[Chimera-users] [chimerax-users] Automate calculation and saving of residues' structural attributes
Tom Goddard
goddard at sonic.net
Thu Feb 3 17:40:10 PST 2022
Hi Christian,
The Python code that runs the MSMS surface calculation (in a separate process since it crashes frequently and we don't want Chimera to crash in that case) is here in our old SVN code repository
https://plato.cgl.ucsf.edu/trac/chimera/browser/trunk/libs/MoleculeSurface
or in your Chimera it would be here on Mac, but in different places on Windows or Linux
Chimera-1.16.app/Contents/Resources/share/MoleculeSurface
You are asking for a world of trouble using MSMS for these calculations because it crashes on many large structures (10,000 atoms) and the ones it doesn't crash on it will often fail in subtler ways giving the wrong surface with self intersections or holes. And that is the bases for the area calculations.
ChimeraX does not use MSMS and computes the SAS area analytically without making a triangulated surface and no failure has ever been reported. Also ChimeraX computes the SES surface and works reliably and correctly on structures of any size. But you would have to write a bit of code to get the SES area from that computed surface -- this is rarely used so we have not written code to get that value. And of course the percent exposure per residue is another calculation not in ChimeraX and requires some methodology to define what that even means (using reference areas).
Tom
> On Feb 3, 2022, at 10:46 AM, Elaine Meng via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
>
> Hi Christian,
> I guess this is a Chimera question now, so I set the "reply to" to chimera-users in case this generates further messages. The following pertains to Chimera.
>
> I believe it uses the MSMS code without modification to do the actual calculation, so it might be needless work to run all this benchmarking. Nevertheless I can understand why you want to figure out the cause of any differences.
>
> (a) I would expect differences in VDW radii to be a major factor. The Chimera default radii are described/listed in the User Guide, see this and links therein:
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/vdwrad.html>
>
> You could try assigning the MSMS-provided radii within Chimera before performing the area calculation in Chimera. See the bottom section of the help page URL above for possible ways of assigning different VDW radius values.
>
> (b) secondarily, it depends how you use the command... if you just used Chimera command "surface" it would automatically just surface the macromolecules and try to omit waters, ligands, etc. However, you can use "surfcat" command to define whatever sets of atoms you like for subsequent "surface" calculations.
> Chimera surface command
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/surface.html>
> Chimera surfcat command
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/msms.html>
>
> (c) dealing with the code is beyond my skill set, so if you still need pointers to look therein, somebody else would have to advise.
>
> I hope this helps,
> Elaine
> -----
> Elaine C. Meng, Ph.D.
> UCSF Chimera(X) team
> Department of Pharmaceutical Chemistry
> University of California, San Francisco
>
>> On Feb 3, 2022, at 12:33 AM, Christian Tüting via ChimeraX-users <chimerax-users at cgl.ucsf.edu> wrote:
>>
>> Hi All,
>>
>> Andre and I continued this conversation in private, as this was not
>> Chimera related anymore.
>>
>> Somehow, we have a question regarding the calculation of the of the SES
>> and SAS.
>>
>> In principle, I looked into the calcsurf.py file in the chimera folder,
>> and figured out, that this calculation is done by msms with a probe
>> radius of 1.4 and a density of 2 and noh. So I copied the code into my
>> workflow and tested this with a xyzr file, generated by the pdb_to_xyzr
>> shell script from msms, and got the same results, as if I am running
>> directly msms from command line. So far not suprisingly.
>>
>> But when we take the same pdb file, the msms code itself gives slightly
>> different values than chimera. Often, it's not siginificant, but Andre
>> benchmarked this, and sometimes the difference is more than 1.2 fold.
>>
>>
>>
>> So this let us come to the conclusion, that chimera somehow calculates a
>> different xyzr file, compared to what the shell script is doing. I was
>> not able to find the code for caluclating the xyzr file. Another
>> explanation is, that chimera treats the results a bit differently. msms
>> results ses and sas per atom, we "just" summed them per residue to get
>> the per residue value. But I don't think, that this is done differently
>> here.
>>
>>
>> So our questions are:
>> (a) is the radius definition different in chimera from what is given
>> with the msms software ( https://ccsb.scripps.edu/msms/downloads/; there
>> is a shellscript and, more importantly a text file, which defines the
>> radius of each atom, with a lot of special cases ).
>> (b) is the structure somehow "filtered" prior to surface calculation.
>> E.g. are waters, ions, ligands considered part of the structure, of are
>> they excluded, as this might introduce also some differences
>> (c) or can you help us find the actual code for this in chimera, than I
>> will be able to anwser the question myself
>>
>> Thanks in advance.
>>
>> Christian
>>
>>
>>
>>>>> André <andre.jfmdm at gmail.com> 01/21/22 7:06 PM >>>
>> Hi Elaine,
>>
>> Thank you so much for your very helpful answer!
>>
>> Yeah, Tom told me before that it wouldn't be possible to directly
>> calculate
>> SESA and relSESA with ChimeraX. So, I'll probably have to stick with
>> Chimera. I've just discovered the "list" command (functional in
>> Chimera),
>> and I think that with it I'll be able to easily get all the attributes
>> by
>> using the --nogui startup.
