On 5/25/10 2:05 AM, Anna Marabotti wrote: > I obtained a model of a protein using Modeller 9v7. Looking at the resulting > PDB file, I saw that several atoms (in some cases the entire sidechain, in > other cases only one or two atoms) show an occupancy<1 (typically 0,5, but > also 0.3 in some cases).
For residues aligned with a template, the atom occupancies will simply match whatever was in the template file.
> In the template, the corresponding residues show > two or sometimes three alternative conformations of their sidechains, with > occupancy<1, so I imagine that Modeller chose one of the conformations and > transferred directly the information about occupancy in the PDB file it > creates. My question is: what is the criterium that Modeller follows to > choose one alternative conformation? It simply chooses the first one, or are > there other criteria?
It picks the first one; see http://salilab.org/modeller/9v8/manual/node170.html
> And what about the fact that a single atom has an > occupancy<1 without any other alternative conformations? Could this be able > to perturb the protein behavior?
Modeller does not use the occupancy information in the model, so this won't affect the generated models.
Ben Webb, Modeller Caretaker