Can 23andMe reinvent itself as a drug discovery company?

Seventy-one years ago, Francis Crick burst into the Eagle pub in Cambridge to announce to startled lunchtime drinkers that he and his fellow researcher James Watson had discovered “the secret of life”. Their deciphering of the structure of DNA launched a new era of scientific research and won them the Nobel Prize for medicine.

Nowadays, it has become ridiculously cheap and easy to discover your own DNA. Spit into a tube, post it to a lab and a genetics testing company will email you the results for as little as $99. The collapse in the cost of this technology is staggering, outstripping the exponential advances in computing power, known as Moore’s Law. At the beginning of the century it cost more than $95mn to sequence one human-sized genome. 

This data revolution has enabled millions of people to trace their ancestry and discover lost relatives. It has also helped doctors diagnose genetic diseases and created an astonishing research resource for scientists. While researchers have been salivating about this informational treasure trove, investors seem less impressed. But intriguingly, several billionaires have laid contrarian bets that 23andMe, one of the biggest testing companies, can reinvent itself as a valuable drug discovery business. Can it succeed?

Since being founded in 2006, 23andMe has experienced a wild roller-coaster ride, even by the dippy standards of West Coast start-ups. Once viewed as one of the world’s coolest companies, 23andMe had little trouble raising $1.4bn to fund the expansion of its direct-to-consumer DNA-testing business. It was closely connected with Silicon Valley royalty: Anne Wojcicki, the company’s co-founder and chief executive, received early backing from Sergey Brin, the co-founder of Google whom she went on to marry (and later divorce). 

Amid great acclaim, the company listed on Nasdaq in 2021 via a special purpose acquisition company backed by Sir Richard Branson and its valuation topped $6bn. But its shares have since plunged more than 95 per cent and it is in danger of being delisted. 

The trouble with DNA testing is that you only need to do it once and the early adopters have now adopted. 23andMe’s attempts to build a healthcare subscription business have not scaled as fast as hoped. The company also suffered a damaging data breach last year. Hence the need for a corporate reinvention.

With foresight, 23andMe created a therapeutics research division nine years ago to exploit its extraordinary cache of genetic data. Its database now includes 15mn users of whom 80 per cent have consented to share their anonymised data for research. The company has been working with the pharmaceutical giant GSK, focusing on immuno-oncology drugs. “We are like a telescope that can see 10 times further into the biological universe,” Jennifer Low, head of therapeutics development at 23andMe, tells me.

Low says that what distinguishes the company’s approach is the scale of its database and its interactions with customers, which allow additional health information to be gathered. Last month, 23andMe announced that the first participant had been dosed in a Phase 1 clinical trial for an antibody that counters advanced solid tumours. The company has genetically validated two other drug targets.

The field of computational biology is evolving fast and several AI companies are exploring the possibilities of accelerating drug development in this data-rich field. Google DeepMind, the London-based AI company, even spun out a dedicated company called Isomorphic Labs in 2021 for this purpose.

Sir Demis Hassabis, co-founder of Google DeepMind, told me last week that Isomorphic’s research was going “amazingly well” and it was building prediction models to assess the properties needed to design a good drug, based mostly on insights gained from the AlphaFold protein structure database. But he said there was really big potential for researchers to apply AI to large-scale genomics data, too. “We have a lot of data in that world. The question is: how to make sense of it all. This is where AI comes in,” he said.

The challenge for 23andMe is to combine a fast-paced healthcare company mentality with a patient focus on drug discovery, meshing together two radically different business cultures, skill sets, customer demands and timelines. It can take between seven to 10 years after a drug enters clinical trials for it to be approved, assuming it works. Maybe the billionaire investors are right to detect value in 23andMe’s data. But it might take a different corporate structure — or an acquirer — to realise it.

john.thornhill@ft.com