pdb I/O

Hi there!





I'm struggling to get rid of my ties to insightII for protein site file preparation, and use PMapper to assign protein force field types, like I'd shown at the latest UGM... the tool's FFassign.java, see attached. Actually, FF typing & charge assignment runs smoothly for small proteins like the one attached - the problem is that the output pdb file keeps no track of residue & atoms names!! The file is technically correct, but it ain't anything very useful for further modeling - no backbones, no ribbons, etc. What can be done to fix this?





Cheers!

Hi Dragos,





this behaviour of the PDB export is a serious problem. Not simply a bug that can be fixed with some efforts, but rather the consequence of an ongoing refurbishment of PDB export/import that takes long time. The import of PDB files was changed first (actually, completely rewritten from scratch), and even this has not yet been accomplish. The export was not touched yet, this is why all records are written as HETATM, residue names get lost on the way etc.


We will certainly re-implement PDB export...





In the meantime, I suggest to produce an extra file by FFassign that contains the calculated/predicted atom property data (like charge, extended atom type etc.) and use standard text processing tools (like awk) to add these extras to the orginal pdb file (without actually interpreting the PDB content, just reading and writing it line by line).


Does this make sense?





If you or your team has no experience with these kind of text processing tools we can help.





Does this make sense/help at all?





Regards,


Miklos

Hi, Miklos,





I was afraid you'd say that - in fact, I was actually looking into alternative solutions like you suggested bu the trouble is that atom numbering in the initial file is not the same as atom numbering after pKa calculation and H assignment... so I'm in trouble when trying to make any matches/new atom insertions with respect to the original .pdb - is there any workaround for that?





Cheers!


Dragos

Hi Dragos,





I will check this. I mean I do not doubt that it is as you describe but it is rather surprising that atom indices are messed up. I hope it'd be easy to fix... This would be more straightforward than puzzling with compicated workarounds.





I'll get back to you soon.





Miklos