>> I'll work on it in the following days, then I'll come back with updates.
>>
>> Once again, thank you very much!
>>
>> Best,
>> André.
>>
>> Em sex., 21 de jan. de 2022 às 13:55, Elaine Meng <meng at cgl.ucsf.edu>
>> escreveu:
>>
>>> Hi André,
>>> Although it displays an SES, ChimeraX does not calculate SES areas,
>> only
>>> SAS areas (command "measure sasa" which also assigns a residue
>> attribute).
>>> <https://rbvi.ucsf.edu/chimerax/docs/user/commands/measure.html#sasa>
>>>
>>> For Chimera, another group had provided SES reference areas for the
>>> exposed state represented by G-X-G tripeptides, but I don't know of a
>>> similar SAS reference area set for ChimeraX that could be used to
>> calculate
>>> relative exposure, sorry. In case others are interested, the Chimera
>>> relative exposure calculation giving relSESA is described here:
>>> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/surfnorm.html>
>>>
>>> So depending on the importance of SESA and relSESA, you may need to
>> use
>>> Chimera, or at least some method outside of ChimeraX.
>>>
>>> In both ChimeraX and Chimera, bfactor is read from the input PDB file,
>> and
>>> phi and psi are calculated automatically for peptide/protein
>> structures
>>> when they are read in. ChimeraX atom and residue attributes:
>>> <https://rbvi.ucsf.edu/chimerax/docs/user/attributes.html#atom>
>>> <https://rbvi.ucsf.edu/chimerax/docs/user/attributes.html#residue>
>>>
>>> kdHydrophobicity is not a calculation, but a simple lookup table by
>> amino
>>> acid type. You may not need either program to "calculate" it, but it
>> is
>>> automatically assigned in Chimera. It can be assigned in ChimeraX by
>>> running this comand script:
>>> <
>>>
>> https://rbvi.ucsf.edu/chimerax/docs/user/kyte-doolittle> values along with some alternative hydrophobicity scales (lookup
>> tables)
>>> are summarized here:
>>>
>> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/hydrophob.html>
>>>
>>> I hope this helps,
>>> Elaine
>>> -----
>>> Elaine C. Meng, Ph.D.
>>> UCSF Chimera(X) team
>>> Department of Pharmaceutical Chemistry
>>> University of California, San Francisco
>>>
>>>> On Jan 21, 2022, at 7:28 AM, André via ChimeraX-users <
>>> chimerax-users at cgl.ucsf.edu> wrote:
>>>>
>>>> Hi Christian,
>>>>
>>>> Thank you so much for your solution! That indeed helps a lot, since
>> it
>>> shows me how to execute ChimeraX scripts independently and directly
>> from
>>> Python!
>>>>
>>>> Now, do you know which commands I should use to calculate the
>> following
>>> structural attributes for all residues: areaSAS, areaSES, relSESA,
>> bfactor,
>>> kdhydro, phi, psi?
>>>>
>>>> I do it manually in Chimera (haven't figured out how to do the same
>> in
>>> ChimeraX yet), by following the route: Tools -> Structure Analysis ->
>>> Render by Attribute -> File -> Save Attributes -> Attribute to Save,
>> which
>>> allows each attribute of residues to be saved as txt files, which
>> contain
>>> the residue number and respective attribute, for all residues in the
>>> protein. For example, the file for areaSAS looks like this:
>>>>
>>>> attribute: areaSAS
>>>> recipient: residues
>>>> :1.B 116.71626508235931
>>>> :2.B 118.39353680610657
>>>> ...
>>>>
>>>> Thus, that's what I wanted to do: figure out which commands I should
>> use
>>> to get the same result as the manual route above (although it is not
>>> really necessary that the txt file is generated directly -- if the
>> needed
>>> information appear in the stdout_as_str string, just like in the
>> example
>>> you showed, I'd be able to parse it to a file, and the problem would
>> be
>>> solved!).
>>>>
>>>> Once again, thank you for your help!
>>>>
>>>> Best,
>>>> André.
>>>>
>>>>
>>>>
>>>> Em sex., 21 de jan. de 2022 às 09:56, Christian Tüting <
>>> christian.tueting at biochemtech.uni-halle.de> escreveu:
>>>> Hi André,
>>>>
>>>> i am pretty sure, that my solution is very crude and that there are
>>>> better ways of doing it (like importing chimerax in the python code
>>>> directly). I guess, you'll get more anwsers soon.
>>>>
>>>> I quickly wrote a script, which performs a chimerax task
>> independently,
>>>> by just running python code.
>>>>
>>>> First: a chimerax script file with the desired commands:
>>>>
>>>> script.cxc:
>>>> open 7ott
>>>> surface #1
>>>> mearuse area #1
>>>> exit
>>>>
>>>> Note: the exit command is needed, otherwise the subprocress will
>> idle in
>>>> the chimerax cmd line.
