[chimerax-users] alphafold contacts output file

Tue Oct 4 11:04:56 PDT 2022

Yes, in agreement with Tom, how to interpret the modeling in the context of your specific research project depends on your own scientific judgment and knowledge of the proteins of interest, and what you are able to glean from the literature.

I'm not an authority on protein-protein interfaces either, but thought I should mention that there have been a number of papers that survey or discuss the properties of known interfaces.  Here are a couple of older examples, but you could probably find dozens more in Pubmed:

Characterization of protein-protein interfaces.
Yan C, Wu F, Jernigan RL, Dobbs D, Honavar V.
Protein J. 2008 Jan;27(1):59-70. doi: 10.1007/s10930-007-9108-x.
PMID: 17851740

Analysis of protein-protein interaction sites using surface patches.
Jones S, Thornton JM.
J Mol Biol. 1997 Sep 12;272(1):121-32. doi: 10.1006/jmbi.1997.1234.
PMID: 9299342

Keep in mind that it may be quite different for multidomain proteins vs. single domain, membrane proteins vs. soluble, etc.

I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.                       
UCSF Chimera(X) team
Department of Pharmaceutical Chemistry
University of California, San Francisco

> On Oct 4, 2022, at 10:44 AM, Tom Goddard via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
> 
> Hi Phil,
> 
>   I think for crystal packing contacts the minimum contact area for judging a biological interface was about 400 Angstroms.  But this is all a matter of research judgement and I am not expert on assessing protein-protein interfaces.
> 
>   You can specify exactly the atoms you want to measure the buried area between in the command, for example,
> 
> 	measure buriedarea /A:40-50 with /B:30-40
> 
> will measure between residue numbers 40-50 in chain A and residues numbers 30-40 in chain B.  Another approach is to just delete parts of the proteins before doing the measurement
> 
> 	delete /A:1-39 /B:1-29
> 
> Our ChimeraX mailing list is primarily for discussing how to the software works, not for advise on how analyze data in your research -- since we are software developers and not experts in your areas of research.
> 
>     Tom
> 
> 
>> On Oct 4, 2022, at 10:33 AM, McClean, Phillip <phillip.mcclean at ndsu.edu> wrote:
>> 
>> Thanks again Tom.
>> 
>> The buried surface area was 1535 angstroms squared.  Is that a "good" value?  Also is that value calculated across the two proteins?  There appears to be specific regions of the two proteins that are in contact.  Is it possible to measure the buried area just for those two regions.
>> 
>> A note on these two proteins.  Both have specific structured regions (that are interacting) but also have disordered regions elsewhere that account for >50% of the proteins distances.  I have looked at those regions using IUPred3.
>> 
>> Phil
>> From: Tom Goddard <goddard at sonic.net>
>> Sent: Tuesday, October 4, 2022 12:23 PM
>> To: McClean, Phillip <phillip.mcclean at ndsu.edu>
>> Cc: chimerax-users at cgl.ucsf.edu <chimerax-users at cgl.ucsf.edu>
>> Subject: Re: [chimerax-users] alphafold contacts output file
>>  
>> Hi Philip,
>> 
>>   The alphafold contacts code as it is currently written finds the pairs of close residues but does not remember which atoms were the closest so it can't easily output the closest atoms between a pair of residues.  But the ChimeraX contacts command or tool is actually intended to find contacting residues and uses more sophisticated criteria that take into account the van der Waals radii of the atoms.
>> 
>> https://www.cgl.ucsf.edu/chimerax/docs/user/tools/clashes.html
>>         https://www.cgl.ucsf.edu/chimerax/docs/user/commands/clashes.html
>> 
>>   Confidence in whether two proteins really make a biological contact in for example a crystal structure where the proteins may must be touching due to crystal packing is usually judge by the total area of contact residues, not by how close individual residues are to each other.  The ChimeraX buried surface area calculation calculates how much surface is in contact (square Angstroms).
>> 
>> measure buriedarea /A with /B
>> > Buried area between /A and /B = 484.71
>> 
>>   To judge if an AlphaFold prediction of a protein-protein interface is correct I suggest looking at both the buried surface area and also the the AlphaFold PAE values between residue pairs at the interface.
>> 
>> Tom
>> 
>> 
>>> On Oct 4, 2022, at 10:08 AM, McClean, Phillip <phillip.mcclean at ndsu.edu> wrote:
>>> 
>>> Hi Tom,
>>> 
>>> Thanks.  Found it on the desktop.
>>> 
>>> A follow-up Chimerax question (hopefully it shouldn't be a different thread).  Is it possible to list the atoms that are in close contact in the output file.
>>> 
>>> And a general newbie question.  You choose a distance of 3 as the default cutoff.  In the computational protein-protein modeling world, is this considered good evidence of contact between two proteins.  The reason I ask is that I have discovered wild type and mutant alleles of a MYB protein which were modeled with Alphafold  multimer along with its functional beta helix-loop-helix protein partner.  If I see changes in distances might that suggest changes in interaction capacity of any sort?
>>> 
>>> Phil McClean
>>> From: Tom Goddard <goddard at sonic.net>
>>> Sent: Tuesday, October 4, 2022 11:57 AM
>>> To: McClean, Phillip <phillip.mcclean at ndsu.edu>
>>> Cc: chimerax-users at cgl.ucsf.edu <chimerax-users at cgl.ucsf.edu>
>>> Subject: Re: [chimerax-users] alphafold contacts output file
>>>  
>>> Hi Phil,
>>> 
>>>   If you start ChimeraX by clicking an icon then it puts files on your desktop.  If you start it from a shell (for example on Linux) it will put the files in the directory you started in.  To find the directory ChimeraX is currently in use the ChimeraX pwd command
>>> 
>>> pwd
>>> Current working directory is: /Users/goddard/Desktop
>>> 
>>> To change the directory ChimeraX is currently in use the cd command
>>> 
>>> cd ~/Documents
>>> Current working directory is: /Users/goddard/Documents
>>> 
>>> Lastly you can give a full path to the output file in your command
>>> 
>>> alphafold contacts /B to /C output /Users/goddard/contact_pae.txt
>>> 
>>> or
>>> 
>>> alphafold contacts /B to /C output ~/contact_pae.txt
>>> 
>>> where "~" means your user account home directory.
>>> 
>>>    Tom
>>> 
>>>> On Oct 4, 2022, at 9:21 AM, McClean, Phillip via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
>>>> 
>>>> Hi Everyone-
>>>> 
>>>> I used the following command to obtain a file with the contacts between two chains:
>>>> 
>>>> alphafold contacts /B to /C output contact_pae.txt
>>>> I cannot seem to find where the output file is located.
>>>> 
>>>> Thanks for any help.
>>>> 
>>>> Phil McClean
>>>> _______________________________________________
>>>> ChimeraX-users mailing list
>>>> ChimeraX-users at cgl.ucsf.edu
>>>> Manage subscription:
>>>> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
> 
> _______________________________________________
> ChimeraX-users mailing list
> ChimeraX-users at cgl.ucsf.edu
> Manage subscription:
> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users