From: Eran H. <era...@gm...> - 2009-03-30 22:47:59
|
Bob, Thank you for the clarification. Everything makes sense. Eran On Tue, Mar 31, 2009 at 1:39 AM, Robert Hanson <ha...@st...> wrote: > Let's clear up the atom serial number business. > > PDB file contain atom serial numbers. These are in the ATOM and HETATM > records. Behind the scenes, in the Jmol code, we call these "atom serial > numbers" as opposed to "atom index numbers", which start at 0 and go > consecutively through ALL models loaded. You can use atom serial numbers > "atomno" in select commands, to select desired atoms, but we can't use them > in Jmol in terms of the state for several reasons: > > 1) They repeat with different structures in an multi-model file or when > multiple files are loaded. > 2) They could be anything, including negative numbers. > 3) They could be in any order and could have gaps. > > Really it doesn't matter what the RCSB writes as the specification. This > atom serial number field has been used in various ways; Jmol just can't use > it as a true "unique identifier" for anything. > > As for the state: Once a selection has been made, that original script is > history. What is selected is a set of atoms. Jmol must use its atom indexing > method, which is 100% reliable based on reading the same file again and is > 100% independent of the history. > > I really think it comes down to this simple tenet: A script file describes > the state for the file that was loaded. That's all that can be asked of it. > You can't ask that the script be able to "autoupdate" if the file it refers > to is modified. > > If you ask me, it's a design flaw of Protopedia to be accessing live PDB > files from RCSB with Jmol state scripts. Of course that should be fine if > the files are never changed, but since they are subject to modification, > that's just not acceptable. Eric, you need to find the original files, copy > them, and access them locally. If there's a way to access all earlier files > before a certain date, then perhaps you could just make a blanket change in > all the Jmol script files (a simple bulk search and replace for the > directory and filename) to change all the pdb file directories to that older > directory and, perhaps, add ".gzip" or whatever to the filename. If not, > well, you just have to go find those files and copy them yourself. > > I'm sorry, but I really think that's the only appropriate solution. Not > fun, I'm sure. > > Bob > > > > On Mon, Mar 30, 2009 at 5:06 PM, Eran Hodis <era...@gm...> wrote: > >> Dear Eric, >> Your email mentioned atom serial numbers but Bob said in a previous email >> "Eric, Jmol does not use atom serial numbers in state scripts. It uses >> atom indexes, but maybe that's what you were referring to." >> >> However, Bob, if atom indexes are the same as atom serial numbers, then I >> too would argue that Jmol state scripts should use a means other than atom >> indexes to save colorings, representations, and so forth, not because it >> would make life more simple for Proteopedia, but rather because John >> Westbrook at the PDB suggested in the email Eric sent us that "Neither the >> V2.6 nor V3.2 format specifications suggests that the atom serial number be >> used as the primary atom identifier." In other words they feel completely >> comfortable changing the atom serial number, and so if any Jmol user has a >> saved state script but hasn't locally saved the PDB file it recalls, that >> state script would not display correctly. >> >> As Eric said, such a change in the state saving function would not help >> Proteopedia with the current problem, but would help with future releases of >> remediated PDB files. >> >> Best regards, >> Eran >> >> >> On Tue, Mar 31, 2009 at 12:16 AM, Eric Martz <em...@mi...>wrote: >> >>> Dear Bob, >>> >>> How difficult would it be to change the state script generator in Jmol to >>> avoid using atom serial numbers? >>> >>> I'm sure none of us ever envisioned the current state of affairs. We have >>> who knows how many state scripts saved in Proteopedia, and now an unknown >>> number of March-17-remediated PDB files have scrambled atom serial numbers. >>> Further, there is no guarantee that such scrambling will not occur again in >>> a future remediation. >>> >>> A change in Jmol to avoid serial number dependency in saved state scripts >>> will not avoid the need for repairs to Proteopedia now, but may avoid such >>> calamities in the future. >>> >>> Regards, -Eric >>> >>> Date: Mon, 30 Mar 2009 15:09:56 -0400 >>> From: John Westbrook <jw...