Fig. 10.

Schematic representation of similarity search. The figure represents the outline of the workflow, which can be divided into three parts: (A) use of deep-learning methodology for the prediction of piperine targets, (B) similarity search from the ZINC and PubChem databases using a ML approach to get top 100 best hits based on piperine, (C) validation and comparison of how well the compounds predicted the FDA-approved drugs for the selected targets. Reprinted (adapted) with permission from (Khan et al. 2019). Copyright (2019) The Royal Society of Chemistry