well i m looking forward for suggestions on the criteria for selecting atoms for angle calculations between a protein and a ligand.. any suggestions would be helpful.
Torsion angle is meant for 4 atoms connected in the same chain. If you want to do for protein and ligand atoms you can consider them as virtual torsion angle , so select the 4 atoms of your choice and measure torsion through any tools.
thanks @Balaji and @Ramin for replying.. but then when we were working on Rosetta we used to define certain angle values between the protein and the ligand in order to fix the ligand to be in the relative position.. what is the logic behind it? and is there any coincidence with my question?
as long as you are interested in a comparison between two states, what Ramin said sounds reasonable because you want to see how it changes (in MD, upon mutation of other components of the complex, ...). Of course if you are mutating some residue in the protein you don't choose any atom in it.
In the logic of Rosetta, fixing such dihedral constraints the ligand to be oriented in the direction chosen, but it can still rotate on such axis (for further rearrangements).