Closed orbeckst closed 8 years ago
Alternative explanation for the small differences: An important step in rotcon is to fit the spin label to the sidechain to "virtually" attach it, so clearly everything depends on this step. Didn't we change slightly the way that RMS-fitting was done, namely how/when mass weighting was done? That might also explain small differences in distribution.
The new MDAnalysis 0.11.0 release breaks the API and RotamerConvolveMD should be adapted accordingly; see the 0.11.0 release guide (and in particular migrating with ten2eleven.py.