The average recov ery rate in case of Temsirolimus chemical structure drug repositioning candidates gener ated by ClusterONE was 95% while in case of Louvain clustering it was 85%. This demonstrates that the drug repositioning candidates we have discovered are robust and that additional edge removal or addition will not affect the output significantly. Drug repositioning candidates and literature based evaluation From the 98 clusters found by Louvain clustering, 11160 drug repositioning candidates were generated. In case of 110 ClusterONE generated clus ters, 2518 drug repositioning candidates were extracted. There were 2501 drug repositioning candidates found by both of the cluster ing approaches. We used these pairs to perform a literature based and clinical trials search using CoPub and a carefully designed PubMed search using NCBIs E Utilities feature.

In the Figure 2 we show the modules which con tained drug repositioning pairs with literature evidence. In the fol lowing sections we discuss two case studies wherein our discovered drug repositioning candidates matched with those in clinical trials and literature. Vismodegib and Gorlin syndrome Two of the drug repositioning candidates in our results that overlapped with the literature reports and clinical trials were derived from a cluster with drugs vismodegib and erismodegib and diseases basal cell carcinoma and Gorlin syndrome. Interestingly, vismodegib, an oral inhibitor of the hedgehog pathway, is the first drug approved by the US Food and Drug Administra tion for the treatment of locally advanced and metastatic BCC.

Additionally, another study reported the efficacy of vismodegib on patients with Gorlin syndrome, a rare autosomal dominant disorder in which those with the disease are prone to developing multiple BCCs at an early age. In our ana lyses, vismodegib and Gorlin syndrome do not share a common gene but are still clustered together because of the pathway based connectivity. This demonstrates the utility of our approach in using feature based heterogeneous networks to identify drug repositioning candidates. g secretase inhibitors, NSAID, Alzheimers and Hidradenitis suppurativa Another interesting set of examples in our study were related to Alzheimers disease and g secretase inhi bitors and NSAID which have been shown as potent reducers of levels of b amyloid.

In our study, AD and hidradenitis suppurativa Drug_discovery were clustered along with the g secretase inhibitors and tarenflurbil. Since several studies have further info implicated b amyloid peptides in the etiology of Alzheimers disease and because Ab is pro duced by the proteolytic cleavage of the amyloid precur sor protein by b and g secretase, g secretase inhibition is thought to have a therapeutic benefit for AD. How ever, all these drugs failed in phase III trials because they either worsened cognition and/or increased the risk of skin cancer.