CCP4BB@JISCMAIL.AC.UK <mailto:CCP4BB@JISCMAIL.AC.UK>
mailto:CCP4BB@JISCMAIL.AC.UK>>
*Subject: *[ccp4bb] To Trim or Not to To Trim

Hi All,

I'm trying to crowdsource an opinion on how people deal with
modelling side chains with poorly resolved electron or cryoEM
  

Re: [ccp4bb] To Trim or Not to To Trim

[Huw Jenkins]

> On 17 Mar 2023, at 08:57, Manfred S. Weiss 
>  wrote:
> 
> In my view, the best approach is to build the side chains in their
> most plausible conformation, or maybe in 2 or 3 or 5 different
> conformations, and let the ADPs refine freely.

One point I don’t think has been mentioned so far in this interesting debate is 
how do you ensure that none of the atoms in these conformations gets moved by 
the refinement program such that one plausible conformation becomes a rotamer 
outlier after refinement? I would think that most structural biologists would 
agree that a rotamer that is an outlier and is unsupported by electron density 
is hard to justify.


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 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] To Trim or Not to To Trim

[Huw Jenkins]

> On 17 Mar 2023, at 15:01, Guillaume Gaullier  
> wrote:
> 
> CryoEM papers often show a map in a main figure, and as a reader I think it 
> is very nice to show me the map that convinced you of some finding before 
> showing me your interpretation of this map.

Surely a key difference is that cryoEM maps (assuming you mean SPA) are not 
calculated with phases derived from the model shown built into the map?

Most maps shown to convince the reader of the validity of the authors 
interpretation of the electron density in X-ray crystallography papers use some 
attempt at removing bias from 2Fo-Fc or Fo-Fc density when the phases have at 
some point in the refinement been derived from a model including the region of 
interest. Few are “discovery maps” to use Dale Tronrud’s term: 
https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg04276.html


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 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] To Trim or Not to To Trim

[Guillaume Gaullier]

Hello,

I think Rams’ last remark is very important. People with every preference 
regarding modelling (trimmed, zero-occupancy or high-B), past their 
disagreement on this particular question, are all concerned about how their 
models will be interpreted by non-experts.

Since all experimental scientists understand the difference between observed 
data and interpretation, I believe the best way to educate non-expert users of 
macromolecular models would be for the PDB to systematically show a map. An 
overview consisting of a still image with map and model is already in 
validation reports, so it could be displayed on the entry page on the web as 
well. We now have in-browser interactive map and model viewers, so the PDB 
should make use of this capability and show the map by default, and use a 
legend that would be clear for non-experts (with terms like "observed density" 
and "atomic model interpretation of the density" or whatever explains it best 
in easy words).

We could also all get into the habit of showing a map (or even only the region 
of interest in a map) in a main figure in our papers, instead of keeping it 
hidden deep in the supplementary figures. CryoEM papers often show a map in a 
main figure, and as a reader I think it is very nice to show me the map that 
convinced you of some finding before showing me your interpretation of this 
map. I hope the cryoEM field will keep doing this, and I hope the MX field 
would start showing more maps in their papers again. Think of a scatter plot 
analogy: what sense would it make to only show the fit line? and leave the data 
points in a supplementary figure or in some database fairly obscure for 
non-experts? This is what we are doing when we only show an atomic model. Then 
no wonder non-expert readers are confused.

Showing maps would go a long way to getting more people "structurally minded", 
until the root problem is fixed with better representations for macromolecular 
models (and viewer programs able to present all the info), like Dale suggested.

Cheers,

Guillaume


On 17 Mar 2023, at 12:25, Subramanian, Ramaswamy 
mailto:subra...@purdue.edu>> wrote:

Dear All,

I am kind of in agreement with Manfred, but I also have concerns.

The RCSB and the databases are public databases used by non-structural biology 
experts a lot.  They take the model as if it is an experimental result - rather 
than it being a model.

Example:  I just got a paper with reviewers' comments.   The reviewer insists 
that there is a crystal structure that clearly shows the conformation is XXX in 
the pdb and our interpretation of the biochemical data is wrong.  The B-factors 
of the atoms that he says contradict our results are over 100, and there is no 
density for them in the ED map.   I am writing with a justification, explaining 
all these - but I will see if it is accepted.  But the fact remains that this 
person who is an expert in a different field - still uses it and makes poor 
interpretations.

Are we doing a dis-service to the large group of non-experts, who think the 
positions are experimental results and interpret data?  They are often the 
audience for who we generate and deposit the data.

Thanks, and cheers!



Rams
subra...@purdue.edu



On Mar 17, 2023, at 4:57 AM, Manfred S. Weiss 
mailto:manfred.we...@helmholtz-berlin.de>> 
wrote:

 External Email: Use caution with attachments, links, or sharing data 


Dear all,

many views have been expressed in this thread, which I have been
following with great interest. Unfortunately, I have to say that many
of the views are more based on personal preferences, than on the
scientific evidence behind.

Here are some facts, that one may want to consider:

1. When you have crystallized your protein, and collected data
from it, the obvious assumption is that it is still intact. Unless you
have additional information, e.g. from mass spec, that some
side chains are cleaved off, they are there.

2. Absence of evidence (meaning missing electron density)
is NOT evidence of absence 

3. If you trim, you give the impression that the respective atoms
are not there. This means that the solvent model will be incorrect,
because the area will be defined as solvent.

In my view, the best approach is to build the side chains in their
most plausible conformation, or maybe in 2 or 3 or 5 different
conformations, and let the ADPs refine freely.

All the best
Manfred
































































































--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. 

Re: [ccp4bb] To Trim or Not to To Trim

[mesters@biochem]

Dear all,

to trim or not to to trim….

- what is the difference between highly disordered side-chain atoms versus 
disordered bulk-solvent atoms? Do not both represent a continuum function with 
a comparable electron-density distribution? Have not solvent and side-chain 
atoms merged/mixed at this position in space and with this, are effectively 
indistinguishable (contrast loss i.e. similar contribution to diffraction) from 
one another?

- are current solvent models comprehensive enough in that modelling of atoms 
with very high B factors is really mutually beneficial for bulk and structure 
model?

- is model bias (overfitting) not increased by modelling highly disordered 
side-chain atoms (Rfree will probably not increase but R will drop)?

- analogous to the disordered side-chain atoms, bulk solvent atoms are of 
course also present in the crystal, so why not fill the AU with solvent 
molecules and let refinement take care of them?
(published methodology, see https://doi.org/10.1107/S0907444906021627 
)


Trimming or non-evidence-based placement and refinement are somehow halfhearted 
options with split opinions among the community as the discussions demonstrate.

Adding/Modelling a plausible side-chain rotamer with zero occupancy still has 
its charm and better reflects the situation "the side-chain atoms must be there 
but we do not know exactly where"……..

Cheers

Jeroen





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] To Trim or Not to To Trim

[Subramanian, Ramaswamy]

Dear All,

I am kind of in agreement with Manfred, but I also have concerns.

The RCSB and the databases are public databases used by non-structural biology 
experts a lot.  They take the model as if it is an experimental result - rather 
than it being a model.

Example:  I just got a paper with reviewers' comments.   The reviewer insists 
that there is a crystal structure that clearly shows the conformation is XXX in 
the pdb and our interpretation of the biochemical data is wrong.  The B-factors 
of the atoms that he says contradict our results are over 100, and there is no 
density for them in the ED map.   I am writing with a justification, explaining 
all these - but I will see if it is accepted.  But the fact remains that this 
person who is an expert in a different field - still uses it and makes poor 
interpretations.

Are we doing a dis-service to the large group of non-experts, who think the 
positions are experimental results and interpret data?  They are often the 
audience for who we generate and deposit the data.

Thanks, and cheers!



Rams
subra...@purdue.edu



On Mar 17, 2023, at 4:57 AM, Manfred S. Weiss 
 wrote:

 External Email: Use caution with attachments, links, or sharing data 


Dear all,

many views have been expressed in this thread, which I have been
following with great interest. Unfortunately, I have to say that many
of the views are more based on personal preferences, than on the
scientific evidence behind.

Here are some facts, that one may want to consider:

1. When you have crystallized your protein, and collected data
from it, the obvious assumption is that it is still intact. Unless you
have additional information, e.g. from mass spec, that some
side chains are cleaved off, they are there.

2. Absence of evidence (meaning missing electron density)
is NOT evidence of absence 

3. If you trim, you give the impression that the respective atoms
are not there. This means that the solvent model will be incorrect,
because the area will be defined as solvent.

In my view, the best approach is to build the side chains in their
most plausible conformation, or maybe in 2 or 3 or 5 different
conformations, and let the ADPs refine freely.

All the best
Manfred
































































































--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseher
Geschäftsführung: Prof. Dr. Bernd Rech, Thomas Frederking

Sitz Berlin, AG Charlottenburg, 89 HRB 5583

Postadresse:
Hahn-Meitner-Platz 1
14109 Berlin
Deutschland

Diese E-Mail kann vertrauliche und/oder rechtlich geschützte Informationen 
enthalten. Wenn Sie diese E-Mail irrtümlich erhalten haben, informieren Sie 
bitte sofort den*die Absender*in und vernichten Sie diese Mail. Das unerlaubte 
Kopieren, die Veröffentlichung sowie die unbefugte Weitergabe dieser Mail ist 
nicht gestattet.
This email may contain confidential and/or proprietary information. If you have 
received this e-mail in error, please inform the sender immediately and destroy 
this e-mail. Unauthorized copying, publishing or distribution of this e-mail is 
not permitted.


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




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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Goldman, Adrian]

I would add that those who insist on modelling the atoms that they can’t see 
display a quasi-religious fervour about it, as if there is only one right way. 
Sorry about the snark.

One example I remember from my graduate student days is this:
https://tinyurl.com/bddefd3t

where intact and truncated gamma-deta resolvase crystallised in *exactly* the 
same crystal form, with two entire DNA-binding domains waving in the breeze as 
far as anyone could tell from dissolving the crystals.  For the 
build-all-the-sidechains approach to be intellectually consistent, as Esko 
pointed out you must do that for all disordered loops and domains, and the N- 
and C-termini unless you have cast-iron evidence (MS on the crystals) that 
these atoms are indeed not present.  If you are working on a glycosylated 
protein, you need to model at the very least all of the first attached NAG and 
NAM and (for preference) run detailed MS on the crystal(s) to identify which 
sites have what levels of glycosylations and model out to the minimal 
identified conserved core. Complete intellectual consistency would require a 
detailed analysis of the fractional extensions at each glycosylation site and 
partial occupancies for those atoms.

At least in the academic circles I have worked in, even some modellers will 
insist on starting from a PDB model with high ADP sidechains as if there is an 
implied truth to them - and often (again my experience, YMMV) turn to you with 
a great cry of ah-ha! that the MD says that they are wrong, so what use is 
crystallography anyway?  Their complete absence sends a rather clear signal to 
the people using the model that the people solving the structure haven’t a clue 
as to where those atoms are.

