Department of Pure Mathematics and Mathematical Statistics
Applications are invited for a four-year, AstraZeneca-funded PhD studentship in the Department of Pure Mathematics and Mathematical Statistics, commencing in October 2027.
The successful candidate will be jointly supervised by Prof Sergio Bacallado (University of Cambridge), Dr Dino Oglic (AstraZeneca), and Dr Jon Paul Janet (AstraZeneca). The studentship offers an opportunity to conduct cutting-edge research at the interface of statistical machine learning, molecular simulation and drug discovery.
Project description
The project, entitled Physics-Guided Representation Learning from Free Energy Perturbation Data for Affinity Prediction and Candidate Down-Selection, aims to develop novel machine-learning methods for predicting molecular binding affinities and supporting decision-making in drug discovery.
Free energy perturbation (FEP) calculations provide valuable information about how strongly candidate drug molecules bind to their biological targets, but these calculations are computationally expensive. This project will investigate how machine-learning models can extract richer information from FEP data and experimental measurements, combining physical principles with modern representation-learning techniques.
The research will involve developing and evaluating new statistical and machine-learning methods, with access to both public benchmark datasets and AstraZeneca's industrial drug-discovery expertise and data.