Investigation of Drug Repurposing Opportunities Using Side-effects data, Topic Modelling and Clustering Algorithms
Pritish Ranjan, Shyam Sundar Das, Narayanan Ramamurthi · 2023
We describe herein, a promising approach for the selection of drug repurposing candidates using drug side-effects, topic modelling and clustering algorithms. It is based on the principle that drugs with similar side-effects profiles are likely to be effective for the same disease. For this purpose, we employed i) side-effects data from FDA drug labels and SIDER database ii) Principal Component analysis (PCA) with feature scaling for the identification of topic number iii) Latent Dirichlet algorithm for topic modelling and iv) clustering algorithms for classifying drugs using side-effects profile. Based on our study, we recommend, 4 novel repurposing candidates and 14 safer alternatives for three diseases namely, epilepsy and Parkinson's and cancer.