The rules for alt. confs. are such that a-a and b-b will always clash. Let A,B chain labels, and a,b,c be altloc labels. You could assign A-a, A-b for the two conformers in chain A, and B-a, B-c for the two conformers in chain B. This way the second conformers will have no nonbonded interactions.

Ralf

On Thu, Dec 1, 2011 at 2:13 PM, Sozanne Solmaz <ssolmaz@mail.rockefeller.edu> wrote:

Dear all,
I have the following problem: I have a homodimer of two protein chains A and B, which I refine in Phenix, and one residue has 2 alternate conformations (a and b) for each chain. If the residues of both chains A and B are in the alternative conformation b, they clash. However, I would expect that never both residues are in conformation b, and therefore only the combinations aa, ab and ba of alternative conformations exist. My problem is, that during the refinement, the two b conformations of the residues clash, which drives the clash-score of the overall structure to a very high value and repulsion pushes the two b-conformations apart during refinement, therefore I cannot model them properly. Is it possible in Phenix to define these two conformations as non-clashing and non-repulsive?
Thank you, best regards, Sozanne Solmaz.
_______________________________________________
phenixbb mailing list
phenixbb@phenix-online.org
http://phenix-online.org/mailman/listinfo/phenixbb