Development of data-driven methods for de novo design of novel enzymes.

De novo protein design is quickly becoming a viable strategy for creating novel protein structures, especially given recent advances in structure prediction using deep-learning based methods such as AlphaFold [1]. However, it remains highly challenging to add complex functionality to these molecules [2]. While there has been some success in designing novel enzymes, their enzymatic activity falls short of most natural systems, making it necessary to perform directed evolution to improve activity [3]. In nature, cofactors are often incorporated into proteins to add a vast array of functionality, and so they offer an attractive route to create highly-active enzymes.

This project aims to develop novel methods for designing proteins that incorporate cofactors. These tools will build on existing technology developed in the Wood lab, that utilises structural analysis and machine learning to produce and evaluate novel-protein sequences [4,5], supported by the Jarvis Lab who are experts in the design of artificial enzymes [6]. Throughout the PhD, the student will develop data-driven methods to design and understand sequences that bind cofactors that have applications in photochemistry. If realised, these proteins will have a transformational impact in the field biocatalysis.

This project is primarily computational, although there may be some opportunity to perform experiments in the lab. While advantageous, experience in programming/machine learning and/or statistics is not required. We are adept at training people in these areas and the student will be well supported. All that is required is enthusiasm and determination to learn these skills.

The student that takes on this project will form part of a larger cross-institutional team with the Universities of Manchester and Bristol, and there will be opportunities to spend some time in these institutions during course of the PhD.