Skip to main content
. 2022 Jun 16;10:902123. doi: 10.3389/fpubh.2022.902123

Figure 1.

Figure 1

An integrated machine learning and molecular docking-based ensemble approach for drug repurposing. (A) It shows that the drug and disease data collected from the public databases are first arranged and then transformed into a feature matrix, (B) followed by training and testing of the predictive models (RF and TE). (C) After training and testing, the data for the COVID-19 host targets are prepared from the literature review and fed to the predictive model for getting the potential repurposed drugs. Potential hits are then validated by using literature mining and collecting the evidence in the form of papers, patents, and database evidence. For the remaining predictions, molecular docking is performed to find the binding affinity of the drugs. Finally, the drugs with the highest binding affinity are prioritized as the potential repurposed candidates ready for the preclinical and clinical tests.