The work in Richard Bryce's group is aimed at developing computational models to improve the accuracy with which molecular processes are described and understood in the condensed phase and then to integrate and employ these methods in the drug design and development process, involving collaboration with medicinal chemists, biochemists and biologists.
His research interests involve the development and application of accurate and efficient computational methods for structure-based drug design. A major focus of his work is the theoretical investigation of carbohydrate structure, energetics, dynamics and reactivity. This requires accurate treatment of aqueous solvent, which as led to the development and application of solvent models to carbohydrate analogue systems. The group employs a variety of computational techniques, including molecular simulation and hybrid quantum mechanical (QM)/molecular mechanical (MM) models.
1. Simulation of biomolecular dynamics and thermodynamics
In addition to hybrid QM/MM models, we also have explored more empirical approaches to modelling biomolecular recognition. For example, we have used a combined molecular mechanics and continuum solvent approach to calculation of thermodynamics, leading to new insights into the recognition of oligosaccharides by the protein, concanavalin A. We have used a similar approach to predict fluorophore-DNA interactions pertinent to molecular diagnostics. This method has been extended to evaluate the free energies of binding of closely-related ligands, with a view to application in rational drug design.
2. Computer-aided drug design
We seek to apply our knowledge of modelling ligand-receptor interactions to adopt a computationally-led design of small molecule modulators. We currently use a range of computational techniques, for example mapping, de novo methods and virtual screening (using molecular docking). In this way, we pursue a number of multidisciplinary design projects, collaborating with medicinal chemists, biochemists and biologists: one example is our structure-based virtual screening of compound libraries which led to the identification of a novel non-nucleobase-derived inhibitor of human thymidine phosphorylase, an anti-cancer target, as a candidate for lead optimisation.
3. Modelling carbohydrates
Carbohydrates, the most naturally abundant biomolecule, function in a number of roles, from maintaining structural integrity to providing energy-storage. However, it is the centrality of carbohydrates in biomolecular recognition, mediating events such as cell-cell interaction and the immune response that offers major opportunities for therapeutic intervention. These polar flexible molecules expose many of the shortcomings of current computational modeling methods.We use a range of methods to study the energetics and dynamics of carbohydrates in the condensed phase. In particular, we have explored the use of combined QM/MM potentials and molecular dynamics. To enable the calculation of free energy surfaces, in collaboration we have derived a focused semi-empirical QM Hamiltonian to specifically improve modelling of carbohydrates.