Structure-based virtual screening for molecules with TRPM8 activity by allying cryo-EM maps and artificial intelligence
This repository contains supplementary files accompanying the manuscript titled 'Structure-based virtual screening for molecules with TRPM8 activity by allying cryo-EM maps and artificial intelligence' by H. Okesade, N. James, T. Rahman, and P. J. Ballester.
Cryo-electron microcsopy (cryo-EM) has become an increasingly powerful tool for the structural determination of membrane-bound proteins. Structure-based virtual screening for transient receptor potential melastatin 8 (TRPM8) – a target of growing therapeutic interest due to its implication in various disease and pain states, including dentin hypersensitivity and cancer – has significantly progressed as a direct result of this. This study explores the potential for improved docking performance using cryo-EM maps as an integral part of deposited TRPM8 structures, using the recently developed ChemEM software by Sweeney and colleagues. We compare ChemEM with the classical docking tool Smina, and subsequently employ machine-learning scoring functions for ranking. Our preliminary findings indicate that while cryo-EM map incorporation shows little significance, the ChemEM scoring function demonstrates superior performance compared to Smina. Future software iterations and/or an automated workflow could enable more comprehensive virtual screening, allowing for the exploration of ultra-large chemical libraries and potentially advancing TRPM8-targeted drug discovery.
- Benchmarking dataset(s)
- Hit lists
- Protein structures
Additional files (cryo-EM density maps used in this study) are available separately on GitLab via. https://gitlab.com/hannahokesade/msc_digi_chem/-/commits/density-maps.