pdbsplit() action

[Former user]

pdbsplit() currently splits raw pdbs by chain but sometimes chain A also includes hetatm records that are included in the split pdb. I believe it would be useful to not include hetatm since this can confound analysis of nma.pdb() although it uses CA only atoms.

[Lars Skjærven]

Hmm.. not sure I agree. Consider e.g. PDB id 1cdk. It contains HETATM residue TPO (in-chain). Omitting this residue would affect the fluctuations of adjacent residues.

Re-introducing the hetatm argument in pdbsplit could be useful for other reasons, but I don't think you want to omit them for NMA.

[Lars Skjærven]

I also note that the method for identifying of c-alpha atom is improved in version 2.2. You might want to check out the updates of function atom.select if you experience trouble in 2.1.

The c-alpha atoms are defined as any "CA" atom in a residue with residue name matching the residue names in the attached aa.table data frame. See aa.table$aa3 for a complete list of residue names. Thus, if your protein contains a HETATM with a CA atom, but with residue name not in aa.table$aa3 pdbaln() will produce a pdbs object without this residue.