On 26/04/2008, Eric Martz <[EMAIL PROTECTED]> wrote: > > Thank you, Rolf, for explaining biological units. It will be tremendously > useful if Jmol can generate them, when asked via script commands, from the > symmetry operations in the PDB (REMARK 350) or mmCIF file header. > > I just wanted to add a few points. > > 1. VERY LARGE BIOLOGICAL UNITS. In some cases, notably virus capsid > proteins, the biological unit is very large (e.g. an entire virus capsid). > As an all-non-hydrogen atom PDB file, these can easily exceed 100 megabytes. > An example would be 1sva, a capsid protein from Simian Virus 40. > > Thus, if Jmol is enabled to use REMARK 350 (or the equivalent in mmCIF), > there will need to be some mechanism to handle cases where the result would > exceed the capacity of java memory. Hopefully this can be done gracefully, > rather than just staying blank forever or freezing/crashing. > > In my tests long ago with Jmol applet 10.00.46 ( > http://molvis.sdsc.edu/fgij/bigpdb.htm ), Jmol applet > failed to display PDB files containing about 200,000 atoms (3ezb with 40 NMR > models, a 17 megabyte PDB file uncompressed), but displayed 1jj2, which > contains 98,000 atoms (an 8 megabyte PDB file uncompressed). I attributed > this cutoff to a java memory issue. > > > 2. BIOLOGICAL UNITS SMALLER THAN THE ASYMMETRIC UNIT. Sometimes the > asymmetric unit (what is published in the PDB file) contains more than one > biological unit. It would be ideal if Jmol could handle these as well. It > could either show only the first biological unit, or possibly could set them > up in its memory as separate models or frames. An example is 1b6b, which > contains a homodimer. The dimerization is believed to be a artifact of > crystallization. Thus, there are two biological units in the asymmetric unit > (PDB file). At > http://www.pdb.org/pdb/explore/explore.do?structureId=1B6B > you will be shown static snapshots of the asymmetric unit, and each monomer. > > The requirement to separate the chains is indicated thus in the PDB format: > > > REMARK 350 GENERATING THE > BIOMOLECULE > REMARK 350 COORDINATES FOR A COMPLETE MULTIMER REPRESENTING THE > KNOWN > REMARK 350 BIOLOGICALLY SIGNIFICANT OLIGOMERIZATION STATE OF > THE > REMARK 350 MOLECULE CAN BE GENERATED BY APPLYING BIOMT > TRANSFORMATIONS > REMARK 350 GIVEN BELOW. BOTH NON-CRYSTALLOGRAPHIC > AND > REMARK 350 CRYSTALLOGRAPHIC OPERATIONS ARE > GIVEN. > REMARK > 350 > > REMARK 350 BIOMOLECULE: > 1 > REMARK 350 APPLY THE FOLLOWING TO CHAINS: > A > REMARK 350 BIOMT1 1 1.000000 > 0.000000 0.000000 > 0.00000 > REMARK 350 BIOMT2 1 0.000000 > 1.000000 0.000000 > 0.00000 > REMARK 350 BIOMT3 1 0.000000 > 0.000000 1.000000 > 0.00000 > REMARK 350 BIOMOLECULE: > 2 > REMARK 350 APPLY THE FOLLOWING TO CHAINS: > B > REMARK 350 BIOMT1 1 1.000000 > 0.000000 0.000000 > 0.00000 > REMARK 350 BIOMT2 1 0.000000 > 1.000000 0.000000 > 0.00000 > REMARK 350 BIOMT3 1 0.000000 > 0.000000 1.000000 > 0.00000 > > > Incidentally, even when you are viewing the snapshot of the biological unit > at RCSB, clicking on the visualizers (e.g. KiNG, Jmol, etc.) does NOT show > you the biological unit. To get it, you must use the Structure tab (upper > left), then open Download Files, and the last item(s) will be the biological > unit(s), gzipped. > > A series of examples of different outcomes of biological units is given at > http://proteinexplorer.org/pqs.htm > This document first explains the operation of the Probable Quaternary > Structure (PQS) Server at the European Bioinformatics Institute, and then > gives the examples. > > PQS examines the contacts within the crystal, and makes an automated > judgement as to whether each contact is the result of a co-evolved specific > oligomeric interaction (large area, often hydrophobic core and/or salt > bridges), or alternatively is an artifactual "crystal contact" (small area > between hydrophilic surfaces). It is usually correct, but sometimes wrong. > It works only when the biological unit forms in the crystal (the usual case, > but not always). > > PQS is in the process of being superceded by a new algorithm implemented as > a server named PISA. > > -Eric
Hi all, Looks like a very useful thread! Below are a few comments. Eric: I was thinking *exactly* the same thing! ;-) Bob: Some friends and I tried to answer that question (what is a BioUnit?) as part of a 'PDB FAQ' that we have been working on. We thought that this question was important and complex enough to get a page all to itself. You can see what we currently have here: http://pdbwiki.org/index.php/Biological_unit The PDB FAQ may be of general interest to this list. It can be found here: http://pdbwiki.org/index.php/PDB_FAQ The FAQ is part of a PDB-Wiki project that we announced on PDB-L: https://lists.sdsc.edu/pipermail/pdb-l/2008-April/004321.html Rolf / Eric: Thanks for the information about BioUnits. If it is OK with you I will try to incorporate your comments into the above FAQ article? Or if you are really keen, you are free to edit the article yourself! Eric: I just realized, probably the easiest way to get the BioUnit data is to use the biounit division of the PDB FTP site: ftp://ftp.wwpdb.org/pub/pdb/data Under this branch you can find exactly the data that you want. Namely, the results of applying the symmetry operations in remark 350 to the ASU. Note that there are many errors in the biounit division of the PDB. Trivially these include things like dimers that are *not* dimers under physiological conditions, monomers that *are* dimers under physiological conditions etc., etc. This is why the 'databases of predicted protein quaternary structures' are so important. My favorite is PiQSi: http://www.supfam.org/elevy/piqsi/piqsi_home.cgi I hope the above helps. Thanks all for the replies, Dan. > ------------------------------------------------------------------------- > This SF.net email is sponsored by the 2008 JavaOne(SM) Conference > Don't miss this year's exciting event. There's still time to save $100. > Use priority code J8TL2D2. > http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone > _______________________________________________ > Jmol-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/jmol-users > > -- hello ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ Jmol-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-users
