作者: Shengyin Gu , Patrice Koehl , Joel Hass , Nina Amenta
DOI: 10.1002/PROT.23192
关键词:
摘要: Determining the structure of protein-protein complexes remains a difficult and lengthy process, either by NMR or X-ray crystallography. Several computational methods based on docking have been developed to support even serve as possible alternatives these experimental methods. In this paper, we introduce new algorithm, shDock, shape complementarity. We characterize local geometry each protein surface with descriptor, surface-histogram. measure complementarity between two surface-histograms, one protein, using modified Manhattan distance. When match is found surfaces, model generated for complex, which then scored checking collision proteins. tested our algorithm Version 3 ZDOCK benchmark. that 110 out 124 test cases bound in benchmark, was able generate top 3600 candidates complex within an RMSD 2.5 A from its native structure. For unbound predictions, models 54 cases. comparison other shape-based algorithms demonstrates approach gives significantly improved performance both