Conformational enhanced sampling the study and development of accelerated sampling algorithms for classical biomolecular dynamics

Abstract

Significant progress in computer hardware and innovations within enhanced sampling methods has led to the ability to study larger and more complex biomolecular systems and problems that rely upon simulating a larger number of slow degrees of freedom that may also be orthogonal to one another. Despite this progress, there remains problems that elude modern enhanced sampling methods. These include the simulation of protein-protein interactions at druggable interfaces, cryptic binding pocket openings and secondary structure formations associated with binding events or mechanisms of regulation within biomolecular pathways of cellular function. This thesis is an exploration of the technical limitations of the popular enhanced sampling method metadynamics (MetaD) and how such limitations can be overcome by the application of novel analysis tools and optimisation procedures. This text commences with a theoretical introduction to MetaD and a categorisation of enhanced sampling methods based upon whether the applied bias acts upon the system coordinates or velocities, focusing upon the similarities between MetaD and other enhanced sampling methods that require the definition of a collective variable (CV) space The technical challenges faced when selecting a CV space and MetaD bias potential parameters for single-replica simulations of folding of the 20-residue peptide TrpCage was studied in detail. The study give way to the development of a set of quantitative quality metrics for MetaD bias performance and CV space quality that accounts for the relationship between the positions of conformationally distinct metastable states within phase space and their positions within the candidate CV space. MetaD single-replica simulations aiming to disrupt the His41-Cys145 catalytic dyad of the SARS-CoV 2 main protease (M^Pro) demonstrates a successful use case of MetaD in sampling a novel pathway for SARS-CoV 2 M^Pro inhibition with respect to disrupting the interaction between His41 and Cys145 responsible for protease activity. The yielded disrupted state about the M^Pro active site can be readily stabilised by ligand interactions with the displaced His41 imidazole group.

More information

Identifiers

ORCID for Khaled Abdel Maksoud: orcid.org/0000-0002-6029-4966

ORCID for Jonathan Essex: orcid.org/0000-0003-2639-2746

Catalogue record

Export record