Advances with deep learning, the growth of databases of molecules for virtual screening and improvements in computational power have supported the emergence of a new field of quantitative structure–activity relationship (QSAR) modelling applications that Tropsha et al. term ‘deep QSAR’. This article discusses key advances in the field, including deep generative and reinforcement learning approaches in molecular design, deep learning models for synthetic planning, and the use of deep QSAR models in structure-based virtual screening.
- Alexander Tropsha
- Olexandr Isayev
- Artem Cherkasov
