Predicting Binding Sites of Hepatitis C Virus Complexes Using Residue Binding Propensity and Sequence Entropy

摘要

Hepatitis C virus (HCV) remains a dangerous health problem for the reason that the mechanism of hepatocyte infection is still unknown. Hence, much attention has been put on the problem of interaction between HCV and human proteins. However, the research is still standing at the beginning point due to the lack of structure information of HCV and human proteins. We extracted the most representative set of 18 complexes all known HCV protein complexes involving human proteins, and computed the binding propensity of each residue and sequence entropy of each HCV protein, analyzed the most representative set of 18 complexes. Using a radial basis function neural network (RBFNN), we predicted binding sites with an overall sensitivity of 77%. The approach will help understand the interaction between HCV and human proteins.