Machine learning in drug discovery is not new, as classical machine learning has been used for decades, however in recent years there has been great advances in artificial neural network applications and a new wave of methods for solving drug discovery problems. In our recent activity we demonstrated state of the art in several predictors relevant for drug discovery, in particular binding affinity and pose predictions. While these methods might be state-of-the-art, they are not always based on deep learning methods, some more in depth analysis of these methods does provide some useful indications.
Copyright. All Rights Reserved.