Binding free energy calculations are commonly used in the pharmaceutical industry to discover molecules which could be usefully developed into drug candidates. Unfortunately, current methods are not always successful. When inaccurate free energies are calculated, the errors may be attributed to either a force field which does not sufficiently represent the molecular interactions in the system, or inadequate sampling, which my project aims to address.

Conventional molecular dynamics is great for sampling a specific binding geometry, but large kinetic barriers separating other potential configurations result in alternative binding modes being rarely observed. A variety of enhanced sampling algorithms have been developed in order to solve this issue, either by reducing the energy barrier or adding energy to the simulation so that the barriers can be crossed naturally, but all of these methods contain various disadvantages that render them inefficient and requiring considerable optimisation.

Through a collaboration with AstraZeneca and GSK, a novel enhanced sampling technique has been developed to address the sampling problem - Fully Adaptive Simulated Tempering (FAST). Current preliminary data suggests that the approach has significant promise, and my research aims to further validate and optimise this technique in order to discover its true potential and the domain of applicability.