Operationally, our job is to come up with an accurate model for the atoms that 
we can see. At least that’s my belief. Everything else we can leave to the 
hallucinations (I use the word advisedly) of AlphaFold2 and its offspring.  So 
let’s live and let live?  You put yours in with high ADPs: I take them out.

In both cases, the (protein structural) world is not going to come to an end.  
We can rely on the Voldemort of modern diplomacy for that.

Adrian

On 17 Mar 2023, at 11:50, Esko Oksanen 
<533d495de740-dmarc-requ...@jiscmail.ac.uk>
 wrote:

Dear Manfred,

In addition to personal preference I think there are some philosophical 
differences in what the model actually means. I would argue that what we see in 
the density is stronger evidence of the chemical identity than what we believe 
we have used as a starting material. After all we don’t know what reactions can 
happen during crystallisation or even data collection. I would also argue that 
the model should not primarily represent the chemical composition as such, but 
rather those atoms that are actually ordered. So if there is no electron 
density, it is not evidence of that particular atom being absent, but it that 
doesn’t mean we should model it. If one would apply the same logic to hydrogen 
atoms we should include them in the model since there is no evidence of their 
absence. Similarly we could explicitly model water molecules in the disordered 
solvent and just let them refine to high ADPs, since we have no evidence of the 
absence of water molecules. Or any disordered region of the protein such as 
termini. By not modelling these termini, hydrogens or water molecules we could 
also give the impression that they are not there, but I think the users of the 
models can and should be expected to understand the limitations of the model. I 
would find it more misleading to model things that we don’t see but assume to 
be there versus not modelling them. I think if the “end users" of the model are 
educated enough that they notice unphysically high ADPs they will also 
understand a truncated lysine side chain.

  Just my 2c,
  Esko

On 17 Mar 2023, at 09:57, Manfred S. Weiss 
mailto:manfred.we...@helmholtz-berlin.de>> 
wrote:


Dear all,

many views have been expressed in this thread, which I have been
following with great interest. Unfortunately, I have to say that many
of the views are more based on personal preferences, than on the
scientific evidence behind.

Here are some facts, that one may want to consider:

1. When you have crystallized your protein, and collected data
from it, the obvious assumption is that it is still intact. Unless you
have additional information, e.g. from mass spec, that some
side chains are cleaved off, they are there.

2. Absence of evidence (meaning missing electron density)
is NOT evidence of absence 

3. If you trim, you give the impression that the respective atoms
are not there. This means that the solvent model will be incorrect,
because the area will be defined as solvent.

In my view, the best approach is to build the side chains in their
most plausible conformation, or maybe in 2 or 3 or 5 different
conformations, and let the ADPs refine 

Re: [ccp4bb] To Trim or Not to To Trim

[Esko Oksanen]

Dear Manfred,

In addition to personal preference I think there are some philosophical 
differences in what the model actually means. I would argue that what we see in 
the density is stronger evidence of the chemical identity than what we believe 
we have used as a starting material. After all we don’t know what reactions can 
happen during crystallisation or even data collection. I would also argue that 
the model should not primarily represent the chemical composition as such, but 
rather those atoms that are actually ordered. So if there is no electron 
density, it is not evidence of that particular atom being absent, but it that 
doesn’t mean we should model it. If one would apply the same logic to hydrogen 
atoms we should include them in the model since there is no evidence of their 
absence. Similarly we could explicitly model water molecules in the disordered 
solvent and just let them refine to high ADPs, since we have no evidence of the 
absence of water molecules. Or any disordered region of the protein such as 
termini. By not modelling these termini, hydrogens or water molecules we could 
also give the impression that they are not there, but I think the users of the 
models can and should be expected to understand the limitations of the model. I 
would find it more misleading to model things that we don’t see but assume to 
be there versus not modelling them. I think if the “end users" of the model are 
educated enough that they notice unphysically high ADPs they will also 
understand a truncated lysine side chain.

  Just my 2c,
  Esko

On 17 Mar 2023, at 09:57, Manfred S. Weiss 
mailto:manfred.we...@helmholtz-berlin.de>> 
wrote:


Dear all,

many views have been expressed in this thread, which I have been
following with great interest. Unfortunately, I have to say that many
of the views are more based on personal preferences, than on the
scientific evidence behind.

Here are some facts, that one may want to consider:

1. When you have crystallized your protein, and collected data
from it, the obvious assumption is that it is still intact. Unless you
have additional information, e.g. from mass spec, that some
side chains are cleaved off, they are there.

2. Absence of evidence (meaning missing electron density)
is NOT evidence of absence 

3. If you trim, you give the impression that the respective atoms
are not there. This means that the solvent model will be incorrect,
because the area will be defined as solvent.

In my view, the best approach is to build the side chains in their
most plausible conformation, or maybe in 2 or 3 or 5 different
conformations, and let the ADPs refine freely.

All the best
Manfred
































































































--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseher
Geschäftsführung: Prof. Dr. Bernd Rech, Thomas Frederking

Sitz Berlin, AG Charlottenburg, 89 HRB 5583

Postadresse:
Hahn-Meitner-Platz 1
14109 Berlin
Deutschland

Diese E-Mail kann vertrauliche und/oder rechtlich geschützte Informationen 
enthalten. Wenn Sie diese E-Mail irrtümlich erhalten haben, informieren Sie 
bitte sofort den*die Absender*in und vernichten Sie diese Mail. Das unerlaubte 
Kopieren, die Veröffentlichung sowie die unbefugte Weitergabe dieser Mail ist 
nicht gestattet.
This email may contain confidential and/or proprietary information. If you have 
received this e-mail in error, please inform the sender immediately and destroy 
this e-mail. Unauthorized copying, publishing or distribution of this e-mail is 
not permitted.



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




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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Manfred S. Weiss]

Dear all,

many views have been expressed in this thread, which I have been
following with great interest. Unfortunately, I have to say that many
of the views are more based on personal preferences, than on the
scientific evidence behind.

Here are some facts, that one may want to consider:

1. When you have crystallized your protein, and collected data
from it, the obvious assumption is that it is still intact. Unless you
have additional information, e.g. from mass spec, that some
side chains are cleaved off, they are there.

2. Absence of evidence (meaning missing electron density)
is NOT evidence of absence 

3. If you trim, you give the impression that the respective atoms
are not there. This means that the solvent model will be incorrect,
because the area will be defined as solvent.

In my view, the best approach is to build the side chains in their
most plausible conformation, or maybe in 2 or 3 or 5 different
conformations, and let the ADPs refine freely.

All the best
Manfred
































































































--
Dr. Manfred S. Weiss
Macromolecular Crystallography
Helmholtz-Zentrum Berlin
Albert-Einstein-Str. 15
D-12489 Berlin



Helmholtz-Zentrum Berlin für Materialien und Energie GmbH

Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.

Aufsichtsrat: Vorsitzender Dr. Volkmar Dietz, stv. Vorsitzende Dr. Jutta 
Koch-Unterseher
Geschäftsführung: Prof. Dr. Bernd Rech, Thomas Frederking

Sitz Berlin, AG Charlottenburg, 89 HRB 5583

Postadresse:
Hahn-Meitner-Platz 1
14109 Berlin
Deutschland

Diese E-Mail kann vertrauliche und/oder rechtlich geschützte Informationen 
enthalten. Wenn Sie diese E-Mail irrtümlich erhalten haben, informieren Sie 
bitte sofort den*die Absender*in und vernichten Sie diese Mail. Das unerlaubte 
Kopieren, die Veröffentlichung sowie die unbefugte Weitergabe dieser Mail ist 
nicht gestattet.
This email may contain confidential and/or proprietary information. If you have 
received this e-mail in error, please inform the sender immediately and destroy 
this e-mail. Unauthorized copying, publishing or distribution of this e-mail is 
not permitted.



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] [Sender Not Verified] Re: [ccp4bb] To Trim or Not to To Trim

[Robbie Joosten]

Hi Markus,

Just to make sure that things are clear: PDB-REDO adds missing side chains 
(that is a design choice) and it also adds missing loops if, and only if, there 
is a homologous template, there is sufficient density (although the criteria 
are rather forgiving), and the geometry of the loop stays okay upon real-space 
refinement. We do not build termini.

Your water example is a very nice one, this also happens for main chain atoms 
unfortunately. So unlike building parts of the protein that are chemically 
attached, please be careful with building waters. Only build ones that you are 
sure are waters. If you know that it is something else, but you don’t know 
what, please don’t build anything. I actually encourage students to build 
waters by hand. It is really simple and fast in Coot and it takes though a tour 
of you structure model enabling you to spot other issues.

HTH,
Robbie

From: CCP4 bulletin board  On Behalf Of Rudolph, Markus
Sent: Tuesday, March 14, 2023 15:58
To: CCP4BB@JISCMAIL.AC.UK
Subject: Re: [ccp4bb] [Sender Not Verified] Re: [ccp4bb] To Trim or Not to To 
Trim

Hello Adrian,

Provided a crystal contains the intact protein, as measured by mass 
spectrometry of a washed and dissolved crystal, leaving out side-chains makes 
little sense, even if some of them are destroyed by X-rays during data 
collection. I realize that most of the time, no such analysis is done on 
crystals, but if you assume all residues are there, then they must fit into the 
asymmetric unit. This includes N- and C-terminal residues, loops, and any 
side-chains without density. If side-chains won't fit, e.g. by crashing into 
symmetry neighbors, then the main-chain trace is very likely wrong. Thus, 
placing side-chains in a model is a good test for a correct main-chain trace. I 
understand many protein architects out there omit the terminal extensions and 
longer loops, and so do I. pdb_redo will automatically complete a model, so one 
could say - don't bother, as a user I'll retrieve the re-refined model from 
there. This is not my personal choice of model building, but a practical option 
from the viewpoint of downstream users. Sometimes, however, I see models where 
side-chains are truncated but the little density still present for the 
side-chain is interpreted by water (see image attached, in which case the water 
comes from a symmetry-related position). If one says "don't build what you 
don't see but build what the data tell you", a well-intended but 
mis-interpreted water molecule may occupy the density belonging to a truncated 
side-chain. I as a user would be confused by this, especially when using such 
models for ligand binding studies where water is important. In our SBDD 
projects, we always take B-values into account, and we welcome complete models 
or we complete and re-refine them ourselves. So yes, my 12 points go to the 
"let the B-factors take care of it” song. But really, I'm thankful for _any_ 
published structure, even with sup-optimal parameterization, experienced 
structural biologists can deal with that and communicate the information to 
their collaborators.


...which brings up the other evergreen, whether the PDB should be more strict 
about ...



Best wishes,
Markus



On Fri, Mar 10, 2023 at 5:33 PM Goldman, Adrian 
mailto:adrian.gold...@helsinki.fi>> wrote:
Maybe simplest just to trim it back. I do worry that the presence of a wrong 
conformation will lead to inaccurate vdw clashes that could negatively affect 
other atoms.

