Megacorp GSK inks AI drug development deal with Brit firm
Exscientia claims approach IDs candidates quicker
GlaxoSmithKline has announced a research deal with British company Exscientia to use artificial intelligence to identify drug targets.
The deal will see GSK fund Exscientia’s research into AI-driven drug discovery, paying out up to £33m if it hits all its targets.
GSK has tasked the firm with identifying molecules that have the potential to treat up to 10 diseases in different areas - and will pay out based on how many of the projects go forward.
The big pharma company is one of many turning to AI to help speed up drug development processes, with this work focusing on the stage of creating a host of drug candidates.
Although some stages of drug discovery have benefited from new technologies, those working in the field want to see more work focused on the early stages of drug development.
This involves identifying molecules that could interact with disease targets; at a simplistic level, this might involve creating a drug that binds to a bacterium in such a way that it can’t produce a protein it needs to survive.
Again, simplistically, the better this binding is, the better the drug works. But the crucial part about drug safety is not just that it interacts efficiently with the disease target - it must be highly specific for that interaction, or risk adverse affects of binding in places it shouldn’t.
The drug industry spends a lot on these early stages, but many targets will fail at a later stage.
Jackie Hunter, CEO of the biological sciences arm of firm BenevolentAI, has said that the pharmaceutical industry loses 50 per cent of compounds in Phase II and Phase III trials - tests on between 100-300, and 300-3,000 patients, respectively - for lack of efficacy.
“That isn’t sustainable; it tells us we’re picking the wrong targets.
"A further quarter of failures in Phase II or III are for strategic or commercial reasons. That also tells us industry is not always making the right decisions about what compounds to prioritize,” she told EY for the consultancy’s recent report on biotechnology.
The industry is trying to cut down on such losses by using AI-driven algorithms trained using academic literature and existing studies.
They will look for patterns in chemical structures and can be used to produce drugs that are specific for the target in question.
It allows researchers to cycle through potential molecules more quickly, and the use of big data allows quicker assessment of candidates; information that is then fed into the AI system and used to generate more - and improved - candidates.
Algorithms can also be used to assess the affect of a molecule on a cell, tissue or organism - projects like this generate masses of data that traditional methods wouldn’t be able to process. This information can then feed into drug discovery.
Chief exec Andrew Hopkins said that Exscientia’s approach could offer up potential drugs in “roughly one-quarter of the time, and at one-quarter of the cost of traditional approaches”.
The firm said that its AI systems are developed to balance the strength - potency - of a drug, how selective it is and its pharmacokinetics - basically how quickly it is absorbed, processed and excreted by the body.
“By applying a rapid design-make-test cycle, the Exscientia AI system actively learns from the preceding experimental results and rapidly evolves compounds towards the desired candidate criteria,” the company said in a statement.
The firm will collaborate with GSK to discover novel and selective small molecules that interact with the disease targets set out by GSK.
A spokesperson for GSK told The Register that the company hadn't disclosed its overall investment in AI, but that it "sees it as a channel to keep on top of" and planned to work with others to advance it.
GSK's other work in the area includes the ATOM initiative - Accelerating Therapeutics for Opportunities in Medicine - in collaboration with the US National Cancer Institute and the Lawrence Livermore National Laboratories.
That project is looking at how to use high-performance computing to replace some of the empirical work used in drug discovery.
Meanwhile, in May, Exscientia announced a partnership with another massive pharmaceutical company, Sanofi, for work on metabolic disease, like diabetes - worth a potential €250m.
Similar to the GSK partnership, that work will develop and validate drug targets, but will focus on creating molecules that work with two distinct drug targets.
This is because drugs used for more complex diseases need to hit a number of targets at the same time to have a sustainable affect on the disease.
That deal brings in more money as any licensed products reaching the market will qualify for recurrent sales milestones. ®
Sponsored: Beyond the Data Frontier