科研之友
微信
新浪微博
Facebook
分享链接
摘要
Docking accuracy is primary criterion to evaluate many molecular docking methods or software. In this study, we present a new method of molecular docking in order to improve the docking accuracy. Residue groups are used in this method to construct receptor model, so the flexibility of the receptor can be considered. Information entropy genetic algorithm is responsible for searching the optimal positions of the docking ligand and receptor. In the process of optimization, empirical scoring function is used to evaluate the docking conformations. Multi-population evolution and narrowing down searching space are also used to improve the stability and convergence rate of genetic algorithm. By using the publicly available GOLD test set (with 134 protein-ligand complexes), numerical experiments were presented to evaluate our method. When comparing with other frequently-used docking software (or methods), we gained the higher accuracy. According to Root-Mean-Square Deviation (RMSD), experiments show that there are more than 65% docking results of this study are less than 2.0 angstrom. The average docking time of our method is only about 1/12 of that of Dock6-F (with default flexible parameters), demonstrating our method more suitable for virtual screening and other related works.
- 出版日期2017-10
- 单位淮海工学院
