Machine Learning and Computational Approach to Repurposing FDA Approved Drugs
Anisha S Jain · SSRN Electronic Journal · 2020
Drug repurposing is a strategy to study drugs that are approved to treat one condition or disease to check if they are effective as well as safe for treating other diseases. For a given indication, this strategy offers wide range of benefits over developing an entirely new drug. Repurposing of drugs is effective enough to minimize the time and costs in drug development since data is already available on their potential toxicity, formulation and pharmacology. Computational research in drug based industries is thought to have an effective approach for Research and Development process to accelerate the rate of drug designing at molecular level. Such an approach has already led to the identification and experimental validation of novel therapeutic indications. Drug repurposing can be forecasted using various machine learning models based on different entities. In this study, we present a strategy to drug repurposing by predicting indication for a specific disease based on expression profiles of drugs, with a focus on oncogenes. The availability of high-performance computing, and databases of various forms have also enhanced the ability to pose reasonable and testable hypotheses for drug repurposing, rescue, and repositioning. Drug repurposing is being considered as a supervised learning problem and applying distinct state-of-the-art machine learning methods for prediction. Drugs that are not initially indicated for a specific disease but have high predicted probabilities serve as superior candidates for repurposing.