[Biopython] PDB Construction Error

[Bhima A van der Molen]

Hi Peter,

At the end of last week you asked me to send you some PDB files that
would have been generated by the hsexpo script, which have been giving
me trouble.  I am sorry it has taken me this long to get back to you,
I've been away from the office since the end of last week. 

The files I am working with are not directly generated by hsexpo. I used
hsexpo to create PDBfiles with solvent accessibility data in the
b-factor column.  I have used these output files as input into my own
program which randomises the solvent accessibility values for each
residue in the file.  This is part of a boot-strapping exercise to
verify my statistical analysis.

There is however a single PDB file constructed by the above method,
which is triggering the PDBConstructionError  I have not been able to
determine which one it is as yet, because I am running an analysis on
more than 25,000 PDB files and it has proven difficult to catch. If I
find something that looks like it is due to something in BioPython I'll
be sure to let you know. 

This is really just to let you know that the hsexpo script is outputting
files which in my experience works well with the PDBParser.  

Thanks for your support.

Bhima

On Thu, 2010-05-06 at 11:34 +0100, Peter wrote:

> 
> On Thu, May 6, 2010 at 11:17 AM, Bhima Auro van der Molen wrote:
> >
> > Hi Peter,
> >
> > Thanks for the response.. I thought it might be a typo somewhere however I don't
> > know enough about the BioPython code to fix it myself yet..
> >
> > I am a bit curious that the is being raised is a PDB Construction error...
> >
> > As I said, I am storing DSSP data in the b-factor column, as is done using the
> > hsexpo.py script, but in order to be thorough in a statistical analysis I need to
> > randomise that data and assign it randomly to residues in the PDB file.. I am
> > not making any changes to the atomic co-ordinates of any of the residues.. the
> > only data that gets re-written in this process is the b-factor column. What would
> > cause a PDBConstruction Error to be raised?
> >
> > Thanks
> >
> > Bhima
> 
> Hi Bhima,
> 
> The PDB Construction error (or rather PDBConstructionException) is being
> raised in  the _parse_coordinates method, and indicated one or more of the
> three atomic coordinates could not be turned into floats. Perhaps they are
> badly aligned (in the wrong column)? Could you send me the problem PDB
> file (off list - sending attachments to mailing lists is a bad idea)?
> 
> If you are creating the problem PDB file with the Biopython hsexpo script
> this may indicate a problem elsewhere in Biopython (perhaps in the PDB
> output code).
> 
> Peter