enhancement: Creating a PDB output file with all atoms that are part of patches labeled in b-factor column

[sgrannem]

I am very keen to use this package but I could not figure out which resides actually belong to the patches. Ideally, each residue that is part of an electrostatic patch would be labeled somehow in the original PDB file. Perhaps the b-factor column could be used for this? Then I can start integrating the tool in my pipeline.

[fwaibl]

Hello,
Thanks for you interest in using our tool!
We have a pull request open (#32), that might be related to your request. Specifically, it would implement a --resout option that would write the residues involved in each patch to a CSV file. Would this solve your problem?
@vhoer should we merge this PR?

Best,
Franz

[sgrannem]

Hi Franz, Yes, that would work. I can then add this information to the b-factor column. We would like to use the patch information to predict putative nucleic acid binding sites and this tool would really help! Sander On 16 Jan 2025, at 13:59, Franz Waibl ***@***.***> wrote: This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Hello, Thanks for you interest in using our tool! We have a pull request open (#32<#32>), that might be related to your request. Specifically, it would implement a --resout option that would write the residues involved in each patch to a CSV file. Would this solve your problem? @vhoer<https://github.com/vhoer> should we merge this PR? Best, Franz — Reply to this email directly, view it on GitHub<#35 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AISKJQNCEOQVYGC64Y3Q2X32K63LVAVCNFSM6AAAAABVFM7BYOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJVG43DONJTGM>. You are receiving this because you authored the thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[fwaibl]

Dear Sander,
I just had a look at #32, and saw that it is not really compatible with our current main branch anymore because we changed some units in the last PR that we merged. To be consistent, I just did another change (already merged that), which updates the units in pep_patch_electrostatics to be nanometer-based. I am aware that this is an unpleasant change for existing users, but it increases our internal consistency, and it means that the PR #32 is again working. Do you want to try out the Patch_Residues_Logging branch, or should I simply merge it?
Best,
Franz

[sgrannem]

Hi Franz, I am happy for you to merge it. I plan to play with the new version this afternoon. Thank you! Sander From: Franz Waibl ***@***.***> Date: Friday, 17 January 2025 at 14:23 To: liedllab/surface_analyses ***@***.***> Cc: Sander Granneman ***@***.***>, Author ***@***.***> Subject: Re: [liedllab/surface_analyses] enhancement: Creating a PDB output file with all atoms that are part of patches labeled in b-factor column (Issue #35) This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Dear Sander, I just had a look at #32<#32>, and saw that it is not really compatible with our current main branch anymore because we changed some units in the last PR that we merged. To be consistent, I just did another change (already merged that), which updates the units in pep_patch_electrostatics to be nanometer-based. I am aware that this is an unpleasant change for existing users, but it increases our internal consistency, and it means that the PR #32<#32> is again working. Do you want to try out the Patch_Residues_Logging branch, or should I simply merge it? Best, Franz — Reply to this email directly, view it on GitHub<#35 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AISKJQIPHOL4USDZKQDEE2D2LEG5LAVCNFSM6AAAAABVFM7BYOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJYGQ3TQNZQGE>. You are receiving this because you authored the thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[sgrannem]

