Computer simulation of mini-fragment binding in drug discovery

Fragment-based drug discovery (FBDD) is proving increasingly important in the pharmaceutical industry. A library of small molecules of low molecular weight (fragments) are screened for binding against a protein target. While there are a number of experimental routes (e.g. SPR, NMR) to assess binding, fragment soaking into protein crystals, with binding being identified through X-ray crystallography, is arguably the most effective. Having identified a low molecular weight hit, computational approaches that build on the protein-ligand crystal structure are used to suggest chemical modifications to the fragment to improve binding affinity, specificity to the specific protein target, and other physico-chemical properties such as solubility.

The aim of this project is to use the experimental binding data of small molecular fragments to test and guide the development of computer simulation approaches to model fragment binding. The development of appropriately validated in silico methods will not only guide experiment, but can also be used in cases unsuited to the experimental approach, where high concentration of ligand is impossible for solubility or aggregation reasons, or where the fragments induce protein denaturation.

Conventional molecular dynamics can be hampered by inadequate sampling due to kinetic barriers i.e. the timescales required for the fragments to bind the receptor by simple diffusion are too slow to be modelled in a micros-length molecular dynamics simulation. To address this sampling limitation, enhanced sampling simulation approaches will be employed.