Our research operates at the intersection of computational molecular biophysics, theoretical physical chemistry, and data science. We design and apply advanced computational approaches — grounded in statistical mechanics, molecular thermodynamics, and machine learning — to problems where molecular dynamics controls biology and function. A defining feature of our group is the co-development of novel in-house algorithms alongside their application to real biological systems.
On the methods side, we develop and release in-house algorithms — PathGennie (direction-guided adaptive sampling for rapid rare-event pathway generation), WeTICA and CoWERA (directed weighted-ensemble enhanced sampling for rare-event kinetics), IceCoder (variational-autoencoder identification of crystal/ice phases), and PPIscout (mixed-solvent mapping of protein–protein interaction hotspots). On the applications side, our work spans dynamic protein allostery and cellular signalling; druggable cryptic-pocket discovery and protein–protein-interaction inhibition (PCSK9, PLK1, Rho GTPases); prion misfolding and aggregation; enzyme promiscuity and selective inhibition (laccase, acetylcholinesterase); the structure of water around solutes and interfaces; and the molecular thermodynamics of soft matter — gas hydrates, designer ionic liquids, self-assembly and nucleation.
We decode how proteins relay signals between distant sites — the basis of cellular regulation and a powerful, selective handle for drug design. Combining all-atom and mixed-solvent molecular dynamics with QM/MM, we established the electrostatics-driven, dynamic nature of allostery in PDZ domains, mapped nucleotide- and phosphorylation-controlled switching in Rho-family GTPases, and uncovered druggable cryptic pockets that let us inhibit otherwise ‘undruggable’ protein–protein interactions such as PCSK9–LDLR and PLK1.
Selected Publications
A defining feature of our group is the co-development of new in-house algorithms alongside their application. We build enhanced-sampling and machine-learning methods that break the timescale and dimensionality barriers of molecular simulation: directed weighted-ensemble samplers (WeTICA, CoWERA), rapid rare-event pathway generators (PathGennie), variational-autoencoder classifiers of physical states (IceCoder), and mixed-solvent interaction-hotspot mappers (PPIscout).
Selected Publications
We connect molecular structure and dynamics to biological function — and dysfunction. Our simulations resolve how prion proteins misfold and aggregate into disease-associated states, how mini-proteins navigate their folding landscapes, and how enzymes such as laccase achieve broad substrate promiscuity or are selectively inhibited (acetylcholinesterase) — bridging fundamental biophysics with bioremediation and drug discovery.
Selected Publications
Water is an active participant in molecular recognition, not a passive backdrop. We quantify the length-scale–dependent structure and dynamics of water around hydrophobic solutes and interfaces, in nanoconfinement (reverse micelles, membrane-protein channels) and in mixed solvents — and connect these microscopic pictures directly to spectroscopic observables such as 2D-IR and vibrational Stark probes.
Selected Publications
We apply molecular thermodynamics to complex, real-world soft-matter systems: the nucleation and growth of gas (methane) hydrates relevant to energy and flow assurance, designer ionic liquids that stabilise biomolecules, surfactant and polymer self-assembly, nanoparticle nucleation, and stimuli-responsive liquid crystals — frequently in close partnership with experiment and industry.
Selected Publications
We are always looking for motivated PhD students and postdoctoral researchers with backgrounds in chemistry, physics, or computational biology.