I studied Chemistry with Materials chemistry at the University of St Andrews, where I was fortunate enough to be exposed to both computational chemistry and chemoinformatics. Throughout my MChem, I became interested in the idea of designing Novel molecules with specific function and form. Following my MChem, I decided to study for a PhD, this project at the University of Sheffield interested me as it combined my interest in chemoinformatics, with the opportunity to utilise large biological datasets to design compounds with specific property profiles.
The goal of the project is to develop new de novo design tools that are effective for the design of novel compounds that can interfere with multiple targets within a biological network, that is the systematic and rational design of compounds with specific biological and chemical profiles. De novo design is divided into three critical components; construction, scoring and search. Each part must be carefully balanced to allow for efficient navigation through the vastness of chemical space to areas of interest. Current research has focused on the design of compounds which are amenable to chemical synthesis. These approaches utilise the in-silico application of common medicinal chemistry reactions to available sets of reagents. However, the sheer size of chemical space that is accessible even using this constrained approach to de novo design is such that it is impossible to enumerate all potential compounds. Hence it is, therefore, necessary to design efficient data-driven search strategies to navigate paths through this space.
Importantly, the project will develop novel search methods by the utilisation of artificial intelligence, coupled with in-silico medicinal chemistry transformations. These search methods are then driven via the use of scoring functions derived from large open-access biological datasets and combined using many-objective optimisation approaches. In doing so, the goal is to design effective small molecules which demonstrate well-defined chemical properties but also specific biological effect across multiple targets within a disease network.