Summary
Deep learning is significantly worse at maintaining the shape of the ligand in general than classical approaches (1). For diffusion-based prediction, this can be partially fixed with mguidance.
Figures
Ref (1)
See also
- DL models excel at finding pockets but not docking into them
- DL structure tools introduce chiral errors such as swapped chiral centers, D-amino acids, and cis-amide bonds
1.
Buttenschoen M, Morris GM, Deane CM. PoseBusters: AI-based docking methods fail to generate physically valid poses or generalise to novel sequences. Chemical Science. 2024;15(9):3130–9. Available from: https://doi.org/10.1039/d3sc04185a