@rc...> >>> Reply-To: jw...@rc... >>> Organization: RCSB - Protein Data Bank >>> To: em...@mi... >>> CC: Helen Berman <be...@rc...>, Kim Henrick < >>> he...@eb...> >>> Subject: Re: Chain order changes: a problem for Proteopedia >>> >>> Dear Eric, >>> >>> In producing the V3.2 wwPDB release, we have tried to preserve as much >>> as possible the PDB chain and residue nomenclature for polymer molecular >>> components. >>> With version 3.2 comes the introduction of more uniform assignment >>> PDB chain identifiers for ligands and solvent, so this has resulted in >>> some nomenclature changes in V3.2 files. >>> >>> Neither the V2.6 nor V3.2 format specifications suggests that the atom >>> serial >>> number be used as the primary atom identifier. I am sure that you are >>> aware that >>> there are now many PDB entries that have been split across multiple PDB >>> data >>> files specifically because of the limitation in the range atom serial >>> numbers. >>> Atom serial numbers are also replicated between models so they do not >>> represent >>> unique atom identifiers for NMR or other multi-model entries. >>> >>> If you have a specific dependency prior atom serial numbers in your >>> software >>> system, then you can always recover the particular version of the PDB >>> entry >>> that you used from our ftp snapshot server ( ftp://snapshots.rcsb.org). >>> >>> I should point out that we provided Jaim with an advanced copy of the >>> V3.2 files for testing on Dec 3, 2008. The issue of atom ordering >>> was not raised as an issue at that time. >>> >>> Regards, >>> >>> John >>> >>> >>> >>> >>> >>> >>> Begin forwarded message: >>> >>> >>> >>>> From: Eric Martz <em...@mi...> >>> >>>> Date: March 29, 2009 7:27:12 PM PDT >>> >>>> To: "in...@rc..." <in...@rc...>, "pd...@rc..." < >>> pd...@rc...> >>> >>>> Subject: pdb-l: Chain order changes: a problem for Proteopedia >>> >>> >>>> >>> >>>> Dear wwPDB: >>> >>>> >>> >>>> The March 17, 2009 remediation of PDB data in the wwPDB (PDB format >>> >>>> 3.20) appears to me to have, in many cases, changed the order of >>> >>>> chains, and hence the atom serial numbers in the PDB files. This has >>> >>>> created a major problem in the wiki Proteopedia.Org, where many >>> >>>> molecular scenes that took hours or weeks to develop are now >>> >>>> nonfunctional. >>> >>>> >>> >>>> The problem arises becaused Jmol uses atom serial numbers for >>> >>>> selecting groups of atoms when it saves a molecular scene (in a >>> >>>> "state script"). Proteopedia's Scene Authoring Tool uses Jmol's >>> state >>> >>>> scripts to capture molecular scenes and attach them to "green >>> links". >>> >>>> >>> >>>> Questions: >>> >>>> >>> >>>> 1. Were the names of ATOM chains ever changed? I assume (and hope) >>> >>>> not, but I have not checked carefully. I see that the chain names >>> >>>> assigned to HETATMs were changed in some cases, e.g. 1e3m, where an >>> >>>> ADP single-residue "chain" originally named chain C (before the 2007 >>> >>>> remediation) is now deemed to be part of chain A (and its position >>> >>>> was moved to the end of the file, after all ATOM records). Since I >>> >>>> have been unable to get pre-March-17 snapshot PDB files (the >>> >>>> snapshot.wwpdb.org server is unresponsive) I am not sure when each >>> of >>> >>>> these changes were made. >>> >>>> >>> >>>> 2. Was the changing of chain orders in the March 17 remediation >>> >>>> intentional? If so, is the new order specified somewhere in the 3.20 >>> >>>> documentation? I can see no pattern to the new chain orders (see >>> >>>> examples below). >>> >>>> >>> >>>> 3. Were chain orders ever changed in files that contain only protein >>> >>>> chains (no nucleic acids)? >>> >>>> >>> >>>> 4. Will the changes in chain order be retained permanently >>> (requiring >>> >>>> substantial repairs to Proteopedia.Org)? >>> >>>> >>> >>>> Observations: >>> >>>> >>> >>>> We first noticed the broken molecular scenes in Proteopedia in cases >>> >>>> that involved DNA. Therefore I have so far limited my inspection of >>> >>>> PDB files to those containing both protein and DNA. >>> >>>> >>> >>>> Since the snapshot ftp server is unresponsive today, my comparisons >>> >>>> were all made between files I had saved before the 2007 remediation >>> >>>> (typically saved 2001-2004), and current files. We have reason to >>> >>>> suspect that changes in chain ordering occurred in the March 17, >>> 2009 >>> >>>> remediation, but I cannot verify this for the cases below. >>> >>>> >>> >>>> Some files have NO CHANGE in chain order: >>> >>>> 1d66: DE (DNA), AB (protein). >>> >>>> 1osl: (an NMR multiple model file) AB (protein), CD (DNA). >>> >>>> 1e3m: old AB (protein), C (single residue ADP HETATM "chain"), EF >>> >>>> (DNA); new AB, EF. (ADP now in chain A at the end, thus changing >>> ATOM >>> >>>> serial numbers.) >>> >>>> Thus there appears to be no requirement for nucleic acid or >>> >>>> protein chains to come first. >>> >>>> >>> >>>> Some files that had protein first were rearranged to put DNA first: >>> >>>> 1aoi: old ABCDEFGH (protein), IJ (DNA); new IJ, ABCDEDFH. >>> >>>> 1fzp: old DB (protein), WK (DNA); new WK, DB. >>> >>>> 1hcr: old A (protein), BC (DNA); new BC, A. >>> >>>> Thus there appears to be no requirement that chains be in >>> >>>> alphabetic order. >>> >>>> >>> >>>> One file had an RNA chain moved to BETWEEN two DNA chains, leaving >>> >>>> protein before DNA: >>> >>>> 1qln: old A (protein), TN (DNA), R (RNA); new A (protein), N >>> >>>> (DNA), R (RNA), T (DNA). >>> >>>> The new order happens to be alphabetical by chain name, but >>> >>>> this is not true in other files (see above). >>> >>>> >>> >>>> I did not happen to come across a case where DNA chains preceded >>> >>>> protein in the old format, with protein being moved before DNA in >>> the >>> >>>> new format. >>> >>>> >>> >>>> There also appears to be no requirement that chains be in the order >>> >>>> given in the COMPND records. Examples where the order differs in the >>> >>>> new files: 1flo, 1qln. >>> >>>> >>> >>>> Sincerely, -Eric >>> >>>> >>> >>>> /* - - - - - - - - - - - - - - - - - - - - - - - - - - - >>> >>>> Eric Martz, Professor Emeritus, Dept Microbiology >>> >>>> U Mass, Amherst -- http://Martz.MolviZ.Org<http://martz.molviz.org/> >>> >>>> >>> >>>> Top Five 3D MolVis Technologies http://Top5.MolviZ.Org<http://top5.molviz.org/> >>> >>>> 3D Wiki with Scene-Authoring Tools http://Proteopedia.Org<http://proteopedia.org/> >>> >>>> Biochem 3D Education Resources http://MolviZ.org<http://molviz.org/> >>> >>>> See 3D Molecules, Install Nothing! - http://firstglance.jmol.org >>> >>>> ConSurf - Find Conserved Patches in Proteins: >>> http://consurf.tau.ac.il >>> >>>> Atlas of Macromolecules: http://atlas.molviz.org >>> >>>> Workshops: http://workshops.molviz.org >>> >>>> World Index of Molecular Visualization Resources: >>> >>>> http://molvisindex.org >>> >>>> PDB Lite Macromolecule Finder: http://pdblite.org >>> >>>> Molecular Visualization EMail List (molvis-list): >>> >>>> http://list.molviz.org >>> >>>> Protein Explorer - 3D Visualization: http://proteinexplorer.org >>> >>>> - - - - - - - - - - - - - - - - - - - - - - - - - - - */ >>> >>>> >>> >>>> >>> >>>> TO UNSUBSCRIBE OR CHANGE YOUR SUBSCRIPTION OPTIONS, please see >>> >>>> https://lists.sdsc.edu/mailman/listinfo.cgi/pdb-l . >>> >>> >>> >>> >>> -- >>> ****************************************************************** >>> John Westbrook, Ph.D. >>> Rutgers, The State University of New Jersey >>> Department of Chemistry and Chemical Biology >>> 610 Taylor Road >>> Piscataway, NJ 08854-8087 >>> e-mail: jw...@rc... >>> Ph: (732) 445-4290 Fax: (732) 445-4320 >>> ****************************************************************** >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> >>> _______________________________________________ >>> Jmol-users mailing list >>> Jmo...@li... >>> https://lists.sourceforge.net/lists/listinfo/jmol-users >>> >>> >> >> >> ------------------------------------------------------------------------------ >> >> _______________________________________________ >> Jmol-users mailing list >> Jmo...@li... >> https://lists.sourceforge.net/lists/listinfo/jmol-users >> >> > > > -- > Robert M. Hanson > Professor of Chemistry > St. Olaf College > 1520 St. Olaf Ave. > Northfield, MN 55057 > http://www.stolaf.edu/people/hansonr > phone: 507-786-3107 > > > If nature does not answer first what we want, > it is better to take what answer we get. > > -- Josiah Willard Gibbs, Lecture XXX, Monday, February 5, 1900 > > > ------------------------------------------------------------------------------ > > _______________________________________________ > Jmol-users mailing list > Jmo...@li... > https://lists.sourceforge.net/lists/listinfo/jmol-users > > |