A KNOWLEDGE-BASED PROTEIN MODELING SYSTEM AND ITS EVALUATION
基于结构比较的蛋白质模建系统及其评估

Based on previous elaborate researches on knowledge-based protein modeling, a new software of modeling a protein of known sequence but unknown structure from structures of homologous proteins is integrated. It consists of two major components, the backbone modeling and side-chain placement, totally including seven modules. The restilts of 11 families of protein modeling show that the system is of the advantages of high pration, less CPU-time and full automation. This paper also predictively models other four proteins, whose structures are unknown nowadays (a neutral phospholipase A2, a lignad for CD40, mouse mast cell protease and human interleukin-13).