Mª Victoria Ruiz Pérez wrote: > I´m a begginer user of MODELLER, and I have a question for you: > I´m making docking experiments, and the protein I have got from the PDB > has some gaps, so the docking can´t go on. So, I thought I could model > the protein using its own cristal as template. So I did, but the models > I got don´t fit very well with the original cristal (when I superimpose > them with MacPyMol, the models are very deviated from the cristal). What > can I do to improve the adjustment between models and template?
Modeller should produce models that look like the template where the two are aligned, by construction. Presumably you have a 1:1 alignment in this case, since you're modeling using its own sequence - the sequence identity is 100% ? So in this case the models should look exactly like the template, differing only in sidechain orientations. If not, something is very wrong - you should check the restraint violations in the log file, because the restraints cannot be satisfied.
Ben Webb, Modeller Caretaker