Computational modelling of long acting injectable formulations.

The application of predictive modelling approaches to formulation design and development is not yet well developed, mostly we still rely on trial-and-error experimentation and extrapolations from existing knowledge. One of the reasons for this is we still have little understanding of the nano (atomistic) structure of the complex materials we generate. In contrast, for early-stage drug discovery there is frequently available - and much used - atomistic structural information on both small molecules and drug targets which with a well-developed understanding of the associated physics/chemistry means that predictive modelling is possible, and frequently very valuable. The aim of this project is to develop a molecular modelling-based method that predicts the atomistic structure of complex nanoparticulate formulations, and then link this with a machine learning/AI approach that takes descriptors from this 3D model and uses them to predict experimental properties of the formulation (such as particle size, stability, ease of manufacture, rate of drug release, etc.). The reason for the ML/AI approach is that, in contrast to the situation in drug discovery, it is very hard - maybe currently impossible - to generate models for nanoparticle nanostructure that are large enough and dynamic enough that the emergent properties that are experimentally discernible can be extracted from them directly.