[BioPython] Extracting residue list from PDB

[Boris Steipe]

SEQRES and ATOM records have different semantics: the SEQRES is what  
the crystallographer puts into the experiment, the ATOM records is  
what she sees. Presumably you have covalent chain-breaks between  
residues, or parts of the polypeptide chain were not traceable in  
electron density and were omitted.

So even though they numbers are inconsistent, they are both right.

A related issue may be what the natural protein sequence is, as  
opposed to the perhaps truncated molecule in the crystal (presumably  
you are not interested in the propensity of fragments to crystallize,  
but in some biological property), or what the translated sequence is,  
that may have been posttranslationally processed, or even what the  
gene sequence is, that may have been translated with e.g.  
selenocysteine, etc. etc.

So, (as usual) "the way to go" is determined by where you want to go to.

HTH

Boris


On 28 Oct 2005, at 22:14, Gad Abraham wrote:



> On Fri, Oct 28, 2005 at 09:07:53AM -0700, Iddo Friedberg wrote:
>
>
>
>> Hi Gad,
>>
>> The reason the chains seem to be of the wrong length, is that they  
>> are
>> generated from the ATOM records, rather than the SEQRES records.  
>> Those
>> disagree often enough in PDB files.
>> This problem does not exists in mmCIF, I believe.
>>
>> Using your code, I got only six chains, so I cannot comment on the
>> second problem.
>>
>>
>>
>>
>
> I see. I'm consistently getting 10 chains (lengths 161, 804, 97, 176,
> 11, 71, 87, 662, 133, 286) for 1N62.
>
> It seems that parsing the SEQRES is the way go.
>
> Thanks,
> Gad
>
> -- 
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> gabraham at cs.rmit.edu.au
> http://yallara.cs.rmit.edu.au/~gabraham
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> _______________________________________________
> BioPython mailing list  -  BioPython at biopython.org
> http://biopython.org/mailman/listinfo/biopython
>
>
>