I have some known protein complex(two chains) and would like to calculate the binding energy of each complex. I don't want to dock any more but just score the binding energy. I have done a fast relax for chain A and chain B, and then combine them into the same pdb file, then use docking prepack to generate the input structure. (should I avoid the fast relax for individual chain but direct use dock prepack to generate the input structure?)
But I don't know which option to calculate the binding energy. Thanks for help.
You should use the InterfaceAnalyzer application (probably with the pack_separated and pack_together options. note see the app documentation for the actual names of the options). The prep of your structure seems ok for the app, but I would probably relax the combined chains. Preferably, the together and individual chains would be relaxed, but that is not currently an option in IAM.
You could also use the score_jd2 app to just get the score for each chain apart (add this together) and the score of the complex to calculate dG.
If I understand correctly, the binding energy can be calculated by
E_complex - E_chainA - E_chainB.
So I can relax the complex first and then extract individual chains from the relaxed complex. Then I can use score_jd2 to get E_complex, E_chainA and E_chainB.
Am I correct? Thank you!
Yes, if you use the score_jd2 app, I would relax the complex together and then relax each chain separately to get the energies. You'll probably want to relax with coordinate constraints as to not mess with the structure too much. Or just relax the whole complex with coordinate constraints and run it through InterfaceAnalyzer app, packing together and packing separately via options - that way you have at least some rearrangement of the interface side chains that are no longer in the interface upon separation.
Thank you very much! But I am confused now.
Should I use docking_prepack_protocol.linuxgccrelease to relax the complex structure or use fast-relax program to relax the complex structure (also the monomer structure)? I am using score_jd2 app to calcuate the energy. Thanks!
docking_prepack_protocol is definitely not what you want here. FastRelax of both complex and individual chains (split before the relax) with coordinate constraints should work well. See the fast relax docs for how to enable it. Typically I will FastRelax the complex with coordinate constraints first, and then use the InterfaceAnalyzerMover application to get dGs, making sure to pack separately. I am not aware of any papers correlating experimental dGs by using one method vs the another.