Ensemble Machine Learning Approaches in Molecular Fingerprint based Virtual screening

Drug discovery is the process of identifying new biologically active compounds which has drug likeness characteristics from a large set of chemical compounds. Virtual screening technique expedites drug discovery by ruling out most of the non drug candidates at an early processing stage. Molecular fingerprints are the choice of molecular data representation for the past few years. Here we are discussing some of the recent advancements in fingerprint based virtual screening and exploring the perks of implementing the concept of hybridism in various crucial components of similarity based virtual screening. We have also conducted a comparative study on the accuracy on the 2 commonly used classification models in similarity based virtual screening, that is support vector machines[SVM] and random forest and discussed the results that we observed.