A new method to identify repositioned drugs for prostate cancer

With the merits of faster development time and reduced risk, identifying new indications for marketed drugs draws more and more attention. In particular, repositioning drugs with known indications has become an hot topic in the area of computational systems biology. However, one of the common shortcomings for most of the previous methods is the ignorance of side effect, i.e., drug through primary targets and off targets might induce both desired and unintended effects respectively, which could not appropriately evaluated in most of existing methods. In this paper with a new measure considering both efficacy and side effect, we developed a new method for identifying the repositioned drugs against prostate cancer by evaluating the mutual relations of the gene expression levels between prostate cancer samples and those induced by bioactive compounds. In this measure, the overlap between gene sets that were oppositely regulated in disease state and drug treatment state was quantified by jaccard index as drug's efficacy while the overlap between essential genes and positively correlated genes (or regulated just after drug treatment) was quantified by jaccard index as drug's side effect, which were balanced with a parameter λ. The preliminary results on repositioning drugs for prostate cancer verify the effectiveness and efficiency of the new method.