Exscientia wants to bring its AI/ML approach into the world of biologics. It's starting with antibodies
Exscientia is looking at breaking out of small molecule R&D.
The UK biotech announced Thursday that it is expanding its R&D efforts into biologics, more specifically antibodies, looking to implement precision engineering and utilize in silico screening. And in order to put the expansion into place, Exscientia also announced that it will be adding new building space to its headquarters to house the new laboratory.
The project, which will be around 8,000 square feet, is expected to be finished in the first half of 2023. Exscientia has been flush with cash over the last few years, thanks to a nine-figure IPO last year and a billion-dollar deal with Sanofi inked earlier in 2022, but did not disclose how much it’s sinking into the new efforts.
CEO Andrew Hopkins tells Endpoints News that the main goal of this expansion was to take the “natural next step” as a company.
“And that’s why we’re moving and expanding from small molecules also going to include biologics…It’s about how we could potentially double the target universe we could go after with our technology, when we think about all the possible antibody and biologic targets we could go after, as well as small molecule targets,” Hopkins said.
The AI chief added that the heart of the expansion is around a new engine that Exscientia’s chief scientist of biologics AI, Charlotte Deane, and her team have been working on developing. Deane, who joined Exscientia earlier this year, is the former head of Oxford’s statistics department.
Hopkins said that the goal of this new engine was to not have just an AI platform, but an automated laboratory system that can have continual back-and-forth between generating antibody designs and then physically testing them in the lab — “biologics by design, not discovery,” Hopkins quipped.
Exscientia has already started looking at antibodies from Sanofi’s pipeline, Hopkins said, essentially applying a precision medicine platform to identify biomarkers and further identify the patient populations for these antibodies to move forward into clinical testing.
The first step of the process is thinking about which targets to go after, Hopkins said, including identifying the epitope, the part of the protein sequence the antibody needs to target. Then comes designing specific antibodies from scratch, de novo, that could try to go after that specific epitope in a virtual library.
Finally there’s the production of those antibodies that can be tested in the real world. Researchers examine how the antibodies bind, as well as the properties of said antibodies. That ties in with new, in-house tech being developed that would allow for higher throughput and, according to Hopkins, for more data generation that can be fed into Exscientia’s machine learning models, forming a positive feedback loop.
As for what’s next, Hopkins says that the outfit has already started working on the engineering — designing and commissioning the technology required to get the automated lab up and running. The first projects, however, are underway, with Exscientia working with various undisclosed partners.
“We do hope this will be a significant part of accenting this pipeline going forward in the future,” Hopkins concluded.