Computational drug repositioning by ranking and integrating multiple data sources

摘要: Drug repositioning helps identify new indications for marketed drugs and clinical candidates. In this study, we proposed an integrative computational framework to predict novel drug both approved molecules by integrating chemical, biological phenotypic data sources. We defined different similarity measures each of these sources utilized a weighted k-nearest neighbor algorithm transfer similarities nearest neighbors prediction scores given compound. A large margin method was used combine individual metrics from multiple into global metric. large-scale study conducted repurpose 1007 against 719 diseases. Experimental results showed that the outperformed similar previously developed approaches. Moreover, also ranked information based on their contributions prediction, thus paving way prioritizing building more reliable models.