Sent from my iPhone

> On 10 Mar 2023, at 18:25, Phil Jeffrey 
> mailto:pjeff...@princeton.edu>> wrote:
>
> On 3/10/23 4:05 AM, Julia Griese wrote:
>> Hi all,
>> My impression has been that the most common approach these days is to “let 
>> the B-factors take care of it”, but I might be wrong. Maybe it’s time to run 
>> another poll?
>> Personally, I call any other approach R-factor cosmetics. The goal in model 
>> building is not to achieve the lowest possible R-factors, it’s to build the 
>> most physically meaningful, most likely to be correct, model.
>
> And I could call your approach "model cosmetics".
>
> If you can't see the side-chain, you don't know where it is and you probably 
> don't even know where the centroid of the distribution is. Only in the case 
> of very short side-chains with few rotamers can you make a reasonable volume 
> approximation to where the side-chain is and "let the B-factors" smear out 
> the density to cover a range of the projected conformations.
>
> For longer side-chains, if you put it in a single conformation, you are very 
> likely NOT coming close to correctly modeling the actual distribution of the 
> conformations.  So let's circle back on "most likely to be correct model" and 
> ask what we *actually* know about where the atoms are.
>
> Put your disordered

Re: [ccp4bb] To Trim or Not to To Trim

[Tom Peat]

Gotta love the extra wood on the fire.
In the bad old days, I agree, we had no real idea.
But these days I usually have mass spec done on most everything and 
additionally look at different crystal forms. Often I find that in one space 
group more will be seen than in another (and that one isn't necessarily the 
highest resolution) and these are even often from the same plate/ screen. 
Additionally, many proteins form oligomers or are found in multiple copies in 
the asymmetric unit, and these again vary as to how much of that mystical 
N-terminus or C-terminus one sees. I would be pretty confident that it isn't 
just one protomer in a trimer/ tetramer that has that extension and the rest of 
have selectively proteolysed...
Good point though. cheers, tom

From: CCP4 bulletin board  on behalf of Jurgen Bosch 

Sent: Saturday, March 11, 2023 8:37 AM
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] To Trim or Not to To Trim

You don't often get email from jxb...@case.edu. Learn why this is 
important<https://aka.ms/LearnAboutSenderIdentification>
Well Tom,

Your missing N-terminus might just be degraded same with the C-terminus two do 
you know it is actually there waving at you? Do you have mass spec evidence for 
the full existence of your protein? You might just have crystalized that 
magical fragment thinking that your full length construct is actually 
crystallized.

Just throwing some wood into the fire …

Jürgen

On Mar 10, 2023, at 4:01 PM, Debanu Das  wrote:

Hi Tom,

-"so no matter how much we try to educate people, the vast (vast, vast) 
majority of people will take the models as the 'truth'
-"So if we don't see something, the conservative approach is to probably avoid 
putting it in, otherwise it will get propagated forever"

I absolutely agree. I have a recent anecdote about this to share. I was at a 
workshop where the majority of attendees were medicinal chemists, comp 
chemists, biologists. I found out that some (many?) were even performing comp 
chem studies (where ligands and waters were important) on structures downloaded 
straight from the PDB without even checking associated electron density for 
such ligands/waters or even realizing potential side chain issues (not even 
including about any potential main chain or other quality/validation issues).

Best regards,
Debanu
--
Debanu Das,
https://bio.site/debanu_das

On Fri, Mar 10, 2023 at 12:41 PM Tom Peat 
mailto:t.p...@unsw.edu.au>> wrote:
Hello All,

I agree with Dale that we don't have a good way to model these things and it 
has been a discussion for a very long time with no proper answer.

Two more small points- we (or at least I do this all the time) don't model the 
N- or C-terminal residues that I don't see. They are most likely there 
somewhere (waving goodbye in the solvent mask) and that seems to be 
(relatively) standard practice, so I'm not sure how different this is to not 
putting in side chains that can't be seen? We don't replace with Ala, so people 
should know what the actual sequence is, but there is no evidence for a side 
chain. A multi-model is likely the better way to go, but isn't as feasible 
currently as the single model we normally deposit. As mentioned by others, we 
shouldn't try to put in ligands or co-factors that we don't see either...

One other point- although we should educate, the PDB has estimated that 99% of 
the users are not structural biologists (nor modellers or others with 
training), so no matter how much we try to educate people, the vast (vast, 
vast) majority of people will take the models as the 'truth'. So if we don't 
see something, the conservative approach is to probably avoid putting it in, 
otherwise it will get propagated forever.

My two cents. cheers, tom

From: CCP4 bulletin board mailto:CCP4BB@JISCMAIL.AC.UK>> 
on behalf of Dale Tronrud mailto:de...@daletronrud.com>>
Sent: Saturday, March 11, 2023 7:15 AM
To: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> 
mailto:CCP4BB@JISCMAIL.AC.UK>>
Subject: Re: [ccp4bb] To Trim or Not to To Trim

Hi

As a frequent contributor to prior discussions on this same topic I
would like to broaden the discussion a bit.  I'm sorry to say that most
of the comments on this threat are exactly the same positions that have
been expressed many times over the years.  I don't want to spent time,
again, retyping my opinions on how I prefer to torment the parameters of
my models to express what I believe is going on inside my crystals.

The fundamental problem is that the parameters we are forced to use,
in our PDB depositions and in the refinement and model building programs
available to us, are wholly inadequate.  We cannot accurately (or
precisely) describe what we are are envisioning for the surface side
chains, and sometimes entire stretches of main chain, in our proteins.
We can continue to argue with each other year after year, 

Re: [ccp4bb] To Trim or Not to To Trim

[Quyen Hoang]

Hi Debanu,In your example, I would guess that models with truncated side chains would probably be problematic too, so would alternative conformations, unmodelled segments, asymmetric unit consisting of only a fraction of the biological/functional unit, etc.I am not sure if it is possible to build a model in a way that would avoid all the potential pitfalls in the circumstance you mentioned. Or should one even try?I don't remember ever solving a structure for the purpose of providing a model for computational chemists to use. We solve structures to answer certain questions and I would rather deposit the models we used to gain the insight we seeked instead of the one artificially made in the attempt to avoid misuses or understanding by non-structural biologists. Cheers,QuyenOn Mar 10, 2023, at 4:01 PM, Debanu Das  wrote:Hi Tom,-"so no matter how much we try to educate people, the vast (vast, vast) majority of people will take
 the models as the 'truth'-"So if we don't see something, the conservative approach is to probably 
avoid putting it in, otherwise it will get propagated forever"I absolutely agree. I have a recent anecdote about this to share. I was at a workshop where the majority of attendees were medicinal chemists, comp chemists, biologists. I found out that some (many?) were even performing comp chem studies (where ligands and waters were important) on structures downloaded straight from the PDB without even checking associated electron density for such ligands/waters or even realizing potential side chain issues (not even including about any potential main chain or other quality/validation issues).   Best regards,Debanu--Debanu Das,https://bio.site/debanu_dasOn Fri, Mar 10, 2023 at 12:41 PM Tom Peat <t.p...@unsw.edu.au> wrote:






Hello All, 




I agree with Dale that we don't have a good way to model these things and it has been a discussion for a very long time with no proper answer. 




Two more small points- we (or at least I do this all the time) don't model the N- or C-terminal residues that I don't see. They are most likely there somewhere (waving goodbye in the solvent mask) and that seems to be (relatively) standard practice, so I'm
 not sure how different this is to not putting in side chains that can't be seen? We don't replace with Ala, so people should know what the actual sequence is, but there is no evidence for a side chain. A multi-model is likely the better way to go, but isn't
 as feasible currently as the single model we normally deposit. As mentioned by others, we shouldn't try to put in ligands or co-factors that we don't see either... 




One other point- although we should educate, the PDB has estimated that 99% of the users are not structural biologists (nor modellers or others with training), so no matter how much we try to educate people, the vast (vast, vast) majority of people will take
 the models as the 'truth'. So if we don't see something, the conservative approach is to probably avoid putting it in, otherwise it will get propagated forever. 




My two cents. cheers, tom 


From: CCP4 bulletin board <CCP4BB@JISCMAIL.AC.UK> on behalf of Dale Tronrud <de...@daletronrud.com>
Sent: Saturday, March 11, 2023 7:15 AM
To: CCP4BB@JISCMAIL.AC.UK <CCP4BB@JISCMAIL.AC.UK>
Subject: Re: [ccp4bb] To Trim or Not to To Trim
 


Hi

    As a frequent contributor to prior discussions on this same topic I 
would like to broaden the discussion a bit.  I'm sorry to say that most 
of the comments on this threat are exactly the same positions that have 
been expressed many times over the years.  I don't want to spent time, 
again, retyping my opinions on how I prefer to torment the parameters of 
my models to express what I believe is going on inside my crystals.

    The fundamental problem is that the parameters we are forced to use, 
in our PDB depositions and in the refinement and model building programs 
available to us, are wholly inadequate.  We cannot accurately (or 
precisely) describe what we are are envisioning for the surface side 
chains, and sometimes entire stretches of main chain, in our proteins. 
We can continue to argue with each other year after year, but there is 
no solution to this problem other than changing the nature of PDB models 
and allowing a reasonable description of multi-conformation models.

