Join our community of smart investors

UK pharma turns to AI for drug discovery

Super smart computers could play a crucial role in a new era for big pharma
August 30, 2017

In March 2016 the prestigious scientific journal Nature published a paper describing how an advanced artificial intelligence (AI) system had beaten the world’s best Go player – a game said to be exponentially more complex than Chess. The findings sparked excitement: if computers can become smart enough to win at Go, their potential in many other fields could be enormous, including in drug discovery. 

AstraZeneca (AZN) has entered into a partnership with Boston group Berg, aiming to identify potential candidates for drugs to treat Parkinson’s disease and other neurological conditions. Berg will examine chemical fragments provided by AstraZeneca and used tissue samples from both healthy and diseased patients in an attempt to identify a molecule that has potential therapeutic effects in a range of neurological conditions.

Financial terms of the partnership have not been disclosed, but Astra will have the right to secure an exclusive licence for any of the candidates identified. Iain Chessell, Astra’s head of neuroscience, told the Financial Times that the research collaboration would allow it to “approach drug discovery in an innovative new way”. 

This follows news in July that GlaxoSmithKline (GSK) had teamed up with Dundee-based group Exscientia to identify small molecules that have the potential to treat multiple illnesses.

Mark Swindells, chief operating officer at Exscientia, said that compared to traditional methods of identifying new drug compounds, AI “is more efficient and productive and the quality of the compounds should be higher”. The company uses a series of algorithms to screen molecules for characteristics that could make them good candidates for new drugs. In 2015 these algorithms helped Japanese pharma group Sumitomo identify a molecule with the potential to treat psychiatric diseases. The drug candidate was discovered after just over a year of testing, compared to the industry standard of two to five years. The process examined a quarter of the molecules that would normally be tested in early-stage drug discovery, saving Sumitomo both time and money.  

These British giants are not alone, Merck (US: MRK), Sanofi (FR: SAN) and Johnson and Johnson (US: JNJ) are a handful of global pharma groups which have turned to AI to help with the drug development process.

For GSK, partnering with Exscientia could provide a crucial plug to its thinning new drugs pipeline. AstraZeneca too could benefit from reducing the costs of early stage research and development (R&D) – the single most expensive and time consuming part of creating a new drug. Its R&D costs recently rose to 27 per cent of total revenue.