Merge pull request #378 from RosettaCommons/vmullig/fix_omega
Vmullig/fix omega
Fixes behaviour of omega score term for beta-amino acids. (Thanks to Andy for pointing out the problem, and for ultimately coming up with the solution.) No integration test changes expected; all unit tests pass (including newly-added one).
Tasks:
Switches the test for beta-amino acids to the is_beta_aa() function in ResidueType (instead of my earlier kludge of looking for a "CM" atom).
Fixes the behaviour for beta-amino acids.
Moves all of the code for figuring out which mainchain torsion is omega to one place.
Adds a unit test for beta-amino acids.
Alters the derivative calculation WRT phi and psi in the special case of beta-amino acids to use mainchain torsions 2 and 3 instead of 1 and 2.