To follow up on one of Gilles least important questions: I was convinced that the PDB must contain structures for which there are more than two alternate conformations. I downloaded all structures with resolutions between 0.5 and 1.0A and with a bit of regexp grepping, quickly found the following set of examples with at least three alternate conformers (I didn't find any with more than four, and the few that had four looked "suspicious"):
1BXO, 1BYI, 1CBN, 1CEX, 1DY5, 1EB6, 1EJG, 1EM0, 1EN8, 1EN9, 1ENE, 1ENN, 1ET1, 1EXR, 1F9Y, 1FN8, 1FY4, 1FY5, 1GCI, 1GDN, 1GWE, 1IX9, 1IXB, 1JFB, 1JXT, 1JXU, 1JXW, 1JXX, 1JXY, 1KTH, 1M40, 1MSO, 1MUW, 1MWQ, 1MXT, 1N1P, 1N4U, 1N4V, 1N4W, 1N55, 1N9B, 1OD3, 1PJX, 1PQ5, 1PQ8, 1R6J, 1S5M, 1S5N, 1SSX, 1TG0, 1US0, 1UZV, 1V0L, 1V6P, 1VB0, 1W0N, 1X6X, 1X6Z, 1XG0, 1XVO, 1YK4, 2AT0, 2BV4, 2BW4, 2CNQ, 2CWS, 2DDX, 2E4T, 2FWH, 2GEW, 2GG2, 2H5C, 2H5D, 2I16, 2I17, 2IXT, 2IZQ, 2NLS, 2NMZ, 2NRL, 2O7A, 2P74, 2PEV, 2PF8, 2PFH, 2PPP, 2PWA, 2PYA, 2QCP, 2QXW, 2RH2, 2V8B, 2VB1, 2Z6W, 3AL1, 3B3R, 3C78, 3CNJ, 3D1P, 8A3H Cheers! -Tom -----Original Message----- From: Gilles Truan [mailto:gilles.tr...@cgm.cnrs-gif.fr] Sent: Mon 9/22/2008 6:14 AM To: Nathan Baker Cc: David Gohara; apbs-us...@lists.sourceforge.net list; pymol-users@lists.sourceforge.net Subject: Re: [PyMOL] [Apbs-users] Strange behaviour with APBS Dear all, Sorry for the late answer but I finally managed to find what the error is... First thing, I work on a Windows XP environment (nobody's perfect!!), and some of the problems were related to that. When I tried running everything from the command line (ie pdb2pqr and apbs) I did get the files 1CJC.pqr and the corresponding map pot.dx. I was able to read them with pymol and everything works. The problem arises only with the APBS pymol plugin. In the CJC file, residues with the alternate configuration A (and frankly I do not see the reason for that as the alt B is not present....) existed and the handling of these alt conformations is different if pymol or pdb2pqr do the charge calculation. When the pdb file is transformed by pdb2pqr (command line version 1.3.0), the alternate sidechains conformations disappear. When Pymol transforms the file into a pqr one the alt field is still present and apbs does not understand the atom that looks something like "NH1AARG". If I use the file generated by pdb2pqr (with the --apbs-input option) it works, apbs reads and does the calculation. The trick is the to remove alternate configurations (specially when no alt B exists!!!). I have done it on the CJC file and everything works perfectly from the apbs plugin and without the need to generate the pdb2pqr file. I was not able to pinpoint the problem with the pymol output when running APBS. I noticed that the error messages differ a bit if I run Pymol from a link (ie without a shell window) or from the pymol command in a DOS shell. In the Pymol command window the error message is : * APBS Tools: coarse grid: (301.633,318.546,307.146) APBS Tools: fine grid: (197.431,207.380,200.674) APBS Tools: center: (11.993,-11.506,14.548) APBS Tools: fine grid points (97,97,97) ObjectMapLoadDXFile-Error: Unable to open file! ObjectMapLoadDXFile: Does 'pymol-generated.dx' exist? * while in the cmd window the message is *Reading PQR-format atom data from pymol-generated.pqr. Valist_readPQR: Error parsing atom 3526! Error reading molecules! ObjectMapLoadDXFile-Error: Unable to open file! ObjectMapLoadDXFile: Does 'pymol-generated.dx' exist?