    I believe it is fair to say that the consensus after a previous 
round of this discussion was that, at the very least, we need a flag for 
each atom which indicates whether that atom was placed based on electron 
density or simply to make a chemically complete set of atoms for that 
type of monomer.  I haven't looked but I think that was about five or 
ten years ago.  Since then the PDB has made major changes to the 
structure of PDB entries that will require most software for analysis of 
macromolecular models be rewritten and right now that organization is 
making a major push to get us to virtually attend a wo

Re: [ccp4bb] To Trim or Not to To Trim

[CCP4BB]

: “Don’t blindly trust 
>> anything.”
>> Cheers
>> /Julia
>> -- 
>> Dr. Julia Griese
>> Associate Professor (Docent)
>> Principal Investigator
>> Department of Cell and Molecular Biology
>> Uppsala University
>> BMC, Box 596
>> SE-75124 Uppsala
>> Sweden
>> email: julia.gri...@icm.uu.se
>> phone: +46-(0)18-471 4982
>> http://www.icm.uu.se/structural-biology/griese-lab/ 
>> <http://www.icm.uu.se/structural-biology/griese-lab/>
>> *From: *CCP4 bulletin board  on behalf of Bernhard 
>> Lechtenberg <968307750321-dmarc-requ...@jiscmail.ac.uk>
>> *Reply-To: *Bernhard Lechtenberg 
>> *Date: *Friday, March 10, 2023 at 05:07
>> *To: *"CCP4BB@JISCMAIL.AC.UK" 
>> *Subject: *Re: [ccp4bb] To Trim or Not to To Trim
>> I found the poll I wrote about earlier. This actually is way older than I 
>> had expected (2011). You can see the poll results (which was run by Ed 
>> Pozharski) and discussion at the time here in the CCP4BB archive:
>> https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html 
>> <https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html>
>> In brief, the results of 240 respondents were:
>> Delete the atoms 43%
>> Let refinement take care of it by inflating B-factors41%
>> Set occupancy to zero12%
>> Other 4%
>> Bernhard
>> *From: *CCP4 bulletin board  on behalf of Debanu Das 
>> 
>> *Date: *Friday, 10 March 2023 at 2:56 pm
>> *To: *CCP4BB@JISCMAIL.AC.UK 
>> *Subject: *Re: [ccp4bb] To Trim or Not to To Trim
>> We dealt with this in-depth during structural genomics days when we 
>> deposited over 1500 novel, high-quality, experimentally-phased structures 
>> into the PDB. Think it’s prudent to trim/truncate side chains without 
>> reliable density.
>> Non-structural biologists using PDB structures without expert help can err 
>> in any of these scenarios: misinterpreting most common/random rotamer, zero 
>> occupancy atoms, B-factors, etc.
>> What is the value of populating the PDB, which is a structural model 
>> repository, with such information that is not there, i.e., reliable 
>> structural model?
>> Any trained crystallographer/structural biologist can easily add in side 
>> chain information if needed for modeling/computational chemistry reasons.
>> Best regards,
>> Debanu
>> On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch > <mailto:jxb...@case.edu>> wrote:
>>I’d say no trimming to side chains for the following reason: There
>>are non-structural biologists using PDB files and if atoms are
>>missing they don’t know what to do. A better approach is where no
>>side chain density allows support of placement, pick the most common
>>rotamer and set the occupancy to zero for those atoms lacking
>>density support. More work for you but more accurate in my opinion.
>>Jürgen
>>___
>>Jürgen Bosch, PhD, MBA
>>Center for Global Health & Diseases
>>Case Western Reserve University
>>Cleveland, OH 44106
>>https://www.linkedin.com/in/jubosch/
>><https://www.linkedin.com/in/jubosch/>
>>CEO & Co-Founder at InterRayBio, LLC
>>On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg
>><968307750321-dmarc-requ...@jiscmail.ac.uk
>><mailto:968307750321-dmarc-requ...@jiscmail.ac.uk>> wrote:
>>Hi Rhys,
>>I am also all for leaving side chains and letting the B-factors
>>deal with the weak/absent density.
>>I don’t think there is a consensus, but I kind of remember that
>>somebody did a poll a few years ago and if I remember correctly
>>the main approaches were the one described above, or trimming
>>the side-chains.
>>Bernhard
>>*Bernhard C. Lechtenberg* PhD
>>NHMRC Emerging Leadership Fellow
>>Laboratory Head
>>Ubiquitin Signalling Division​​
>>E lechtenber...@wehi.edu.au <mailto:lechtenber...@wehi.edu.au>
>>T +61 3 9345 2217
>>*From: *CCP4 bulletin board ><mailto:CCP4BB@JISCMAIL.AC.UK>> on behalf of Rhys Grinter
>><22087c81e8c6-dmarc-requ...@jiscmail.ac.uk
>><mailto:22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>&

Re: [ccp4bb] To Trim or Not to To Trim

[Tom Peat]

Hello All,

I agree with Dale that we don't have a good way to model these things and it 
has been a discussion for a very long time with no proper answer.

Two more small points- we (or at least I do this all the time) don't model the 
N- or C-terminal residues that I don't see. They are most likely there 
somewhere (waving goodbye in the solvent mask) and that seems to be 
(relatively) standard practice, so I'm not sure how different this is to not 
putting in side chains that can't be seen? We don't replace with Ala, so people 
should know what the actual sequence is, but there is no evidence for a side 
chain. A multi-model is likely the better way to go, but isn't as feasible 
currently as the single model we normally deposit. As mentioned by others, we 
shouldn't try to put in ligands or co-factors that we don't see either...

One other point- although we should educate, the PDB has estimated that 99% of 
the users are not structural biologists (nor modellers or others with 
training), so no matter how much we try to educate people, the vast (vast, 
vast) majority of people will take the models as the 'truth'. So if we don't 
see something, the conservative approach is to probably avoid putting it in, 
otherwise it will get propagated forever.

My two cents. cheers, tom

From: CCP4 bulletin board  on behalf of Dale Tronrud 

Sent: Saturday, March 11, 2023 7:15 AM
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] To Trim or Not to To Trim

Hi

As a frequent contributor to prior discussions on this same topic I
would like to broaden the discussion a bit.  I'm sorry to say that most
of the comments on this threat are exactly the same positions that have
been expressed many times over the years.  I don't want to spent time,
again, retyping my opinions on how I prefer to torment the parameters of
my models to express what I believe is going on inside my crystals.

The fundamental problem is that the parameters we are forced to use,
in our PDB depositions and in the refinement and model building programs
available to us, are wholly inadequate.  We cannot accurately (or
precisely) describe what we are are envisioning for the surface side
chains, and sometimes entire stretches of main chain, in our proteins.
We can continue to argue with each other year after year, but there is
no solution to this problem other than changing the nature of PDB models
and allowing a reasonable description of multi-conformation models.

I believe it is fair to say that the consensus after a previous
round of this discussion was that, at the very least, we need a flag for
each atom which indicates whether that atom was placed based on electron
density or simply to make a chemically complete set of atoms for that
type of monomer.  I haven't looked but I think that was about five or
ten years ago.  Since then the PDB has made major changes to the
structure of PDB entries that will require most software for analysis of
macromolecular models be rewritten and right now that organization is
making a major push to get us to virtually attend a workshop to help us
make this transition.  And yet I don't think there is anything in this
new data dictionary to help us with this important but intractable
problem.

Unless the PDB gives us the parameters we need to properly describe
a macromolecular model, and the refinement/model building developers
give us the tools to make use of them, we will be back here again, every
five years or so, rehashing this debate over exactly the same,
irreconcilably poor, solutions to this problem.

Dale E. Tronrud


On 3/10/2023 1:05 AM, Julia Griese wrote:
> Hi all,
>
> My impression has been that the most common approach these days is to
> “let the B-factors take care of it”, but I might be wrong. Maybe it’s
> time to run another poll?
>
> Personally, I call any other approach R-factor cosmetics. The goal in
> model building is not to achieve the lowest possible R-factors, it’s to
> build the most physically meaningful, most likely to be correct, model.
> So if you know that the side chain is part of the protein, you should
> model it the best way you can. If it’s there, just disordered, then the
> most correct way to model it is to let it have high B-factors. Most
> molecular graphics programs don’t flag zero-occupancy atoms, so the user
> might never notice. Truncation of a side chain, unless there is evidence
> that it really physically isn’t there, is also misleading, in my
> opinion. I don’t believe that it is more helpful to the non-expert user
> than high B-factors either.
>
> If people who are not structural biologists themselves don’t know how to
> use a structure, then we need to educate them better. It is very
> straightforward these days to look at electron density in the PDB
> viewer. It used to be difficult, but nowadays there’s no excuse for not
> checking the electron den

Re: [ccp4bb] To Trim or Not to To Trim

[Dale Tronrud]

Hi

   As a frequent contributor to prior discussions on this same topic I 
would like to broaden the discussion a bit.  I'm sorry to say that most 
of the comments on this threat are exactly the same positions that have 
been expressed many times over the years.  I don't want to spent time, 
again, retyping my opinions on how I prefer to torment the parameters of 
my models to express what I believe is going on inside my crystals.


   The fundamental problem is that the parameters we are forced to use, 
in our PDB depositions and in the refinement and model building programs 
available to us, are wholly inadequate.  We cannot accurately (or 
precisely) describe what we are are envisioning for the surface side 
chains, and sometimes entire stretches of main chain, in our proteins. 
We can continue to argue with each other year after year, but there is 
no solution to this problem other than changing the nature of PDB models 
and allowing a reasonable description of multi-conformation models.


   I believe it is fair to say that the consensus after a previous 
round of this discussion was that, at the very least, we need a flag for 
each atom which indicates whether that atom was placed based on electron 
density or simply to make a chemically complete set of atoms for that 
type of monomer.  I haven't looked but I think that was about five or 
ten years ago.  Since then the PDB has made major changes to the 
structure of PDB entries that will require most software for analysis of 
macromolecular models be rewritten and right now that organization is 
making a major push to get us to virtually attend a workshop to help us 
make this transition.  And yet I don't think there is anything in this 
new data dictionary to help us with this important but intractable 
problem.


   Unless the PDB gives us the parameters we need to properly describe 
a macromolecular model, and the refinement/model building developers 
give us the tools to make use of them, we will be back here again, every 
five years or so, rehashing this debate over exactly the same, 
irreconcilably poor, solutions to this problem.


Dale E. Tronrud


On 3/10/2023 1:05 AM, Julia Griese wrote:

Hi all,

My impression has been that the most common approach these days is to 
“let the B-factors take care of it”, but I might be wrong. Maybe it’s 
time to run another poll?


Personally, I call any other approach R-factor cosmetics. The goal in 
model building is not to achieve the lowest possible R-factors, it’s to 
build the most physically meaningful, most likely to be correct, model. 
So if you know that the side chain is part of the protein, you should 
model it the best way you can. If it’s there, just disordered, then the 
most correct way to model it is to let it have high B-factors. Most 
molecular graphics programs don’t flag zero-occupancy atoms, so the user 
might never notice. Truncation of a side chain, unless there is evidence 
that it really physically isn’t there, is also misleading, in my 
opinion. I don’t believe that it is more helpful to the non-expert user 
than high B-factors either.


If people who are not structural biologists themselves don’t know how to 
use a structure, then we need to educate them better. It is very 
straightforward these days to look at electron density in the PDB 
viewer. It used to be difficult, but nowadays there’s no excuse for not 
checking the electron density. The PDB validation flags RSRZ outliers. 
You can easily colour a structure by B-factors. It doesn’t take that 
much effort to teach students how to validate structures. The main point 
you need to get across is that it is necessary to do so. And this needs 
to be done not only in courses aimed at prospective experimental 
structural biologists, of course, but whenever students use structures 
in any way.


This is just the opinion of someone who feels very strongly about 
teaching structure validation and rejoices when students’ reply to the 
question “What was the most important thing you learned today?” is: 
“Don’t blindly trust anything.”


