We aim to understand the mechanistic details of biological processes of therapeutic relevance, using molecular dynamics simulations. To this end, we develop and apply path sampling and enhanced sampling techniques based on statistical mechanics and machine learning to predict the free energy landscape and kinetics of biomolecular rare events, the timescales of which go beyond the capability of conventional simulation techniques. Our areas of interest include drug-target interaction, antigen-antibody recognition, and conformational transitions in proteins and nucleic acids.
Methodological Developments
- Development of Path Sampling and Enhanced Sampling Algorithms
- Machine Learning for Collective Variable Discovery
Applications
- Kinetics of Drug-Target Binding
- Recognition of RNA Targets by Small Molecules
- Allosterism in Antigen-antibody Recognition
- Protein Folding
