We have previously developed screening and prognostic models of colorectal cancer (Palaniappan et al., COADREADx: apalanialab.shinyappps.io/COADREADx ). In this project, we propose a data-driven approach combining the latest techniques in AI-based drug design to develop generalized frameworks for discovering drugs for colorectal cancer. The biomarkers identified in COADREADx will be expanded with STRING and searched in multiple sources including DepMap and DRUGNOMEAI to validate and finally prioritise by Robust Rank Aggregation. After the identification of the potential targets, SMILES-processed databases of small natural compounds like DrDukes, IMPPAT, COCONUT and Enamine will be used to identify the possible lead / scaffold molecules. Promising selective scaffolds will be functionalized, derivatized and extended to improve selectivity and ADME profile as well as minimize off-target effects and toxicity. Cheminformatics for dual action will be explored (e.g, active site, allostery).
Ashok intends to achieve the following milestones:
- Rational target identification, dataset preprocessing and Chemical Space Docking against targets
- QSAR / QSPR Optimization of lead molecules with scope for dual action against multiple targets
- Analysis for further improvement of protocol, selectivity and toxicity, profiles.