>> This might make it somewhat easier ... in fact, most of the data
>> structures to support this are already in place.
>>
> then perhaps we don't need to read the MODRES at all.

I think that would be good. Esp if MODRES records are often missing from
.pdb file.

> as long as Jmol
> can separate modified residues from covalently-bound ligands.

I do not understand

>> I also build a data structure called viewer.pdb.PdbPolymer which is
>> similar, but not the same. It is the data structure for a real
>> *chain*.
>>
> we spoke about polymer vs chain in the autumn, as I recall.  this sounds
> like a Very Useful Thing.

Correct, you helped me figure this out ... leading to
dramatic/breathtaking code restructuring :-)

>> In order for a residue to make it into a PdbPolymer it needs to have:
>>  - amino nitrogen
>>  - alpha carbon
>>  - carbonyl carbon
>>  - carbonyl oxygen
>>
>> This should happen independent of the group name.
>>
> sounds good.  the alpha carbon should be easy; it has a unique
> identifier CA.

Unfortunately, many .pdb files do not follow the PDB spec.

The current code tries to be flexible and will accept the first 'C' as a 'CA'

>  the other three are simply defined as N, C, and O.  in
> standard amino acid residues this is a non-issue - since all other atoms
> are well-defined by their chemical position (CB, NZ, OD1, OD2, etc.),
> the backbone atoms are obvious.  I'm not sure this is true for modified
> residues - I'll have to check the PDB spec.
>
>
>
>> Here is how you can test this out:
>>  1. load up a model that has a modified residue
>>     (with or without a MODRES record ...
>>     this record type is currently ignored)
>>  2. turn on backbone
>>  3. see whether or not the residue appears in the backbone
>>
>> If it does, then most of the work is done.
>>
> ok; I'll check.
>
>
>> Q: Can you please add a small file that contains a modified protein
>> residue to the samples directory ... with a name like
>> modifiedProteinResidue.pdb
>>
> will do this.

Eric has recommended that we start holding on to strange/test cases. Let's
organize this a little better by creating some subdirectories:

  samples/pdbtest/modifiedGroup

Put your sample files in there, with their pdbId names.

We'll add other test cases to sibling subdirectories.

>> you said:
>> > anything with an alpha carbon and amino group bound in the
>> > backbone of a protein chain would be considered protein
>>
>> Q: Are there cases where something can get into the protein chain
>> without having
>> N-C-C=O ?
>>
> yes; any covalently-bound ligand.  I'll scare up some examples from the
> PDB.

OK, stick them in the modifiedGroup subdirectory (or another subdirectory
of pdbtest if that is better)

Something for you to think about ... What happens to
backbone/trace/strands/ribbons/cartoons when you stick one of these
'covalently-bound ligands' in the pdbpolymer?


>> Q: If so, what are the criteria for determining whether or not
>> something is in the polymer chain?
>>
> assuming you mean the PDB criteria, I'll have to look.

Actually, I didn't mean the PDB criteria. I think you gave the answer
above ... 'any covalently-bound ligand'


> I'll get back with you after doing some more research.

Good


Miguel





-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
Jmol-developers mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jmol-developers

Reply via email to