* which of course is much more informative because it points directly to the error in the pqr file that I had not been able to find because the pqr looked ok to me (Unfortunately I did not jump to the 3526th atom that has the alt A conformation!!). Now the last problem: running the APBS plugin in WinXP generate path errors because of the spaces in the paths. *Parsed input file. Got paths for 1 molecules Reading PQR-format atom data from **D:\Documents. VASSERT: ASSERTION FAILURE! filename vmem.c, line 241, ((num > 0) && (size > 0)) This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information. ObjectMapLoadDXFile-Error: Unable to open file! ObjectMapLoadDXFile: Does 'pymol-generated.dx' exist?* So the path is not complete (D:\Documents and Settings\Gilles\Mes Documents\blahblahblah\ is missing). Again, the error message in the pymol console was not complete: * APBS Tools: coarse grid: (103.210,121.009,109.487) APBS Tools: fine grid: (80.712,91.182,84.404) APBS Tools: center: (17.315,-2.473,13.671) APBS Tools: fine grid points (97,97,97) ObjectMapLoadDXFile-Error: Unable to open file! * Unfortunately I thought I had the same error while trying to run APBS with the 1CJC file and using the pymol generated pqr (error with alt conformations) or using the pqr file generated independantly (PATH error with spaces in path names). It is all sorted out now. I can use the pdb2pqr web site to generate the pqr and apbs input files or directly pymol APBS plugin with the b factor and alternate conformation "blanked out" before use!!! I finally looked at the apbs.py file to see where the pqr calculation was and added a few commands after line 1491: here they are: pymol.cmd.remove('%s and alt b' %sel) pymol.cmd.alter ('%s' %sel,'b=0') pymol.cmd.alter ('%s' %sel,'alt=""') pymol.cmd.sort('%s' %sel) I tried the APBS script on the 1CJC file and it works. I am not 100% sure it will work for every file as I only deleted the "b" conformation (do PDB files with more than 2 alt conformations exist?). And BTW it is a bit crude to decide to switch to the "a" conformation instead of letting the user choose his preferred one!! Anyway, I think the problem is solved for me, except for this PATH problem in APBS trying to read an external pqr file. But so far, I think I will definitely manage with the different workarounds. Gilles Nathan Baker a écrit : > Hi Gilles -- > >> I installed APBS 0.4.0 and my pymol version is the latest build for >> windows, ie 1.1 r0, the apbs plugin is the latest also I believe. > > That's a very old version of APBS (2005); the latest is 1.0.0 and can > be downloaded from http://apbs.sourceforge.net/#download > >> The b factor error is only visible when calculating the grid. I >> already use the b factor change (alter 1cjc, b=0), and the workaround >> works perfectly when the error message complains about strange >> numbers in the file (ValueError: invalid literal for float(): >> 1.00101.34). I also tried to do a pdb2pqr (via the web interface) >> before doing apbs, but I have never been able to read the pqr file >> through the APBS interface. > > What sorts of errors do you receive? Almost all of my PyMOL > calculations are performed with pre-generated PQR files from PDB2PQR > without problem. > >> In my hands, the pymol pdb to pqr conversion works quite well. My >> second problem is apparently not related to the b factor or the pqr >> file: it seems that for some pdb files, APBS cannot (or does not) see >> (or write) the pymol-generated.dx file... >> If I fetch 1EHB pdb file, I can calculate an APBS map (removing water >> and hem molecules). If I fetch 1CJC pdb file, I cannot (again >> removing water and fad molecules)!! If I fetch again 1EHB and redo >> the calculation, everything goes fine. Here is the output of the >> commands... > > This is generally due to an APBS error; although the error messages > often get obscured through the "APBS Tools" interface (Warren, is > there some way to make these errors more transparent?). If you visit > your working directory for the APBS calculation, you should find the > input and PQR files used with the calculation. Can you send those to us? > > Thanks, > > Nathan > >> >> PyMOL>fetch 1ehb >> PyMOL>remove resn hoh+hem >> APBS Tools: coarse grid: (61.146,53.431,62.444) >> APBS Tools: fine grid: (55.968,51.430,56.732) >> APBS Tools: center: (16.504,0.863,7.287) >> APBS Tools: fine grid points (97,97,97) >> DXStrToMap: Dimensions: 97 97 97 >> DXStrToMap: Origin -11.480 -24.852 -21.079 >> DXStrToMap: Grid 0.583 0.536 0.591 >> DXStrToMap: 912673 data points. >> PyMOL>delete all >> PyMOL>fetch 1cjc >> PyMOL>remove resn hoh+fad >> Maximum number of grid points exceeded. Old grid dimensions were >> [129, 161, 161] >> Fine grid points rounded down from [91, 115, 115] >> New grid dimensions are [65, 97, 97] >> APBS Tools: coarse grid: (100.841,119.466,106.848) >> APBS Tools: fine grid: (79.318,90.274,82.852) >> APBS Tools: center: (17.545,-2.019,13.919) >> APBS Tools: fine grid points (65,97,97) >> ObjectMapLoadDXFile-Error: Unable to open file! >> ObjectMapLoadDXFile: Does 'pymol-generated.dx' exist? >> PyMOL>delete all >> PyMOL>fetch 1ehb >> PyMOL>remove resn hoh+hem >> APBS Tools: coarse grid: (61.146,53.431,62.444) >> APBS Tools: fine grid: (55.968,51.430,56.732) >> APBS Tools: center: (16.504,0.863,7.287) >> APBS Tools: fine grid points (97,97,97) >> DXStrToMap: Dimensions: 97 97 97 >> DXStrToMap: Origin -11.480 -24.852 -21.079 >> DXStrToMap: Grid 0.583 0.536 0.591 >> DXStrToMap: 912673 data points. >> >> DeLano Scientific a écrit : >>> Nathan, >>> >>> Strictly speaking, the issue may be with PDB2PQR itself, not APBS. >>> Furthermore, we don't yet know whether a combined development >>> version of all >>> three packages would still reproduce the problem. >>> >>> As an aside, a good reason for PyMOL to see PQR files is for diagnostic >>> visualization: to enable the user to inspect and modify or >>> manipulate what >>> is being fed directly into APBS. >>> PyMOL can read & write PQR directly, though we may have issues reading >>> non-PDB-like PQR files. I would generally agree that use of PDB2PQR to >>> convert PDBs to PQRs is the preferred route, since PyMOL doesn't do any >>> checking or optimization. >>> >>> Anyway, based on reported symptoms alone, the problem appears to be >>> that >>> some component in the PyMOL->APBS pipeline assumes that the input >>> PDBs are >>> space-delimited rather than column-delimited. This guess is based >>> on the >>> observation that a B-factor of 100 or greater eliminates the space >>> between >>> the occupany and the B factor in spec-compliant PDB files, and the >>> calculation fails. >>> >>> As a starting point for troubleshooting, I just pulled current pdb2pqr >>> source from trunk and CANNOT reproduce the problem, so either this is >>> already a solved problem with current code, or something else is >>> going on. >>> Gilles, what versions of PyMOL and APBS are you using? >>> >>> Cheers, >>> Warren >>> -- >>> DeLano Scientific LLC >>> Subscriber Support Services >>> mailto:supp...@delsci.com >>> >>> >>> -----Original Message----- >>> From: Nathan Baker [mailto:nathanaba...@mac.com] Sent: Wednesday, >>> September 17, 2008 6:28 PM >>> To: DeLano Scientific >>> Cc: gilles.tr...@cgm.cnrs-gif.fr; pymol-users@lists.sourceforge.net; >>> apbs-us...