Capital at Scale Panel #CamTechWeek

So the first question that I wanna explore is how are, and we're gonna start small and go broader, how are you using AI? So one of the things about scale-ups that do get to be globally consequential is they are using AI differently than large companies that they're running circles around so really want to understand how is it helping you to continue to to scale what are you doing with it uh and maybe what lessons we you know others might have but how is it helping you solve problems um that give you confidence about the the future and that resonate with investors because then we're going to talk about what investors are looking for in those that deploy ai

Yeah, that's a really important topic.

So at Helix, we use AI to do broadly two things.

One is to understand complex rare disease biology and predict drugs that work on that biology in a much more complicated way and much more depth.

And then the other is to really run operational model that's incredibly cost effective and effectively not seen before.

And so the part that's incredibly exciting right now is the intersection of AI, deep tech and biology.

And so to use that to find, in our case, breakthrough drugs like an anti-tumor drug with no side effects, that is a breakthrough drug.

Talk about from nuts and bolts, where are you getting the datasets from? How are you using AI in that sense? Because I'm really curious about how that differs from some of the things that Tom might be investing in or Amelia might be investing in.

Yeah, so for us, for example, we invest a lot in what's called multi-omic data sets, so your genome, proteome, metabolome, et cetera.

And then we use that to define complex disease signatures and then predict patterns of drugs that would reverse basically a disease signature to a healthy one, but in a much more complex way.

And just to put this into perspective, That's a paradigm shift compared to the current drug discovery paradigm that's called target-based.

And that started in 2001.

So Cambridge actually had a very important role for the first human genome project, which resulted from the Sanger technology, so Frederick Sanger, who got the Nobel Prize.

And the naivety that came with that was that basically now we have sequenced all the genes.

Every gene is going to be linked to one disease.

And we're going to take this one gene, we're going to target it, or the protein it's sequenced for, and that's going to be your magic bullet.

And now fast forward more than 20 years, we're in an industry where it costs $2 to $3 billion to get a drug to approval and more than 90% failure rate.

And that is because the current drug discovery paradigm was too simple, so a single target.

Now you can use AI to not dump down your biology at the start, really multimodal complexity.

We have 22,000 genes, 60,000 proteins, and that's super exciting.

And so this is the thing that we were pioneering at the chemistry department.

It was the foundation for the company.

And yeah, now we have... real drugs in the clinic, breakthrough drugs.

Next year we're going to start registration trial to approval.

And if you look at the economics, the next trial, so registration to approval, will cost about $50 million.