Meta-Path Based Gene Ontology Profiles for Predicting Drug-Disease Associations

Drug repositioning, discovering new indications for existing drugs, is known to solve the bottleneck of drug discovery and development. To support a task many in silico methods have been proposed predicting drug-disease associations. A meta-path based approach, which extracts network-based information through paths from disease, can produce comparable performance with less required when compared other approaches. However, typically use counts extracted discard intermediate nodes those although they are very important indicators, such as drug- disease-associated proteins. Herein, we propose an ensemble learning method Meta-path Gene ontology Profiles Drug-Disease Associations (MGP-DDA). We exploit gene (GO) terms link drugs diseases their associated functions act drug-GO-disease tripartite network. For each pair, MGP-DDA utilizes meta-paths generate novel profiles GO functions, termed profiles. train bagging boosting classifiers features recognize (positive) unknown (unlabeled) Consequently, outperforms state-of-the-art yields precision 88.6%. By MGP-DDA, eminent number associations supporting evidence ClinicalTrials.gov (37.7%) ensures practicality our repositioning.