@lists.sourceforge.net >>> Subject: Re: [Apbs-users] [PyMOL] Strange behaviour with APBS >>> >>> Hi Warren -- >>> >>> The PQR format (which doesn't really have a standard) is basically just >>> whitespace delimited. I'm curious, though -- why would a PyMOL APBS >>> invocation ever see the PDB file? Wouldn't a user first convert it >>> to PQR >>> with PDB2PQR (ideally) or with PyMOL's built-in functionality? >>> >>> Thanks, >>> >>> Nathan >>> >>> On Sep 17, 2008, at 2:21 PM, DeLano Scientific wrote: >>> >>> >>>> Gilles, >>>> >>>> The PyMOL and APBS programs have differing views as to whether PDB >>>> & PQR files are column-based or space-delimited, respectively. We >>>> must eventually sort this out, because it causes continued grief >>>> for our users in common. >>>> >>>> For now, however, the workaround is to make sure no B-factors >>>> exceed 100. >>>> >>>> alter all, b=min(b,99.9) >>>> >>>> should about do it. >>>> >>>> Cheers, >>>> Warren >>>> >>>> -----Original Message----- >>>> From: pymol-users-boun...@lists.sourceforge.net >>>> [mailto:pymol-users-boun...@lists.sourceforge.net] On Behalf Of >>>> Gilles Truan >>>> Sent: Wednesday, September 17, 2008 1:09 AM >>>> To: pymol-users@lists.sourceforge.net >>>> Subject: [PyMOL] Strange behaviour with APBS >>>> >>>> I have noticed a strange behaviour of the APBS module. I used it a >>>> lot recently and it works perfectly for 95% of the structures I >>>> analyzed. I usually fetch the pdb code, then remove waters, >>>> heteroatoms and so on. I also noticed that in some cases, when the >>>> b factor is quite big (over >>>> 100) APBS cannot work the pdb file because it mixes up the b factor >>>> and the number which is just before (occupancy). But I managed with >>>> that, I usually erase the b factor to 0 before doing the APBS >>>> calculation and everything works fine then. >>>> The other strange behavious is related to nothing I can pinpoint >>>> unfortunately. For some files (I always fetch them with the fetch >>>> command from Pymol), I can set the grid OK and when I launch APBS, >>>> I can see the file being processed (I see the select and hydrogens >>>> go on!) and then it stops with the following error message: >>>> >>>> ObjectMapLoadDXFile-Error: Unable to open file! >>>> ObjectMapLoadDXFile: Does 'pymol-generated.dx' exist? >>>> >>>> All the paths for APBS are correctly setup, the temp files are >>>> setup as well. >>>> If I then fetch a new file, APBS works, even for several files (one >>>> after >>>> another) and if I come back to the file I wanted to do first, it >>>> gives me the same error message (BTW, the pdb file is 1CJC). >>>> >>>> Any help would be greatly appreciated!!! >>>> >>>> >>>> ---------------------------------------------------------------------- >>>> --- This SF.Net email is sponsored by the Moblin Your Move >>>> Developer's challenge Build the coolest Linux based applications >>>> with Moblin SDK & win great prizes Grand prize is a trip for two to >>>> an Open Source event anywhere in the world >>>> http://moblin-contest.org/redirect.php?banner_id=100&url=/ >>>> _______________________________________________ >>>> apbs-users mailing list >>>> apbs-us...@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/apbs-users >>>> >>> >>> -- >>> Associate Professor, Dept. of Biochemistry and Molecular Biophysics >>> Center >>> for Computational Biology, Washington University in St. Louis >>> Web: http://cholla.wustl.edu/ >>> >>> >>> >>> >>> >>> >> <gilles_truan.vcf> > > -- > Associate Professor, Dept. of Biochemistry and Molecular Biophysics > Center for Computational Biology, Washington University in St. Louis > Web: http://cholla.wustl.edu/ > > > > > > This email (including any attachments) may contain material that is confidential and privileged and is for the sole use of the intended recipient. Any review, reliance or distribution by others or forwarding without express permission is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies. Exelixis, Inc. reserves the right, to the extent and under circumstances permitted by applicable law, to retain, monitor and intercept e-mail messages to and from its systems.
