AI drug discovery

EXPERIMENT

Timely experimental confirmation of predictions from our AI-driven Centaur system is critical for success. Only with continual feedback between experiment and prediction can Machine Learned models be refined and projects progressed.

Our systems have been built with the flexibility to incorporate a range of experimental data. Here we describe three areas of particular interest; Exscientia’s in-house biophysical fragment screening using Surface Plasmon Resonance, application to 3D structure enabled projects and opportunities for high-content Phenotypic drug discovery.

FRAGMENT BASED LEAD DISCOVERY

For novel targets with little to no pre-existing data we want to generate seed data quickly. Our preferred approach is fragment-based screening using Surface Plasmon Resonance. The approach is faster, richer and more sensitive than traditional High throughput screening techniques and does not require the determination of a 3D structure by x-ray crystallography or NMR spectroscopy.

Where 3D structures can be generated in reasonable time, X-ray driven fragment screening is also conducted. This combines perfectly with biophysical screening, with x-ray providing structural information and SPR the binding kinetics

Phenotypic Drug Discovery

Exscientia’s AI system can support data from any experimental technique and high content experiments are no exception. Whether it be cell morphology or higher-level behavioural data, the reduction of high dimensionality readouts to the signal types used by Exscientia’s existing systems is all that is required to apply Centaur technology to phenotypic drug discovery.