Understanding Side-Chain Motions using Statistical Torsion Angle Potential

To understand protein functions, an accurate protein structure is required. Because protein side-chain affects protein-protein interaction and protein-molecule interaction in tertiary protein structure, side-chain mobility is particularly important for understanding protein function. In general, determination of side-chain population is experimentally difficult, so we used computational method: Statistical Torsion Angle Potential (STAP) with Φ-χ1, Ψ-χ1 and χ1-χ2 combinations and simulated annealing protocol are used to predict side-chain mobility. Order parameters (S2) of methyl carbon and their correlation coefficients are evaluated for accuracy check. The computational prediction of χ1, χ2 rotamer population is performed for 17 target proteins. Seven target proteins are used for training set to find optimal values of van der Waals and STAP force constant. Ten target proteins are used for test set. This study can be used to characterize the χ1, χ2 conformation of exposed residues and to understand side-chain motions.