The project: This is an industrially funded 4-year PhD project, in collaboration with Boehringer Ingelheim (BI), a leading pharmaceutical company. It includes a one-year placement in Oxford, during which in-depth training will be provided in fundamental theory, software development, and computational chemistry simulations. This training will be delivered by academics from the Universities of Southampton, Bristol and Oxford via the Centre for Doctoral Training in Theory and Modelling in the Chemical Sciences. Successful completion of year one will lead to the award of an Oxford MSc, and progression to a 3-year PhD research project, which will be based in Southampton, and will include close collaboration with BI and visits to their research labs in Germany.

This project will offer unique exposure to both academic and industrial research environments and is ideally suited to applicants who wish to make an impact in problems of real industrial relevance.

Developing free energy methods with large-scale quantum ensembles for applications in drug optimisation

Computational simulation plays an important role in the early stages of the development of new drugs by identifying molecules (potential drugs) which can bind to biomolecular targets (e.g. sites in a protein) with high affinity and selectivity. This process involves several stages. Crude but computationally inexpensive methods (e.g. docking) are initially used to scan huge libraries of molecules and reduce the number of candidates. Eventually a small number of the “best” leads can be refined with computationally more demanding but also more accurate approaches which compute relative free energies of binding. The most rigorous methods for calculation of relative free energies of binding are based on statistical mechanics and are an active field of research. This is a particularly challenging area as these methods require generating and sampling a large number of biomolecular configurations (in order to capture the entropic contributions). Each of these configurations needs to be a correct (in terms of statistical probabilities) representative of the thermodynamic state and we need to have a very accurate evaluation of its energy. Finally, the accurate description of the solvent and its interactions with the solutes is also crucial as the ligand de-solvation free energy is required in the free energy thermodynamic cycles. Commonly used rigorous approaches for free energy calculations depend on the use of empirical force fields for the generation of configurations and their energies. While force fields are computationally more tractable than ab initio quantum chemistry calculations they have limited accuracy and transferability, as they cannot capture explicitly important energy contributions such as the electronic polarisation and charge transfer that occur in a biomolecular association event. The limitations of force fields are more severe when the molecules considered are different from the scope of parameterisation of the force field, which is often the case when searching for new drugs.

The goal of this project is to overcome the force field limitations in biomolecular free energy calculations by employing large-scale ab initio calculations. To achieve this goal we will develop hybrid free energy methods which start with force fields to compute free energy differences between different ligands but then compute the free energy of mutation from the classical to the quantum description (free energies are thermodynamic state functions so such a transition is well-defined). This work will build on our previous experience in this area [1,2] and will use the ONETEP linear-scaling DFT program [3], which we develop in our group. Particular challenges in this project will be the development of free energy methods that have high configurational overlap between the classical and the quantum description and produce correct ensembles of structures in both descriptions. Also, the development of quantum methods that provide the most accurate description of biomolecular interactions (such as new generations of DFT approaches) while reducing the computational demands, and the calibration of explicit and implicit models for the solvent. The project will involve development of new theory and code within ONETEP and in stand-alone free energy methods programs. Energy Decomposition Analysis (EDA) [4] will be used on the DFT calculations to dissect the protein-drug interaction in terms of energy components (such as electrostatic, exchange, polarisation, charge transfer) and into particular chemical functional groups providing thus information for subsequent chemical modifications to improve the activity.

The new methods will be validated in actual protein-ligand targets of relevance to the pharmaceutical industry. The project is supported by Boehringer Ingelheim (BI) and will be co-supervised by researchers from BI and by Professor Chris-Kriton Skylaris, and will involve periods of work with the BI computational chemists in Germany.

[1] S. J. Fox, J. Dziedzic, T. Fox, C. S. Tautermann, and C.-K. Skylaris, Proteins 82(2014) 3335.

[2] C. Sampson, T. Fox, C. S. Tautermann, C. J. Woods, and C.-K. Skylaris, J. Phys. Chem. B119(2015) 7030-7040.

[3] C.-K. Skylaris, P. D. Haynes, A. A. Mostofi and M. C. Payne, J. Chem. Phys.122(2005) 084119.

[4] M. J. S. Phipps, T. Fox, C. S. Tautermann and C.-K. Skylaris, Chem. Soc. Rev.44(2015) 3177-3211

Candidate requirements:

UK or EU citizen fully funded
Successful applicants to TMCS typically hold a first class honours degree (or equivalent) in Chemistry or a closely related discipline.

Funding: Full funding of fees and stipend for 4 years. Applications are accepted from self-funded international applicants, but funding restricted to UK/EU applicants only.

Contacts:

Project queries: Professor Chris-Kriton Skylaris, [email protected]

TMCS queries: [email protected]