Cheers

/Julia

--

Dr. Julia Griese

Associate Professor (Docent)

Principal Investigator

Department of Cell and Molecular Biology

Uppsala University

BMC, Box 596

SE-75124 Uppsala

Sweden

email: julia.gri...@icm.uu.se

phone: +46-(0)18-471 4982

http://www.icm.uu.se/structural-biology/griese-lab/ 
<http://www.icm.uu.se/structural-biology/griese-lab/>


*From: *CCP4 bulletin board  on behalf of 
Bernhard Lechtenberg <968307750321-dmarc-requ...@jiscmail.ac.uk>

*Reply-To: *Bernhard Lechtenberg 
*Date: *Friday, March 10, 2023 at 05:07
*To: *"CCP4BB@JISCMAIL.AC.UK" 
*Subject: *Re: [ccp4bb] To Trim or Not to To Trim

I found the poll I wrote about earlier. This actually is way older than 
I had expected (2011). You can see the poll results (which was run by Ed 
Pozharski) and discussion at the time here in the CCP4BB archive:


https://www.mail-archive.com/ccp4bb@jiscma

Re: [ccp4bb] To Trim or Not to To Trim

[Julia Griese]

Hi Phil,

I don't think that this is model cosmetics, but I certainly didn't claim that 
this isn't controversial. That's why we're having this discussion again and 
again after all.

Have you tried to model a disordered Arg with 10 (why only 10?) alternate 
rotamers? Did it refine to something sensible? If so, then I would agree that 
this would be better. I haven't tried it, but I suspect they would all converge 
more or less on top of each other. (If anyone wants to give it a try, please 
let me know what happened!)

I still maintain that modeling a disordered Arg as an Arg with a single rotamer 
with high B-factors is more correct than modeling it as an Ala or setting its 
side chain occupancy to zero, because it is not an Ala, and its side chain 
atoms are not absent. (Yes, I see Harry's point, but let's assume you've 
collected your data carefully and don't have such serious radiation damage that 
you've lost arginine side chains. If that was the case, you'd be missing a 
whole lot of other side chains too.) I do not claim that such a model 
accurately reflects where the Arg really is, I only claim that it's the best we 
can do (at present). Hence why we need to educate users.

And absolutely, none of the current solutions is ideal, that's also why we're 
having this discussion again and again. It seems no one has found a better 
solution yet.

/Julia

On 3/10/23, 17:15, "Phil Jeffrey" mailto:pjeff...@princeton.edu>> wrote:


On 3/10/23 4:05 AM, Julia Griese wrote:
> Hi all,
>
> My impression has been that the most common approach these days is to
> “let the B-factors take care of it”, but I might be wrong. Maybe it’s
> time to run another poll?
>
> Personally, I call any other approach R-factor cosmetics. The goal in
> model building is not to achieve the lowest possible R-factors, it’s to
> build the most physically meaningful, most likely to be correct, model.


And I could call your approach "model cosmetics".


If you can't see the side-chain, you don't know where it is and you
probably don't even know where the centroid of the distribution is.
Only in the case of very short side-chains with few rotamers can you
make a reasonable volume approximation to where the side-chain is and
"let the B-factors" smear out the density to cover a range of the
projected conformations.


For longer side-chains, if you put it in a single conformation, you are
very likely NOT coming close to correctly modeling the actual
distribution of the conformations. So let's circle back on "most likely
to be correct model" and ask what we *actually* know about where the
atoms are.


Put your disordered Arg in with 10 alternate conformations, each with a
refined relative occupancy, and then let the B-factors smear that lot
out, and that's your better model.


Phil Jeffrey
Princeton






VARNING: Klicka inte på länkar och öppna inte bilagor om du inte känner igen 
avsändaren och vet att innehållet är säkert.
CAUTION: Do not click on links or open attachments unless you recognise the 
sender and know the content is safe.











När du har kontakt med oss på Uppsala universitet med e-post så innebär det att 
vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du 
läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/

E-mailing Uppsala University means that we will process your personal data. For 
more information on how this is performed, please read here: 
http://www.uu.se/en/about-uu/data-protection-policy



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] To Trim or Not to To Trim

[Julia Griese]

Surely not if the locations of those other atoms are strongly supported by 
density? And surely you would always select a rotamer that does not clash with 
its surroundings?

On 3/10/23, 17:32, "CCP4 bulletin board on behalf of Goldman, Adrian" 
mailto:CCP4BB@JISCMAIL.AC.UK> on behalf of 
adrian.gold...@helsinki.fi > wrote:


Maybe simplest just to trim it back. I do worry that the presence of a wrong 
conformation will lead to inaccurate vdw clashes that could negatively affect 
other atoms.


Sent from my iPhone


> On 10 Mar 2023, at 18:25, Phil Jeffrey  > wrote:
>
> On 3/10/23 4:05 AM, Julia Griese wrote:
>> Hi all,
>> My impression has been that the most common approach these days is to “let 
>> the B-factors take care of it”, but I might be wrong. Maybe it’s time to run 
>> another poll?
>> Personally, I call any other approach R-factor cosmetics. The goal in model 
>> building is not to achieve the lowest possible R-factors, it’s to build the 
>> most physically meaningful, most likely to be correct, model.
>
> And I could call your approach "model cosmetics".
>
> If you can't see the side-chain, you don't know where it is and you probably 
> don't even know where the centroid of the distribution is. Only in the case 
> of very short side-chains with few rotamers can you make a reasonable volume 
> approximation to where the side-chain is and "let the B-factors" smear out 
> the density to cover a range of the projected conformations.
>
> For longer side-chains, if you put it in a single conformation, you are very 
> likely NOT coming close to correctly modeling the actual distribution of the 
> conformations. So let's circle back on "most likely to be correct model" and 
> ask what we *actually* know about where the atoms are.
>
> Put your disordered Arg in with 10 alternate conformations, each with a 
> refined relative occupancy, and then let the B-factors smear that lot out, 
> and that's your better model.
>
> Phil Jeffrey
> Princeton
>
> 
>
> 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/ 



VARNING: Klicka inte på länkar och öppna inte bilagor om du inte känner igen 
avsändaren och vet att innehållet är säkert.
CAUTION: Do not click on links or open attachments unless you recognise the 
sender and know the content is safe.











När du har kontakt med oss på Uppsala universitet med e-post så innebär det att 
vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du 
läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/

E-mailing Uppsala University means that we will process your personal data. For 
more information on how this is performed, please read here: 
http://www.uu.se/en/about-uu/data-protection-policy



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] To Trim or Not to To Trim

[Quyen Hoang]

As with Jurgen, I’ve never trimmed a residue.
In my view the trimmed residues do not exist.
If we were to build models consisting of only atoms defined by the experimental 
density, then I wonder what the original model of the DNA double helix would 
have looked like.

BTW, hey Jurgen, long time no talk.

Cheers,
Quyen

Quyen Hoang, PhD
Professor of Biochemistry and Molecular Biology
Director of IUSM Center for Electron Microscopy (iCEM)
Adjunct Professor of Neurology
Primary Investigator of the Stark Neuroscience Research Institute
Indiana University School of Medicine
635 Barnhill Drive, MS0013C
Indianapolis, IN, 46202
(317)274-4371
https://qqhoang.pages.iu.edu


> On Mar 10, 2023, at 12:09 PM, Jurgen Bosch  wrote:
> 
> I’m sure James H. Is preparing a philosophical dissertation on the "State of 
> the atoms to B or not to B that is not only a refinement question” that he 
> will share momentarily with the board.
> 
> Jürgen 
> 
>> On Mar 10, 2023, at 12:06 PM, DEBANU DAS  
>> wrote:
>> 
>> Yes, zero occupancy would reflect that. But not the coordinates in any 
>> proper way. 
>> Debanu
>> 
>> On Fri, Mar 10, 2023 at 8:58 AM Jurgen Bosch > > wrote:
>> Going back to RIP phasing methods :-)
>> So Harry in your particular case occupancy of zero would actually reflect 
>> reality for those “combusted” atoms.
>> 
>> Jürgen 
>> 
>> > On Mar 10, 2023, at 11:56 AM, Harry Powell 
>> > <193323b1e616-dmarc-requ...@jiscmail.ac.uk 
>> > > wrote:
>> > 
>> > Hi folks
>> > 
>> > One other thing that I haven’t noticed anyone mentioning yet (sorry to 
>> > those who have mentioned it!!) is that you may not see your sidechain 
>> > atoms in density because they are not there at all, in spite of what you 
>> > may have had in the original protein, or even if the atoms were really 
>> > there in the crystal _before_ exposure to the beam.
>> > 
>> > The coordinates are supposed to be what you actually find, not what you 
>> > hope is there.
>> > 
>> > Just my two ha’porth
>> > 
>> > 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 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/ 
>> 
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1 
> 



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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Jurgen Bosch]

I’m sure James H. Is preparing a philosophical dissertation on the "State of 
the atoms to B or not to B that is not only a refinement question” that he will 
share momentarily with the board.

Jürgen 

> On Mar 10, 2023, at 12:06 PM, DEBANU DAS  
> wrote:
> 
> Yes, zero occupancy would reflect that. But not the coordinates in any proper 
> way. 
> Debanu
> 
> On Fri, Mar 10, 2023 at 8:58 AM Jurgen Bosch  > wrote:
>> Going back to RIP phasing methods :-)
>> So Harry in your particular case occupancy of zero would actually reflect 
>> reality for those “combusted” atoms.
>> 
>> Jürgen 
>> 
>> > On Mar 10, 2023, at 11:56 AM, Harry Powell 
>> > <193323b1e616-dmarc-requ...@jiscmail.ac.uk 
>> > > wrote:
>> > 
>> > Hi folks
>> > 
>> > One other thing that I haven’t noticed anyone mentioning yet (sorry to 
>> > those who have mentioned it!!) is that you may not see your sidechain 
>> > atoms in density because they are not there at all, in spite of what you 
>> > may have had in the original protein, or even if the atoms were really 
>> > there in the crystal _before_ exposure to the beam.
>> > 
>> > The coordinates are supposed to be what you actually find, not what you 
>> > hope is there.
>> > 
>> > Just my two ha’porth
>> > 
>> > 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 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/




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] To Trim or Not to To Trim

[Harry Powell]

Hi Jürgen

You might think so, but I’d disagree. Not going too far away from your line of 
reasoning I could also put in a completely fictitious ligand or cofactor and 
assign its occupancies to zero (I really, really knew it was there but I just 
couldn’t find any evidence… :-))

Harry

> On 10 Mar 2023, at 16:58, Jurgen Bosch  wrote:
> 
> Going back to RIP phasing methods :-)
> So Harry in your particular case occupancy of zero would actually reflect 
> reality for those “combusted” atoms.
> 
> Jürgen 
> 
>> On Mar 10, 2023, at 11:56 AM, Harry Powell 
>> <193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
>> 
>> Hi folks
>> 
>> One other thing that I haven’t noticed anyone mentioning yet (sorry to those 
>> who have mentioned it!!) is that you may not see your sidechain atoms in 
>> density because they are not there at all, in spite of what you may have had 
>> in the original protein, or even if the atoms were really there in the 
>> crystal _before_ exposure to the beam.
>> 
>> The coordinates are supposed to be what you actually find, not what you hope 
>> is there.
>> 
>> Just my two ha’porth
>> 
>> 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 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/

Re: [ccp4bb] To Trim or Not to To Trim

[Debanu Das]

Yes, zero occupancy would reflect that. But not the coordinates in any
proper way. Then what would be the point of associating occupancy and
B-factors with totally incorrect coordinates?
Debanu

On Fri, Mar 10, 2023 at 8:58 AM Jurgen Bosch  wrote:

> Going back to RIP phasing methods :-)
> So Harry in your particular case occupancy of zero would actually reflect
> reality for those “combusted” atoms.
>
> Jürgen
>
> > On Mar 10, 2023, at 11:56 AM, Harry Powell <
> 193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
> >
> > Hi folks
> >
> > One other thing that I haven’t noticed anyone mentioning yet (sorry to
> those who have mentioned it!!) is that you may not see your sidechain atoms
> in density because they are not there at all, in spite of what you may have
> had in the original protein, or even if the atoms were really there in the
> crystal _before_ exposure to the beam.
> >
> > The coordinates are supposed to be what you actually find, not what you
> hope is there.
> >
> > Just my two ha’porth
> >
> > 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 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/
>
-- 
---
LinkedIn: www.linkedin.com/in/debanudas
Cal Alumni: cal.berkeley.edu/debanudas



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] To Trim or Not to To Trim

[Jurgen Bosch]

Going back to RIP phasing methods :-)
So Harry in your particular case occupancy of zero would actually reflect 
reality for those “combusted” atoms.

