Скачать или смотреть Application of an automated machine learning pipeline to accelerate drug discovery in Africa

Abstract: Drug discovery and development is a time-consuming and extremely expensive process as a molecule must satisfy numerous efficacy and safety criteria before it can be considered for use as a medicine. As the vast majority of compounds will not meet these criteria, there is a need to identify the most promising compounds as early as possible and to deprioritize the rest. As machine learning algorithms are adept at identify patterns in data, these methods can produce models that learn from past data to then prioritize prospective compounds for further research and testing. In this work, we present an automated modelling pipeline for chemical datasets, ZairaChem, that produces state-of-the-art property-prediction models to support the drug discovery research at the Holistic Drug Discovery and Development (H3D) Centre at the University of Cape Town.

Bio: Dr Jason Hlozek is a computational chemist at the University of Cape Town who is passionate about leveraging computational methods to address various scientific questions, particularly in the context of global health. During his postgraduate studies, he focused on applying molecular dynamics simulations for the study and optimisation of carbohydrate polymer anti-bacterial vaccines. Dr Hlozek is currently a researcher at the Holistic Drug Discovery and Development (H3D) Centre where he develops and implements artificial intelligence (AI)-based tools in order to accelerate infectious disease research and drug development on the African continent. This involves working closely with lab-based scientists to understand their individual data science needs and to then build solutions that facilitate scientific advances in low-resource settings.