Abstract:

Machine learning strategies in drug discovery are becoming increasingly popular and can be used in various areas. In the Schrödinger Suite DeepAutoQSAR serves as the main tool for training machine learning models to predict activity, ADMET, and other compound properties. In order to leverage both the proven accuracy and wide applicability domain of physics-based computational models, such as QM and FEP, together with the speed and scale of machine learning, we have combined our physics-based modeling technologies with an active learning framework. This framework can effectively speed up virtual screening methods such as in Active Learning -Glide, Active Learning-FEP, and Active Learning-ABFEP, or to improve the accuracy and applicability domain of models such as pKa prediction in Epik and machine learned force fields such as QRNN. We will also discuss how to utilize machine learning protein structure prediction methods to enable new targets for structure-based drug design.

Speaker:

Dr. Marton Vass, Principal Scientist II, Schrödinger