Hi Franz, My sincere apologies for bothering you with this, but I just installed #32 and for some reason I cannot see the new output options: /home/pyrbdome/anaconda3/envs/peppatch/bin/pep_patch_electrostatic -h usage: pep_patch_electrostatic [-h] [--ref REF] [--protein_ref PROTEIN_REF] [--stride STRIDE] [--dx [DX]] [--apbs_dir APBS_DIR] [--probe_radius PROBE_RADIUS] [-o OUT] [-c PATCH_CUTOFF PATCH_CUTOFF] [-ic INTEGRAL_CUTOFF INTEGRAL_CUTOFF] [--surface_type {sas,ses,gauss}] [--ply_out PLY_OUT] [--pos_patch_cmap POS_PATCH_CMAP] [--neg_patch_cmap NEG_PATCH_CMAP] [--ply_cmap PLY_CMAP] [--ply_clim PLY_CLIM PLY_CLIM] [--check_cdrs] [-n N_PATCHES] [-s SIZE_CUTOFF] [--gauss_shift GAUSS_SHIFT] [--gauss_scale GAUSS_SCALE] [--pH PH] [--ion_species [ION_SPECIES ...]] parm trajs [trajs ...] positional arguments: parm trajs optional arguments: -h, --help show this help message and exit --ref REF Reference structure with the SAME atoms (default: None) --protein_ref PROTEIN_REF Reference structure for protein alignment using TMalign (default: None) --stride STRIDE --dx [DX] Optional dx file with the electrostatic potential. If this is omitted, you must specify --apbs_dir (default: None) --apbs_dir APBS_DIR Directory in which intermediate files are stored when running APBS. Will be created if it does not exist. (default: None) --probe_radius PROBE_RADIUS Probe radius in nm (default: 0.14) -o OUT, --out OUT Output csv file. (default: None) -c PATCH_CUTOFF PATCH_CUTOFF, --patch_cutoff PATCH_CUTOFF PATCH_CUTOFF Cutoff for positive and negative patches. (default: (2.0, -2.0)) -ic INTEGRAL_CUTOFF INTEGRAL_CUTOFF, --integral_cutoff INTEGRAL_CUTOFF INTEGRAL_CUTOFF Cutoffs for "high" and "low" integrals. (default: (0.3, -0.3)) --surface_type {sas,ses,gauss} Which type of molecular surface to produce. (default: sas) --ply_out PLY_OUT Base name for .ply output for PyMOL. Will write BASE-pos.ply and BASE-neg.ply. (default: None) --pos_patch_cmap POS_PATCH_CMAP Matplotlib colormap for .ply positive patches output. (default: tab20c) --neg_patch_cmap NEG_PATCH_CMAP Matplotlib colormap for .ply negative patches output. (default: tab20c) --ply_cmap PLY_CMAP Matplotlib colormap for .ply potential output. (default: coolwarm_r) --ply_clim PLY_CLIM PLY_CLIM Colorscale limits for .ply output. (default: None) --check_cdrs For an antibody Fv region as input: check whether patches belong to CDRs. (default: False) -n N_PATCHES, --n_patches N_PATCHES Restrict output to n patches. Positive values output n largest patches, negative n smallest patches. (default: 0) -s SIZE_CUTOFF, --size_cutoff SIZE_CUTOFF Restrict output to patches with an area of over s A^2. If s = 0, no cutoff is applied (default). (default: 0.0) --gauss_shift GAUSS_SHIFT --gauss_scale GAUSS_SCALE --pH PH Specify pH for pdb2pqr calculation. If None, no protonation is performed. (default: None) --ion_species [ION_SPECIES ...] Specify ion species and their properties (charge, concentration, and radius). Provide values for multiple ion species as charge1, conc1, radius1, charge2, conc2, radius2, etc. (default: None) Am I missing something? Sander From: Franz Waibl ***@***.***> Date: Friday, 17 January 2025 at 14:23 To: liedllab/surface_analyses ***@***.***> Cc: Sander Granneman ***@***.***>, Author ***@***.***> Subject: Re: [liedllab/surface_analyses] enhancement: Creating a PDB output file with all atoms that are part of patches labeled in b-factor column (Issue #35) This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Dear Sander, I just had a look at #32<#32>, and saw that it is not really compatible with our current main branch anymore because we changed some units in the last PR that we merged. To be consistent, I just did another change (already merged that), which updates the units in pep_patch_electrostatics to be nanometer-based. I am aware that this is an unpleasant change for existing users, but it increases our internal consistency, and it means that the PR #32<#32> is again working. Do you want to try out the Patch_Residues_Logging branch, or should I simply merge it? Best, Franz — Reply to this email directly, view it on GitHub<#35 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AISKJQIPHOL4USDZKQDEE2D2LEG5LAVCNFSM6AAAAABVFM7BYOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJYGQ3TQNZQGE>. You are receiving this because you authored the thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[fwaibl]

Dear Sander,
when I install the branch of #32, I get the following output (note the -ro or --resout option). Are you sure that you installed the package correctly in your environment?
I am still waiting for a comment from @vhoer, after that we'll merge this PR.

