How Four Letters and AI Can Create a Cure

The impetus for François Vigneault’s decades-long love affair with biology was actor Dustin Hoffman. Vigneault watched the movie Outbreak in 1995, in which Hoffman plays an Army epidemiologist searching for a cure for a quickly spreading and deadly virus. “That’s what I want to do,” he thought.

The movie motivated Vigneault, who was serving in the Canadian Navy at the time, to pursue an undergrad degree in biology, eventually leading to a Ph.D. and a fellowship at Harvard. Today, he boasts a track record of success. He sold his first startup, oncology therapeutics company AbVitro, to Juno Therapeutics in 2016 for $125 million.

His current venture, Seattle-based Shape Therapeutics, combines artificial intelligence and RNA to program and develop new medicines. Founded in 2018 by Vigneault, John Suliman, and Prashant Mali, ShapeTX has grown into a 125-employee company with $147 million in Series B funding.

François Vigneault

Courtesy Of Shape Therapeutics

The aspiration was really to think about what would impact patients’ lives the most in the coming decades. The thing that was pretty obvious to us at the time was AI applied to large data sets. We said if we could do that on very large genomic data sets, this is bound to eventually develop drugs that we can’t do right now with classic biology. The other component was RNA.

DNA and RNA are composed of the TCGA. Four letters make your whole genomic code. The idea was looking at RNA more as a software, bits and codes, you could make every possible permutation. If you were to design every variation of these four nucleotides on RNA, you could possibly find a solution to fix mutations and help patients.

At ShapeTX, the more we did this, the more we understood how the system works to a point where we’re now able to use AI to purely design and mRNA (messenger RNA) to fix causal mutations and go after specific diseases.

I’m attracted to understanding how things work. I really liked the intersection of trying to break down problems to their bare components. That’s why I was attracted to biology. These tiny little parts — they’re not intelligent, but it’s so well designed and so efficient.

That’s also why ShapeTX was built with this idea: Can you boil down RNA to a computer code and make sense of it with AI?

Every company starts with a couple of people who are a bit like lunatics, and think that they can change the world and it will be easy. It never is, right? You start with three founders, and you recruit one person after the other. You attract great talent and you let them flourish.

Very interesting science is extremely risky, and it’s extremely difficult. You have to be able to fail and keep going. The common trait with great scientists is they tend to be very persistent.

It doesn’t feel like a job because everybody’s passionate about the cause. We’re trying to cure diseases like Alzheimers, Parkinson’s, Tourette Syndrome. It’s easy to go to bed in the evening and feel you’re doing something good.

Boston is the mecca of life science. Seattle is definitely smaller. But Seattle is at the right size for biotech. It has seen a biotech boom in the past eight years or so, I would say. Amgen used to be here. Juno did really well. This created a pool of people wanting to do the work.

We’re really trying to merge tech and biotech, so it’s the right city to be in. There’s quite a bit of focus on software engineering and AI in the city, and just the right amount of biotech.

We signed a quite sizable partnership two years ago with Roche to work on Alzheimer’s and Parkinson’s and other diseases we can’t disclose, worth north of $3.5 billion. And the reason why we partner is that there’re five million people in the U.S. with Alzheimer’s. It’s useful to have one of the biggest formal partners to eventually be able to commercialize the drug and deploy it properly.

The long-term goal is bringing these drugs to the clinic. We’re still a few years away, but the work is going really well.

What we are trying to achieve is a way to make drugs more efficient and cheaper. We’re trying to fix underlying genetic disorders that are incurable. There’s only a few approved on the market, and one of them is a drug fighting spinal muscular atrophy by Novartis. The cost of that drug is $2.5 million per treatment. So that’s a bit of a problem.

The cost of these drugs is very high. The way you save is by changing the underlying technology so you can make it cheaper. This is why I think AI is going to matter a lot. Instead of having, you know, 10 people in the lab, screening tens of thousands of guided RNA over months and months and months, we’ve discovered the role that makes this enzyme work really well. So now we can have a computer do it.

The gain in cost and efficiency is significant. The dream is one day for every drug, antibody, RNA, and small molecule, an AI software should be able to design that from scratch, for any disease. That changes, completely, the equation of cost for society.