Jürgen 

> On Mar 10, 2023, at 11:56 AM, Harry Powell 
> <193323b1e616-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi folks
> 
> One other thing that I haven’t noticed anyone mentioning yet (sorry to those 
> who have mentioned it!!) is that you may not see your sidechain atoms in 
> density because they are not there at all, in spite of what you may have had 
> in the original protein, or even if the atoms were really there in the 
> crystal _before_ exposure to the beam.
> 
> The coordinates are supposed to be what you actually find, not what you hope 
> is there.
> 
> Just my two ha’porth
> 
> 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 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/

Re: [ccp4bb] To Trim or Not to To Trim

[Harry Powell]

Hi folks

One other thing that I haven’t noticed anyone mentioning yet (sorry to those 
who have mentioned it!!) is that you may not see your sidechain atoms in 
density because they are not there at all, in spite of what you may have had in 
the original protein, or even if the atoms were really there in the crystal 
_before_ exposure to the beam.

The coordinates are supposed to be what you actually find, not what you hope is 
there.

Just my two ha’porth

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 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] To Trim or Not to To Trim

[Jon Agirre]

Are downstream users of models with poorly resolved regions more likely to
spot them if they have huge B-factors or zero occupancy? If the answer is
'neither', then perhaps we need to develop a different solution.

On Fri, 10 Mar 2023 at 16:33, Goldman, Adrian 
wrote:

> Maybe simplest just to trim it back. I do worry that the presence of a
> wrong conformation will lead to inaccurate vdw clashes that could
> negatively affect other atoms.
>
> Sent from my iPhone
>
> > On 10 Mar 2023, at 18:25, Phil Jeffrey  wrote:
> >
> > On 3/10/23 4:05 AM, Julia Griese wrote:
> >> Hi all,
> >> My impression has been that the most common approach these days is to
> “let the B-factors take care of it”, but I might be wrong. Maybe it’s time
> to run another poll?
> >> Personally, I call any other approach R-factor cosmetics. The goal in
> model building is not to achieve the lowest possible R-factors, it’s to
> build the most physically meaningful, most likely to be correct, model.
> >
> > And I could call your approach "model cosmetics".
> >
> > If you can't see the side-chain, you don't know where it is and you
> probably don't even know where the centroid of the distribution is. Only in
> the case of very short side-chains with few rotamers can you make a
> reasonable volume approximation to where the side-chain is and "let the
> B-factors" smear out the density to cover a range of the projected
> conformations.
> >
> > For longer side-chains, if you put it in a single conformation, you are
> very likely NOT coming close to correctly modeling the actual distribution
> of the conformations.  So let's circle back on "most likely to be correct
> model" and ask what we *actually* know about where the atoms are.
> >
> > Put your disordered Arg in with 10 alternate conformations, each with a
> refined relative occupancy, and then let the B-factors smear that lot out,
> and that's your better model.
> >
> > Phil Jeffrey
> > Princeton
> >
> > 
> >
> > 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/
>


-- 
Dr Jon Agirre [image: A button with "Hear my name" text for name playback
in email signature]  
Royal Society University Research Fellow
CCP4 WG2 co-chair and ED champion | instruct-ERIC representative @
3D-Bioinfo (Elixir)
York Structural Biology Laboratory, Department of Chemistry
University of York, Heslington, YO10 5DD, York, UK
http://www.york.ac.uk/chemistry/research/ysbl/people/staff/jagirre/
Office: /B/K/065 Phone: +44 (0) 1904 32 8252
Mastodon: @glycojones@mastodon.world



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] To Trim or Not to To Trim

[Goldman, Adrian]

Maybe simplest just to trim it back. I do worry that the presence of a wrong 
conformation will lead to inaccurate vdw clashes that could negatively affect 
other atoms. 

Sent from my iPhone

> On 10 Mar 2023, at 18:25, Phil Jeffrey  wrote:
> 
> On 3/10/23 4:05 AM, Julia Griese wrote:
>> Hi all,
>> My impression has been that the most common approach these days is to “let 
>> the B-factors take care of it”, but I might be wrong. Maybe it’s time to run 
>> another poll?
>> Personally, I call any other approach R-factor cosmetics. The goal in model 
>> building is not to achieve the lowest possible R-factors, it’s to build the 
>> most physically meaningful, most likely to be correct, model. 
> 
> And I could call your approach "model cosmetics".
> 
> If you can't see the side-chain, you don't know where it is and you probably 
> don't even know where the centroid of the distribution is. Only in the case 
> of very short side-chains with few rotamers can you make a reasonable volume 
> approximation to where the side-chain is and "let the B-factors" smear out 
> the density to cover a range of the projected conformations.
> 
> For longer side-chains, if you put it in a single conformation, you are very 
> likely NOT coming close to correctly modeling the actual distribution of the 
> conformations.  So let's circle back on "most likely to be correct model" and 
> ask what we *actually* know about where the atoms are.
> 
> Put your disordered Arg in with 10 alternate conformations, each with a 
> refined relative occupancy, and then let the B-factors smear that lot out, 
> and that's your better model.
> 
> Phil Jeffrey
> Princeton
> 
> 
> 
> 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/

Re: [ccp4bb] To Trim or Not to To Trim

[Phil Jeffrey]

On 3/10/23 4:05 AM, Julia Griese wrote:

Hi all,

My impression has been that the most common approach these days is to 
“let the B-factors take care of it”, but I might be wrong. Maybe it’s 
time to run another poll?


Personally, I call any other approach R-factor cosmetics. The goal in 
model building is not to achieve the lowest possible R-factors, it’s to 
build the most physically meaningful, most likely to be correct, model. 


And I could call your approach "model cosmetics".

If you can't see the side-chain, you don't know where it is and you 
probably don't even know where the centroid of the distribution is. 
Only in the case of very short side-chains with few rotamers can you 
make a reasonable volume approximation to where the side-chain is and 
"let the B-factors" smear out the density to cover a range of the 
projected conformations.


For longer side-chains, if you put it in a single conformation, you are 
very likely NOT coming close to correctly modeling the actual 
distribution of the conformations.  So let's circle back on "most likely 
to be correct model" and ask what we *actually* know about where the 
atoms are.


Put your disordered Arg in with 10 alternate conformations, each with a 
refined relative occupancy, and then let the B-factors smear that lot 
out, and that's your better model.


Phil Jeffrey
Princeton



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] To Trim or Not to To Trim

[Rob Nicholls]

I believe Julia is absolutely right.

The other consideration that I haven’t seen mentioned yet in this thread is the 
effect on refinement. If you truncate a side chain or set it to zero occupancy 
then the bulk solvent mask would contain that region that is actually occupied 
by the flexible side chain in the crystal. The solvent mask is supposed to 
represent flat unordered solvent, and shouldn’t contain such ordered (though 
flexible) parts of the macromolecule. This is unavoidable in the case of larger 
unmodelled regions (e.g. missing surface loops, even whole domains at low 
resolution) and this hard problem is not accounted for using existing tools. 
But the simpler case of flexible/disordered side chains is something that we 
can do something to avoid: just build the side chain in and let the B-factors 
be accordingly inflated - those atoms must be approximately in that position in 
the crystal due to being restrained by covalent bonding, and the high B-factors 
will reflect the uncertainty/flexibility of the positioning.

And these days there are sufficient regularisers in use that the additional 
parameters from explicitly modelling the side chain shouldn’t be a problem, 
even at low resolution. Having more parameters *should* reduce Rwork if 
anything… although it will of course depend on refinement protocol/weights etc.

Zero occupancy atoms are not physically meaningful, and the fact that they are 
ignored during refinement means that the B-factors wouldn’t get refined and so 
wouldn’t reflect positional uncertainty, which could cause further issues with 
downstream misinterpretation. Similarly partial occupancies shouldn’t be used 
unless to specifically reflect the modelling assumption that those atoms are 
only physically present in a portion of the crystal.

Regards,
Rob



> On 10 Mar 2023, at 09:05, Julia Griese  wrote:
> 
> 
> CAUTION: This email originated from outside of the LMB.
> Do not click links or open attachments unless you recognize the sender and 
> know the content is safe.
> .-owner-ccp...@jiscmail.ac.uk <mailto:owner-ccp...@jiscmail.ac.uk>-.
> 
> Hi all,
>  
> My impression has been that the most common approach these days is to “let 
> the B-factors take care of it”, but I might be wrong. Maybe it’s time to run 
> another poll?
>  
> Personally, I call any other approach R-factor cosmetics. The goal in model 
> building is not to achieve the lowest possible R-factors, it’s to build the 
> most physically meaningful, most likely to be correct, model. So if you know 
> that the side chain is part of the protein, you should model it the best way 
> you can. If it’s there, just disordered, then the most correct way to model 
> it is to let it have high B-factors. Most molecular graphics programs don’t 
> flag zero-occupancy atoms, so the user might never notice. Truncation of a 
> side chain, unless there is evidence that it really physically isn’t there, 
> is also misleading, in my opinion. I don’t believe that it is more helpful to 
> the non-expert user than high B-factors either. 
>  
> If people who are not structural biologists themselves don’t know how to use 
> a structure, then we need to educate them better. It is very straightforward 
> these days to look at electron density in the PDB viewer. It used to be 
> difficult, but nowadays there’s no excuse for not checking the electron 
> density. The PDB validation flags RSRZ outliers. You can easily colour a 
> structure by B-factors. It doesn’t take that much effort to teach students 
> how to validate structures. The main point you need to get across is that it 
> is necessary to do so. And this needs to be done not only in courses aimed at 
> prospective experimental structural biologists, of course, but whenever 
> students use structures in any way. 
>  
> This is just the opinion of someone who feels very strongly about teaching 
> structure validation and rejoices when students’ reply to the question “What 
> was the most important thing you learned today?” is: “Don’t blindly trust 
> anything.” 
>  
>  
> Cheers
>  
> /Julia
>  
> --
> Dr. Julia Griese
> Associate Professor (Docent)
> Principal Investigator
> Department of Cell and Molecular Biology
> Uppsala University
> BMC, Box 596
> SE-75124 Uppsala
> Sweden
>  
> email: julia.gri...@icm.uu.se <mailto:julia.gri...@icm.uu.se>
> phone: +46-(0)18-471 4982
> http://www.icm.uu.se/structural-biology/griese-lab/ 
> <http://www.icm.uu.se/structural-biology/griese-lab/>
>  
> From: CCP4 bulletin board  <mailto:CCP4BB@JISCMAIL.AC.UK>> on behalf of Bernhard Lechtenberg 
> <968307750321-dmarc-requ...@jiscmail.ac.uk 
> <mailto:968307750321-dmarc-requ...@jiscmail.ac.uk>>
> Reply-To: Bernhard Lechtenberg  <mailto:lechtenber..