pep_patch_electrostatic starting at 2025-01-17 17:41:30.984986
Command line arguments:
/localhome/fwaibl/.conda/envs/py310_2/bin/pep_patch_electrostatic --help
usage: pep_patch_electrostatic [-h] [--ref REF] [--protein_ref PROTEIN_REF] [--stride STRIDE] [--dx [DX]] [--apbs_dir APBS_DIR] [--probe_radius PROBE_RADIUS] [-o OUT] [-ro RESOUT] [-c PATCH_CUTOFF PATCH_CUTOFF]
                               [-ic INTEGRAL_CUTOFF INTEGRAL_CUTOFF] [--surface_type {sas,ses,gauss}] [--ply_out PLY_OUT] [--pos_patch_cmap POS_PATCH_CMAP] [--neg_patch_cmap NEG_PATCH_CMAP] [--ply_cmap PLY_CMAP]
                               [--ply_clim PLY_CLIM PLY_CLIM] [--check_cdrs] [-n N_PATCHES] [-s SIZE_CUTOFF] [--gauss_shift GAUSS_SHIFT] [--gauss_scale GAUSS_SCALE] [--pH PH] [--ion_species [ION_SPECIES ...]]
                               parm trajs [trajs ...]

positional arguments:
  parm
  trajs

options:
  -h, --help            show this help message and exit
  --ref REF             Reference structure with the SAME atoms (default: None)
  --protein_ref PROTEIN_REF
                        Reference structure for protein alignment using TMalign (default: None)
  --stride STRIDE
  --dx [DX]             Optional dx file with the electrostatic potential. If this is omitted, you must specify --apbs_dir (default: None)
  --apbs_dir APBS_DIR   Directory in which intermediate files are stored when running APBS. Will be created if it does not exist. (default: None)
  --probe_radius PROBE_RADIUS
                        Probe radius in nm (default: 0.14)
  -o OUT, --out OUT     Output csv file. (default: None)
  -ro RESOUT, --resout RESOUT
                        Residue csv file. (default: None)
  -c PATCH_CUTOFF PATCH_CUTOFF, --patch_cutoff PATCH_CUTOFF PATCH_CUTOFF
                        Cutoff for positive and negative patches. (default: (2.0, -2.0))
  -ic INTEGRAL_CUTOFF INTEGRAL_CUTOFF, --integral_cutoff INTEGRAL_CUTOFF INTEGRAL_CUTOFF
                        Cutoffs for "high" and "low" integrals. (default: (0.3, -0.3))
  --surface_type {sas,ses,gauss}
                        Which type of molecular surface to produce. (default: sas)
  --ply_out PLY_OUT     Base name for .ply output for PyMOL. Will write BASE-pos.ply and BASE-neg.ply. (default: None)
  --pos_patch_cmap POS_PATCH_CMAP
                        Matplotlib colormap for .ply positive patches output. (default: tab20c)
  --neg_patch_cmap NEG_PATCH_CMAP
                        Matplotlib colormap for .ply negative patches output. (default: tab20c)
  --ply_cmap PLY_CMAP   Matplotlib colormap for .ply potential output. (default: coolwarm_r)
  --ply_clim PLY_CLIM PLY_CLIM
                        Colorscale limits for .ply output. (default: None)
  --check_cdrs          For an antibody Fv region as input: check whether patches belong to CDRs. (default: False)
  -n N_PATCHES, --n_patches N_PATCHES
                        Restrict output to n patches. Positive values output n largest patches, negative n smallest patches. (default: 0)
  -s SIZE_CUTOFF, --size_cutoff SIZE_CUTOFF
                        Restrict output to patches with an area of over s A^2. If s = 0, no cutoff is applied (default). (default: 0.0)
  --gauss_shift GAUSS_SHIFT
  --gauss_scale GAUSS_SCALE
  --pH PH               Specify pH for pdb2pqr calculation. If None, no protonation is performed. (default: None)
  --ion_species [ION_SPECIES ...]
                        Specify ion species and their properties (charge, concentration, and radius). Provide values for multiple ion species as charge1, conc1, radius1, charge2, conc2, radius2, etc. (default: None)

[vhoer]

Hey all,
I'd like to include the fix to #34 as well, which I'll add tomorrow and fix some small issues with the tests as well.
Then I think we can merge the PR.

