Thomas Juettemann wrote: > could someone enlighten me how MODELLER treats non-standard residues. > I am using Dunbrack's S2C as a source for ATOM sequences. In the PDB > file 1M2W, I have several selenomethionine (MSE) residues, marked as > HETATM. Running MODELLER, it seems to convert them to methionine > (which I appreciate) and successfully produces a model. > > My question is: How does MODELLER check / replace HETATM that have a > residue identifier that is in the permitted range (not 'X', '.' or > '-')? I did not not set env.io.hetatm to true.
Modeller does not do any special processing except for the MSE -> MET translation that you have already noticed. If you have env.io.hetatm=False, no HETATMs (except MSE) will be read from the PDB file.
If you set env.io.hetatm=True, Modeller will read all HETATM records (except for waters, which it will only read if you set env.io.water). It doesn't do any replacements. For matching, '.' will match anything, and treat it as a rigid body. 'X' will match only the UNK residue type. Other residue types can be matched if a suitable one letter code is listed in modlib/restyp.lib; for residues not in that file, see http://salilab.org/modeller/FAQ.html#8
Ben Webb, Modeller Caretaker