The MRC Laboratory of Molecular Biology in Cambridge, UK is seeking a doctoral student to join the lab of Joe Greener. The fully-funded position will aim to develop a universal biomolecular force field using machine learning approaches. The student will become an early member of a friendly, multi-disciplinary team.
Molecular dynamics has been successful in obtaining biological insights by simulating the motion of biomolecules. Increases in compute power mean that simulations will soon reach the time scales on which much of biology happens. However the force fields used are developed manually for specific systems and often perform poorly for disordered proteins and protein aggregation.
Recent advances in machine learning and an increase in available data have opened the path to learning universal molecular mechanics force fields that are accurate for proteins, nucleic acids, lipids, small molecules and more. This project aims to develop such force fields using differentiable molecular simulation, an emerging technique where whole force fields can be trained to fit experimental data using automatic differentiation, the technique used to train neural networks. Other recent advances such as continuous atom typing using graph neural networks and exploration of different functional forms for non-bonded interactions will also be incorporated.
The project will involve significant coding using the Julia package Molly.jl, on both CPU and GPU, as well as addressing computational and algorithmic questions surrounding differentiable simulations. An additional aim is to improve the performance of free energy perturbation simulations as used in drug discovery. This is a purely computational project but will involve collaboration with experimentalists at the LMB to run simulations on their data. It is expected that the student’s interests will contribute to the direction of the project.
Entry requirements
Open to any UK or international candidates. Applicants should have, or expect to achieve, at least a 2.1 honours degree or a master's (or international equivalent) in chemistry, physics, computer science, maths or biology.
How to apply
Applicants are required to send a CV and cover letter to Joe Greener along with the names and contact details of two referees.
References
Greener JG. Differentiable simulation to develop molecular dynamics force fields for disordered proteins, Chemical Science (accepted)
Greener JG, Jones DT. Differentiable molecular simulation can learn all the parameters in a coarse-grained force field for proteins, PLoS ONE 16(9), e0256990, 2021
Robustelli P, Piana S, Shaw DE. Developing a molecular dynamics force field for both folded and disordered protein states, Proc Natl Acad Sci USA 115:E4758-E4766, 2018
Wang Y, Fass J, Kaminow B, Herr, JE, et al. End-to-End differentiable construction of molecular mechanics force fields, Chemical Science 13(41):11953-12246, 2022