>>>>
>>>>
>>>> Second: a python script, which execute this script in chimerax and
>>>> fetches the output:
>>>>
>>>> test.py:
>>>>
>>>> import subprocess
>>>> import os
>>>>
>>>> cwd = os.getcwd()
>>>>
>>>> chimerax =
>> "/Applications/ChimeraX_Daily.app/Contents/MacOS/ChimeraX"
>>>> scriptfile = f"{cwd}/script.cxc"
>>>>
>>>> cmd = [chimerax, "--nogui",scriptfile]
>>>> p = subprocess.run(" ".join(cmd),shell = True,capture_output=True)
>>>>
>>>> stdout_as_str = p.stdout.decode("utf-8")
>>>> print(stdout_as_str)
>>>>
>>>> This just runs a subprocess and fetches the output to a string,
>> which
>>>> you can than further analyse.
>>>>
>>>>
>>>> Best
>>>> Christian
>>>>
>>>>
>>>>
>>>> Dr. rer. nat. Christian Tüting
>>>>
>>>> Kastritis Laboratory for Biomolecular Research
>>>> Cryo-Electron Microscopy & Computational Structural Biology
>>>> ________________________________________________
>>>> Martin-Luther-Universität Halle-Wittenberg
>>>> Biozentrum, Room A.2.19
>>>> IWE ZIK HALOmem NWG III
>>>> "Kryo-Elektronenmikroskopie an Membranproteinkomplexen"
>>>> Weinbergweg 22, 06120 Halle
>>>> tel: +49 345 5524985
>>>> web (Lab): https://blogs.urz.uni-halle.de/kastritislab/
>>>> web (HALOmem): https://www.halomem.de/en/
>>>>>>> André via ChimeraX-users <chimerax-users at cgl.ucsf.edu> 01/21/22
>> 1:29
>>>> PM >>>
>>>> Hello!
>>>>
>>>> I'm working on an automated pipeline (in Python) for the structural
>>>> analysis of proteins. In particular, I'm using Chimera to> > areaSES, relSESA, bfactor, kdhydro, phi, psi). These files are the
>>>> inputs
>>>> of the pipeline.
>>>> However, currently the files are exported manually, and, for the
>> purpose
>>>> of
>>>> building a fully automated pipeline, it would be great to also
>> automate
>>>> this step of calculating the attributes and exporting their
>> respective
>>>> files.
>>>>
>>>> I searched a lot, and tried some ways to achieve such an automation
>>>> (using
>>>> Chimera), but nothing worked.
>>>> And I also considered using ChimeraX. In particular, I went through
>>>> ChimeraX Programming Manual (
>>>> https://www.cgl.ucsf.edu/chimerax/docs/devel/index.html#), but, if I
>>>> understood correctly, I'd only be able to import chimerax and use
>> its
>>>> functions in the Python interpreter within ChimeraX (Tools >>
>> General >>
>>>> Shell).
>>>> At first, I'd like to import chimerax as a module in a Python
>>>> interpreter
>>>> outside of ChimeraX, so that manually opening ChimeraX wouldn't be
>>>> necessary in order to save the files with the residues' attributes.
>> But,
>>>> from some other threads in the mailing list, I found that it
>> wouldn't be
>>>> possible (right?)
>>>>
>>>> But I also found out that I could write a .py script which
>> calculates
>>>> and
>>>> exports the files with the residues' properties, and then run it
>> with
>>>> ChimeraX application from the command-line, something like:
>> *chimerax
>>>> --nogui myscript.py*
>>>> If that works, I think that my automation problem would be solved!
>> The
>>>> only
>>>> problem is that I really couldn't figure out the Python code to
>>>> calculate
>>>> and export the aforementioned residues' properties (I took a look in
>>>> ChimeraX Recipes (https://rbvi.github.io/chimerax-recipes/), but I
>>>> couldn't
>>>> find what I need, nor more detailed documentation).
>>>>
>>>> Given that, I'd like to know if anyone has done something like this,
>>>> and/or
>>>> have any guidance on how I could proceed.
>>>>
>>>> I really appreciate any help you can provide!
>>>> (And thanks to Tom Goddard for suggesting moving the discussion here
>> to
>>>> the
>>>> ChimeraX mailing list).
>>>>
>>>> Thank you so much,
>>>> André.
>>>>
>>>> _______________________________________________
>>>> ChimeraX-users mailing list
>>>> ChimeraX-users at cgl.ucsf.edu
>>>> Manage subscription:
>>>> https://plato.cgl.ucsf.edu/mailman/listinfo/chimerax-users
>>>
>>>
>>
>> _______________________________________________
>> ChimeraX-users mailing list
>> ChimeraX-users at cgl.ucsf.edu
>> Manage subscription:
>> https://plato.cgl.ucsf.edu/mailman/listinfo/chimerax-users
>>
>
>
> _______________________________________________
> ChimeraX-users mailing list
> ChimeraX-users at cgl.ucsf.edu
> Manage subscription:
> https://plato.cgl.ucsf.edu/mailman/listinfo/chimerax-users
>
More information about the Chimera-users
mailing list