Robyn Karnauskas : That’s really helpful. And I guess a follow-up, it’s like $1 billion is a lot. So have you been able to negotiate since you’ve been working with these companies, a little bit more disclosure about what you need to see to get those milestones? And then my last question, sorry for so many. You don’t talk a lot about OmniDeep, and I know you believe that nature-based [indiscernible], but I’m just wondering if you’re willing to leverage OmniDeep platforms to AI/ML-based in silico antibody design. It’s a hot topic right now, so I thought I’d ask that question.
Kurt Gustafson : On the first part, just financially, the way these deals are structured, they’re really structured the same way as our other deals, right? So there are typically clinical stage milestones as they progress through the clinic and royalties. The difference is that the magnitude of the payments are larger and mostly because of a function of the exclusivity on which we have written these deals. On the antibody side, all of the things that people are going after, those are nonexclusive targets. Whereas with the ion channels, these targets are being licensed out on an exclusive basis. And as a result, that’s what triggers the larger economics. But there’s nothing unusual necessarily about the types of things they fall within that same sort of deal structure of upfront payment and milestones and royalties.
Matthew Foehr : Yes, and I’ll be happy to comment on the AI question, Robyn. Obviously, given the visibility and use of AI associated with technologies or industries that are, I’ll say, highly visible in a popular sense or counted in a popular sense. This is obviously a question we do get. And OmniDeep, I’ll just say, is our suite of in silico tools for therapeutic antibody discovery and optimization that are really woven throughout our various technologies and capabilities. And these tools include the structural modeling and large — very large multi-species antibody databases, AI and machine learning sequence models and more. And it really — it allows for optimization of identification of candidates that come out of our technology.
Now there’s — to the core of your question, obviously, there’s a lot of discussion around AI and its use — kind of sole AI approaches. And I think the element of that, that might not be as well understood is that there are really some important considerations and limitations of that, and that’s why we think there’s so much power in not only the biological intelligence of our animals, but also the leveraging our AI capabilities. Now we’ve been using in silicon and AI tools in our downstream work for a long time actually, especially on the screening side and some of our work around the ion channels and transporters. So we have deep expertise here in our organization that we really kind of rolled out in the concept of OmniDeep in Q2, but these have been woven throughout our organization and technology from the spirit of cutting edge and good science for quite a long time.
And in fact, more than a couple of years ago, we actually did a deal with landing AI for a vision portion of AI that we incorporated into our exploration platform successfully that has really been kind of a wild success story around how we leverage our screening. But I’ll say on the technical level, there’s a lot of limitations to just AI approaches. So really the power of OmniDeep comes from marrying it to, what I’ll call, the biological intelligence of our transgenic animals because a carefully engineered transgenic animal system really has many of the tests that are needed to select a winning antibody inherently built into them as natural checkpoints. So you can essentially try and test millions of different sequence possibilities rather than just doing it in a model.
And obviously, the biological system can weave those out. Now we do leverage AI and other ways downstream from that with large amounts of data. And I think that’s where a lot of that power comes from. So I know I got a little technical there, but hopefully, that makes sense.
