Predicting drug-target interactions based on an improved semi-supervised learning approach

Identifying interactions between compounds and target proteins is an important area of research in drug discovery and there is thus a strong incentive to develop computational approaches capable of detecting these potential compound-protein interactions efficiently. In this study, two different methods were first utilized to construct chemical and genomic spaces, respectively. Then two spaces were combined into a integrate space to discover the potential compound-target pairs in the known drug-target interaction data by an improved semi-supervised learning method (FLapRLS). The results demonstrated that this prediction method is effective.