Optimized distance-dependent atom-pair-based potential DOOP for protein structure prediction.

摘要: The DOcking decoy-based Optimized Potential (DOOP) energy function for protein structure prediction is based on empirical distance-dependent atom-pair interactions. To optimize the interactions, native structures are decomposed into polypeptide chain segments that correspond to structural motives involving complete secondary elements. They constitute near ligand-receptor systems (or just pairs). Thus, a total of 8609 were prepared from 954 selected proteins. For each these hypothetical systems, 1000 evenly sampled docking decoys with 0-10 A interface root-mean-square-deviation (iRMSD) generated method used before protein-protein docking. neural network-based optimization was applied derive optimized parameters using so mimics funnel-like landscape interaction between systems. our hierarchically models overall structures. resulting tested several commonly decoy sets recognition and compared other statistical potentials. In combination torsion potential term which describes local conformational preference, atom-pair-based outperforms reported functions in correct ranking variety sets. This especially case most challenging ROSETTA set, although it does not take account side orientation-dependence explicitly. DOOP prediction, underlying database their freely available at http://agknapp.chemie.fu-berlin.de/doop/.