rosie4loop Lv 1
So, in my opinion, if the scoring function is more optimized and if everyone is sampling the energy landscape of the same ligand (like sampling protein folding pathway, the only variable is the conformation but not the protein sequence), then sending the score+structure every certain period of time would be useful. Otherwise there's likely too many variables in the same puzzle, and the sample size of each unique ligand would be insufficient (plus bias to the initial state).