[sgrannem]

Hi All, Please find attached my surface analysis test code. I managed to install the code corresponding to the branch that reports the residues that are part of the patch. The attached HTML file shows you exactly what I tried. The pep_patch electrostatic code ran just fine, but I think I may have found a bug where the residue numbers are not the same as the numbers that are reported in the PDB file. See the attached P0132_merged.pdb file, where the first residue is an ALA with residue number 5, whereas the biggest patch reported in the results file (see attached “all_results.csv”) starts with one. Ideally, I’d like the residue numbers for each residue to be exactly the same as reported in the PDB file as it would otherwise be challenging to map the patches back to the original PDB file. Would this be possible? You may be wondering, what am I trying to accomplish? I would like to use your tool as an alternative to BindUP. This tool also reports positively charged patches, and has worked well for us, but we have to use web servers to get the results, making automation challenging. The server is also unreliable, so I would like to integrate your code in our pipeline to replace BindUP. I’ve attached the BindUP results file for the protein that I was analysing (P09132; SRP19). Here you can see that the patch is indicated in the b-factor column (value = 10). You can then visualise the patch easily in pymol/chimera by colouring by b-factor. That is the type of result that I wish to achieve. I have ~600 more proteins to analyse! ;D Sander On 17 Jan 2025, at 18:17, vhoer ***@***.***> wrote: This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Hey all, I'd like to include the fix to #34<#34> as well, which I'll add tomorrow and fix some small issues with the tests as well. Then I think we can merge the PR. — Reply to this email directly, view it on GitHub<#35 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AISKJQMEELQS2L35U34MFPL2LFCNPAVCNFSM6AAAAABVFM7BYOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJYHEZTMNRXGE>. You are receiving this because you authored the thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[fwaibl]

Hi,

I just merged #32, please let me know if this helps. I think the residue number is consistent with the mdtraj object rather than the input PDB. Would it be possible to work with these (continuous) residue numbers? E.g., it would be easy to write another PDB file that also has this numbering. But generally, it is not very easy to retain all residue information after processing with MDTraj, especially with insertion codes etc.

[sgrannem]

Hi Franz, Many thanks. Yes it is important to have the numbering exactly as the PDB file because in the PDB file there are often amino acids missing at the beginning and there are also gaps in the structures because these regions were not resolved or because of alternative splicing. So you could have a sequence like this 5-MKL-7-gap-10-CQT-12 But with the current code it would look like: 1-MKLCQT-6 This could complicate downstream analyses. Thanks! A great tool! Sander From: Franz Waibl ***@***.***> Date: Wednesday, 29 January 2025 at 17:30 To: liedllab/surface_analyses ***@***.***> Cc: Sander Granneman ***@***.***>, Author ***@***.***> Subject: Re: [liedllab/surface_analyses] enhancement: Creating a PDB output file with all atoms that are part of patches labeled in b-factor column (Issue #35) This email was sent to you by someone outside the University. You should only click on links or attachments if you are certain that the email is genuine and the content is safe. Hi, I just merged #32<#32>, please let me know if this helps. I think the residue number is consistent with the mdtraj object rather than the input PDB. Would it be possible to work with these (continuous) residue numbers? — Reply to this email directly, view it on GitHub<#35 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AISKJQLD4QZECCRNAQYEZUT2NEF4LAVCNFSM6AAAAABVFM7BYOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMMRSGM4TOMZTGQ>. You are receiving this because you authored the thread.Message ID: ***@***.***> The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336.

[fwaibl]

Hi,
in the current code, we re-number the residues starting from 1. I believe that this was added in the past to avoid ending up with multiple residues with the exact same name, because the insertion codes are lost in the MDTraj object.
As far as I see, there are multiple options. @sgrannem @vhoer what is your preference?

• We can retain the current behavior and force the user to re-number the PDB. This is always possible because the current numbering is unique, but might be tedious.
• We can remove the re-numbering, which would mostly give the expected result, but might be more tedious in cases where the insertion codes are relevant.
• We can think of a more elaborate solution. E.g., we could re-assign insertion codes when multiple residues have the same number.