O-3: Drug Repositioning by Merging Gene Expression Data Analysis and Cheminformatics Target Prediction Approaches

The transcriptional responses of drug treatments combined with a protein target prediction algorithm was utilised to associate compounds to biological genomic space. This enabled us to predict efficacy of compounds in cMap and LINCS against 181 databases of diseases extracted from GEO. 18/30 of top drugs predicted for leukemia (e.g. Leflunomide and Etoposide) and breast cancer (e.g. Tamoxifen and Ouabain) were already proven effective. LC50 of a predict compound, Fenbendazole, on HL60 and BMSC cell lines was experimentally evaluated to be 127nM and 5066nM (48hrs). Novel predicted drugs for differentiating stem cells to cardiomyocytes such as Famotidine enhanced expansion rate and heart-beat strength of Embryoid Bodies and expression of cardiac precursor markers(Nkx2-5, Gata4, Mef2c) and cardiac markers(Actc1, Tagln, Myh6) considerably during and post cardiac progenitor cell (CPC) formation. Combining target prediction with gene expression analysis enabled us to predict effective compounds and targets for diseases and differentiation of stem cells.