Re: [ccp4bb] To Trim or Not to To Trim

[Julia Griese]

Hi all,

My impression has been that the most common approach these days is to “let the 
B-factors take care of it”, but I might be wrong. Maybe it’s time to run 
another poll?

Personally, I call any other approach R-factor cosmetics. The goal in model 
building is not to achieve the lowest possible R-factors, it’s to build the 
most physically meaningful, most likely to be correct, model. So if you know 
that the side chain is part of the protein, you should model it the best way 
you can. If it’s there, just disordered, then the most correct way to model it 
is to let it have high B-factors. Most molecular graphics programs don’t flag 
zero-occupancy atoms, so the user might never notice. Truncation of a side 
chain, unless there is evidence that it really physically isn’t there, is also 
misleading, in my opinion. I don’t believe that it is more helpful to the 
non-expert user than high B-factors either.

If people who are not structural biologists themselves don’t know how to use a 
structure, then we need to educate them better. It is very straightforward 
these days to look at electron density in the PDB viewer. It used to be 
difficult, but nowadays there’s no excuse for not checking the electron 
density. The PDB validation flags RSRZ outliers. You can easily colour a 
structure by B-factors. It doesn’t take that much effort to teach students how 
to validate structures. The main point you need to get across is that it is 
necessary to do so. And this needs to be done not only in courses aimed at 
prospective experimental structural biologists, of course, but whenever 
students use structures in any way.

This is just the opinion of someone who feels very strongly about teaching 
structure validation and rejoices when students’ reply to the question “What 
was the most important thing you learned today?” is: “Don’t blindly trust 
anything.”


Cheers

/Julia

--
Dr. Julia Griese
Associate Professor (Docent)
Principal Investigator
Department of Cell and Molecular Biology
Uppsala University
BMC, Box 596
SE-75124 Uppsala
Sweden

email: julia.gri...@icm.uu.se
phone: +46-(0)18-471 4982
http://www.icm.uu.se/structural-biology/griese-lab/

From: CCP4 bulletin board  on behalf of Bernhard 
Lechtenberg <968307750321-dmarc-requ...@jiscmail.ac.uk>
Reply-To: Bernhard Lechtenberg 
Date: Friday, March 10, 2023 at 05:07
To: "CCP4BB@JISCMAIL.AC.UK" 
Subject: Re: [ccp4bb] To Trim or Not to To Trim

I found the poll I wrote about earlier. This actually is way older than I had 
expected (2011). You can see the poll results (which was run by Ed Pozharski) 
and discussion at the time here in the CCP4BB archive:
https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html

In brief, the results of 240 respondents were:
Delete the atoms 43%
Let refinement take care of it by inflating B-factors41%
Set occupancy to zero12%
Other 4%


Bernhard

From: CCP4 bulletin board  on behalf of Debanu Das 

Date: Friday, 10 March 2023 at 2:56 pm
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] To Trim or Not to To Trim
We dealt with this in-depth during structural genomics days when we deposited 
over 1500 novel, high-quality, experimentally-phased structures into the PDB. 
Think it’s prudent to trim/truncate side chains without reliable density.

Non-structural biologists using PDB structures without expert help can err in 
any of these scenarios: misinterpreting most common/random rotamer, zero 
occupancy atoms, B-factors, etc.

What is the value of populating the PDB, which is a structural model 
repository, with such information that is not there, i.e., reliable structural 
model?

Any trained crystallographer/structural biologist can easily add in side chain 
information if needed for modeling/computational chemistry reasons.

Best regards,
Debanu

On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch 
mailto:jxb...@case.edu>> wrote:
I’d say no trimming to side chains for the following reason: There are 
non-structural biologists using PDB files and if atoms are missing they don’t 
know what to do. A better approach is where no side chain density allows 
support of placement, pick the most common rotamer and set the occupancy to 
zero for those atoms lacking density support. More work for you but more 
accurate in my opinion.

Jürgen


___
Jürgen Bosch, PhD, MBA
Center for Global Health & Diseases
Case Western Reserve University
Cleveland, OH 44106
https://www.linkedin.com/in/jubosch/

CEO & Co-Founder at InterRayBio, LLC



On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg 
<968307750321-dmarc-requ...@jiscmail.ac.uk<mailto:968307750321-dmarc-requ...@jiscmail.ac.uk>>
 wrote:

Hi Rhys,

I am also all for leaving side chains and letting the B-factors deal with the 
weak/absent density.

I don’t think th

Re: [ccp4bb] To Trim or Not to To Trim

[Bernhard Lechtenberg]

I found the poll I wrote about earlier. This actually is way older than I had 
expected (2011). You can see the poll results (which was run by Ed Pozharski) 
and discussion at the time here in the CCP4BB archive:
https://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg20268.html

In brief, the results of 240 respondents were:
Delete the atoms 43%
Let refinement take care of it by inflating B-factors41%
Set occupancy to zero12%
Other 4%


Bernhard

From: CCP4 bulletin board  on behalf of Debanu Das 

Date: Friday, 10 March 2023 at 2:56 pm
To: CCP4BB@JISCMAIL.AC.UK 
Subject: Re: [ccp4bb] To Trim or Not to To Trim
We dealt with this in-depth during structural genomics days when we deposited 
over 1500 novel, high-quality, experimentally-phased structures into the PDB. 
Think it’s prudent to trim/truncate side chains without reliable density.

Non-structural biologists using PDB structures without expert help can err in 
any of these scenarios: misinterpreting most common/random rotamer, zero 
occupancy atoms, B-factors, etc.

What is the value of populating the PDB, which is a structural model 
repository, with such information that is not there, i.e., reliable structural 
model?

Any trained crystallographer/structural biologist can easily add in side chain 
information if needed for modeling/computational chemistry reasons.

Best regards,
Debanu

On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch 
mailto:jxb...@case.edu>> wrote:
I’d say no trimming to side chains for the following reason: There are 
non-structural biologists using PDB files and if atoms are missing they don’t 
know what to do. A better approach is where no side chain density allows 
support of placement, pick the most common rotamer and set the occupancy to 
zero for those atoms lacking density support. More work for you but more 
accurate in my opinion.

Jürgen


___
Jürgen Bosch, PhD, MBA
Center for Global Health & Diseases
Case Western Reserve University
Cleveland, OH 44106
https://www.linkedin.com/in/jubosch/


CEO & Co-Founder at InterRayBio, LLC




On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg 
<968307750321-dmarc-requ...@jiscmail.ac.uk<mailto:968307750321-dmarc-requ...@jiscmail.ac.uk>>
 wrote:

Hi Rhys,

I am also all for leaving side chains and letting the B-factors deal with the 
weak/absent density.

I don’t think there is a consensus, but I kind of remember that somebody did a 
poll a few years ago and if I remember correctly the main approaches were the 
one described above, or trimming the side-chains.

Bernhard

Bernhard C. Lechtenberg PhD
NHMRC Emerging Leadership Fellow
Laboratory Head
Ubiquitin Signalling Division​​
E lechtenber...@wehi.edu.au<mailto:lechtenber...@wehi.edu.au>
T +61 3 9345 2217


From: CCP4 bulletin board mailto:CCP4BB@JISCMAIL.AC.UK>> 
on behalf of Rhys Grinter 
<22087c81e8c6-dmarc-requ...@jiscmail.ac.uk<mailto:22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>>
Date: Friday, 10 March 2023 at 12:26 pm
To: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> 
mailto:CCP4BB@JISCMAIL.AC.UK>>
Subject: [ccp4bb] To Trim or Not to To Trim
Hi All,

I'm trying to crowdsource an opinion on how people deal with modelling side 
chains with poorly resolved electron or cryoEM density.

My preference is to model the sidechain and allow the B-factors to go high in 
refinement to represent that the side chain is flexible. However, I'm aware 
that some people truncate sidechains if density is not present to justify 
modelling. I've also seen models where the sidechain is modelled but with zero 
occupancy if density isn't present.

Is there a consensus and justifying arguments for why one approach is better?

