Only with continual feedback between experiment and prediction can machine learning models be refined and deliver against complex therapeutic profiles.
To deliver experimental results in a robust manner, we have invested in over 10,000 sq ft (~100 sq m) of cutting edge laboratory that house highly sensitive Surface Plasmon Resonance (SPR), x-ray crystallography and Transducerome techniques. Our laboratories are co-located with AI Drug Design teams to promote a strong interplay between these core disciplines.
Our advanced biophysical and x-ray techniques provide the ideal experimental underpinnings for AI-Design
Transducerome and biophysical experiments allow us to optimise designs for this key target family
High content data allows us to develop the broadest range of machine learning models and AI Design opportunities
Novel targets need high-quality seed data generated quickly for our algorithms whilst established targets may require experiment to uncover new binding sites and additional chemotypes.
Fragment-based screening is ideal for Project Seeding and can be implemented by two complementary high-precision experiments.
Faster, richer and more sensitive than traditional High throughput screening techniques, Surface Plasmon Resonance systems screen Exscientia fragment libraries to seed projects and provide a more direct route to drug-like molecules in early design cycles.
Where available, structure-based fragment screening will directly reveal which fragments occupy which pockets. Successful combination of 3D information with SPR binding kinetics provides an ideal data package to progress early stage projects
Initiation & Progression
Once established, the same SPR and X-ray systems can provide rapid feedback to our Centaur Chemist® systems throughout subsequent Design-Make-Test optimise cycles
G-protein Coupled Receptors are one of the most successful families for small molecule drug discovery. Exscientia has already demonstrated the success of automated design on GPCRs and now has one of the world’s most advanced experimental platforms alongside.
Our transducerome assay technologies investigate detailed G-protein coupling preferences with industry-beating assay sensitivity. This allows us to uncover precise mechanisms of disease biology and develop candidate molecules that decouple therapeutic effects from unwanted side-effects.
We are uniquely able to fragment-screen native-state GPCR targets with biophysical techniques.
By retaining their full range of native conformations agonist, antagonist and inverse agonist forms can be explored equally.
For under-studied GPCRs we can use these technologies to uncover new pharmacology opportunities and develop first-in-class molecules. For well-studied target we can now tease apart complex G-protein responses in order to design improved medicines by minimising unwanted G-protein action.
Some discovery projects are best served by cell based experiments rather than a target-driven approach. Exscientia’s AI systems are entirely flexible and can support any data types including those from high content experiments.
High Content Data
Whether it be cell morphology or higher-level behavioural data, the conversion of high dimensionality readouts to Centaur Chemist® input is all that is required to implement phenotypic drug discovery.
With phenotypic experiments in place, our Centaur Chemist® systems focus on optimising against phenotypic endpoints rather than conventional target-based profiles.
Through companion experiments, target-based mechanisms may subsequently become validated, In such cases Centaur Chemist® continues to drive projects forward using the combined power of target and phenotypic data.