Cheers,

Rhys




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
--
---
LinkedIn: www.linkedin.com/in/debanudas<http://www.linkedin.com/in/debanudas>
Cal Alumni: cal.berkeley.edu/debanudas<http://cal.berkeley.edu/debanudas>



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



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

This message was issued to members of www.jiscmail.ac.u

Re: [ccp4bb] To Trim or Not to To Trim

[Debanu Das]

In fact, for non-SBs using the PDB, it is far likelier that they will
recognize truncated side chains and thus seek help or know what to do
themselves, compared to recognizing incorrectly modeled rotamers (as they
won’t be checking electron density likely) and atoms with zero occupancy
and B-factors, which could lead to fairly incorrect biological
interpretations.
Best,
Debanu

On Thu, Mar 9, 2023 at 7:55 PM Debanu Das  wrote:

> We dealt with this in-depth during structural genomics days when we
> deposited over 1500 novel, high-quality, experimentally-phased structures
> into the PDB. Think it’s prudent to trim/truncate side chains without
> reliable density.
>
> Non-structural biologists using PDB structures without expert help can err
> in any of these scenarios: misinterpreting most common/random rotamer, zero
> occupancy atoms, B-factors, etc.
>
> What is the value of populating the PDB, which is a structural model
> repository, with such information that is not there, i.e., reliable
> structural model?
>
> Any trained crystallographer/structural biologist can easily add in side
> chain information if needed for modeling/computational chemistry reasons.
>
> Best regards,
> Debanu
>
> On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch  wrote:
>
>> I’d say no trimming to side chains for the following reason: There are
>> non-structural biologists using PDB files and if atoms are missing they
>> don’t know what to do. A better approach is where no side chain density
>> allows support of placement, pick the most common rotamer and set the
>> occupancy to zero for those atoms lacking density support. More work for
>> you but more accurate in my opinion.
>>
>> Jürgen
>>
>>
>> ___
>> Jürgen Bosch, PhD, MBA
>> Center for Global Health & Diseases
>> Case Western Reserve University
>> Cleveland, OH 44106
>> https://www.linkedin.com/in/jubosch/
>>
>> CEO & Co-Founder at InterRayBio, LLC
>>
>>
>>
>> On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg <
>> 968307750321-dmarc-requ...@jiscmail.ac.uk> wrote:
>>
>> Hi Rhys,
>>
>> I am also all for leaving side chains and letting the B-factors deal with
>> the weak/absent density.
>>
>> I don’t think there is a consensus, but I kind of remember that somebody
>> did a poll a few years ago and if I remember correctly the main approaches
>> were the one described above, or trimming the side-chains.
>>
>> Bernhard
>>
>> *Bernhard C. Lechtenberg* PhD
>> NHMRC Emerging Leadership Fellow
>> Laboratory Head
>> Ubiquitin Signalling Division​​
>> E lechtenber...@wehi.edu.au
>> T +61 3 9345 2217
>>
>>
>>
>> *From: *CCP4 bulletin board  on behalf of Rhys
>> Grinter <22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>
>> *Date: *Friday, 10 March 2023 at 12:26 pm
>> *To: *CCP4BB@JISCMAIL.AC.UK 
>> *Subject: *[ccp4bb] To Trim or Not to To Trim
>> Hi All,
>>
>> I'm trying to crowdsource an opinion on how people deal with modelling
>> side chains with poorly resolved electron or cryoEM density.
>>
>> My preference is to model the sidechain and allow the B-factors to go
>> high in refinement to represent that the side chain is flexible. However,
>> I'm aware that some people truncate sidechains if density is not present to
>> justify modelling. I've also seen models where the sidechain is modelled
>> but with zero occupancy if density isn't present.
>>
>> Is there a consensus and justifying arguments for why one approach is
>> better?
>>
>> Cheers,
>>
>> Rhys
>>
>>
>> --
>>
>> 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
>>
> --
> ---
> LinkedIn: www.linkedin.com/in/debanudas
> Cal Alumni: cal.berkeley.edu/debanudas
>
-- 
---
LinkedIn: www.linkedin.com/in/debanudas
Cal Alumni: cal.berkeley.edu/debanudas



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] To Trim or Not to To Trim

[Debanu Das]

We dealt with this in-depth during structural genomics days when we
deposited over 1500 novel, high-quality, experimentally-phased structures
into the PDB. Think it’s prudent to trim/truncate side chains without
reliable density.

Non-structural biologists using PDB structures without expert help can err
in any of these scenarios: misinterpreting most common/random rotamer, zero
occupancy atoms, B-factors, etc.

What is the value of populating the PDB, which is a structural model
repository, with such information that is not there, i.e., reliable
structural model?

Any trained crystallographer/structural biologist can easily add in side
chain information if needed for modeling/computational chemistry reasons.

Best regards,
Debanu

On Thu, Mar 9, 2023 at 6:50 PM Jurgen Bosch  wrote:

> I’d say no trimming to side chains for the following reason: There are
> non-structural biologists using PDB files and if atoms are missing they
> don’t know what to do. A better approach is where no side chain density
> allows support of placement, pick the most common rotamer and set the
> occupancy to zero for those atoms lacking density support. More work for
> you but more accurate in my opinion.
>
> Jürgen
>
>
> ___
> Jürgen Bosch, PhD, MBA
> Center for Global Health & Diseases
> Case Western Reserve University
> Cleveland, OH 44106
> https://www.linkedin.com/in/jubosch/
>
> CEO & Co-Founder at InterRayBio, LLC
>
>
>
> On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg <
> 968307750321-dmarc-requ...@jiscmail.ac.uk> wrote:
>
> Hi Rhys,
>
> I am also all for leaving side chains and letting the B-factors deal with
> the weak/absent density.
>
> I don’t think there is a consensus, but I kind of remember that somebody
> did a poll a few years ago and if I remember correctly the main approaches
> were the one described above, or trimming the side-chains.
>
> Bernhard
>
> *Bernhard C. Lechtenberg* PhD
> NHMRC Emerging Leadership Fellow
> Laboratory Head
> Ubiquitin Signalling Division​​
> E lechtenber...@wehi.edu.au
> T +61 3 9345 2217
>
>
>
> *From: *CCP4 bulletin board  on behalf of Rhys
> Grinter <22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>
> *Date: *Friday, 10 March 2023 at 12:26 pm
> *To: *CCP4BB@JISCMAIL.AC.UK 
> *Subject: *[ccp4bb] To Trim or Not to To Trim
> Hi All,
>
> I'm trying to crowdsource an opinion on how people deal with modelling
> side chains with poorly resolved electron or cryoEM density.
>
> My preference is to model the sidechain and allow the B-factors to go high
> in refinement to represent that the side chain is flexible. However, I'm
> aware that some people truncate sidechains if density is not present to
> justify modelling. I've also seen models where the sidechain is modelled
> but with zero occupancy if density isn't present.
>
> Is there a consensus and justifying arguments for why one approach is
> better?
>
> Cheers,
>
> Rhys
>
>
> --
>
> 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
>
-- 
---
LinkedIn: www.linkedin.com/in/debanudas
Cal Alumni: cal.berkeley.edu/debanudas



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] To Trim or Not to To Trim

[Jurgen Bosch]

I’d say no trimming to side chains for the following reason: There are 
non-structural biologists using PDB files and if atoms are missing they don’t 
know what to do. A better approach is where no side chain density allows 
support of placement, pick the most common rotamer and set the occupancy to 
zero for those atoms lacking density support. More work for you but more 
accurate in my opinion.

Jürgen


___
Jürgen Bosch, PhD, MBA
Center for Global Health & Diseases
Case Western Reserve University
Cleveland, OH 44106
https://www.linkedin.com/in/jubosch/

CEO & Co-Founder at InterRayBio, LLC



> On Mar 9, 2023, at 9:45 PM, Bernhard Lechtenberg 
> <968307750321-dmarc-requ...@jiscmail.ac.uk> wrote:
> 
> Hi Rhys,
>  
> I am also all for leaving side chains and letting the B-factors deal with the 
> weak/absent density.
>  
> I don’t think there is a consensus, but I kind of remember that somebody did 
> a poll a few years ago and if I remember correctly the main approaches were 
> the one described above, or trimming the side-chains.
>  
> Bernhard
>  
> Bernhard C. Lechtenberg PhD
> NHMRC Emerging Leadership Fellow
> Laboratory Head
> Ubiquitin Signalling Division​​
> E lechtenber...@wehi.edu.au <mailto:lechtenber...@wehi.edu.au>
> T +61 3 9345 2217
>  
>  
> From: CCP4 bulletin board  on behalf of Rhys Grinter 
> <22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>
> Date: Friday, 10 March 2023 at 12:26 pm
> To: CCP4BB@JISCMAIL.AC.UK 
> Subject: [ccp4bb] To Trim or Not to To Trim
> 
> Hi All,
>  
> I'm trying to crowdsource an opinion on how people deal with modelling side 
> chains with poorly resolved electron or cryoEM density.
>  
> My preference is to model the sidechain and allow the B-factors to go high in 
> refinement to represent that the side chain is flexible. However, I'm aware 
> that some people truncate sidechains if density is not present to justify 
> modelling. I've also seen models where the sidechain is modelled but with 
> zero occupancy if density isn't present. 
>  
> Is there a consensus and justifying arguments for why one approach is better?
>  
> Cheers,
>  
> Rhys
>  
>  
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 
> 
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1
> 




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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Bernhard Lechtenberg]

Hi Rhys,

I am also all for leaving side chains and letting the B-factors deal with the 
weak/absent density.

I don’t think there is a consensus, but I kind of remember that somebody did a 
poll a few years ago and if I remember correctly the main approaches were the 
one described above, or trimming the side-chains.

Bernhard

Bernhard C. Lechtenberg PhD
NHMRC Emerging Leadership Fellow
Laboratory Head
Ubiquitin Signalling Division​​
E lechtenber...@wehi.edu.au<mailto:lechtenber...@wehi.edu.au>
T +61 3 9345 2217


From: CCP4 bulletin board  on behalf of Rhys Grinter 
<22087c81e8c6-dmarc-requ...@jiscmail.ac.uk>
Date: Friday, 10 March 2023 at 12:26 pm
To: CCP4BB@JISCMAIL.AC.UK 
Subject: [ccp4bb] To Trim or Not to To Trim
Hi All,

I'm trying to crowdsource an opinion on how people deal with modelling side 
chains with poorly resolved electron or cryoEM density.

My preference is to model the sidechain and allow the B-factors to go high in 
refinement to represent that the side chain is flexible. However, I'm aware 
that some people truncate sidechains if density is not present to justify 
modelling. I've also seen models where the sidechain is modelled but with zero 
occupancy if density isn't present.

Is there a consensus and justifying arguments for why one approach is better?

Cheers,

Rhys




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



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

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

Re: [ccp4bb] To Trim or Not to To Trim

[Jon Cooper]

Hello, yes, the trimming and occupancy tricks can be confusing and they can be 
a bit of a cheeky way to improve the model statistics. I prefer to build the 
whole side chain and if you can't see all or part of it, then that's just too 
bad, but I have erred.

Best wishes, Jon Cooper. jon.b.coo...@protonmail.com

Sent from Proton Mail mobile

 Original Message 
On 10 Mar 2023, 01:25, Rhys Grinter wrote:

> Hi All,
>
> I'm trying to crowdsource an opinion on how people deal with modelling side 
> chains with poorly resolved electron or cryoEM density.
>
> My preference is to model the sidechain and allow the B-factors to go high in 
> refinement to represent that the side chain is flexible. However, I'm aware 
> that some people truncate sidechains if density is not present to justify 
> modelling. I've also seen models where the sidechain is modelled but with 
> zero occupancy if density isn't present.
>
> Is there a consensus and justifying arguments for why one approach is better?
>
> Cheers,
>
> Rhys
>
> ---
>
> To unsubscribe from the CCP4BB list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB=1



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

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

[ccp4bb] To Trim or Not to To Trim

[Rhys Grinter]

Hi All,

I'm trying to crowdsource an opinion on how people deal with modelling side
chains with poorly resolved electron or cryoEM density.

My preference is to model the sidechain and allow the B-factors to go high
in refinement to represent that the side chain is flexible. However, I'm
aware that some people truncate sidechains if density is not present to
justify modelling. I've also seen models where the sidechain is modelled
but with zero occupancy if density isn't present.

Is there a consensus and justifying arguments for why one approach is
better?

Cheers,

Rhys



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/