AbbVie’s Fresh Look At How AI And Quantum Computing Will Transform Biotech
February 11, 2020 | After nearly 15 years building networks and technology systems for the high-frequency trading and proprietary trading industries, Brian Martin brings a fresh perspective to pharma. Now he’s the head of artificial intelligence and senior principal data scientist within AbbVie's R&D Information Research group and is applying his financial sector expertise to the data problems and opportunities facing biopharma.
Our innovations must produce value, Martin argues, operational value first, then analytic, and finally experiential. Operational value may be the lowest hanging fruit, but progressing through that pipeline to deliver greater value should be the goal in all of our innovation efforts.
On behalf of Bio-IT World, Bridget Kotelly spoke with Martin about the data challenges facing biopharma today, how we can improve, and what should be our next focus after AI.
Editor’s note: Martin will be speaking at the Bio-IT World Conference & Expo, April 21-23 in Boston, both during the Emerging AI Technologies program and in a preconference workshop on Tuesday, April 21.
Bio-IT World: Welcome Brian. I appreciate your time. If you could just start off by telling us a little about yourself and your background and what your role at AbbVie is.
Brian Martin: Information Research is part of our research and development organization that is focused on helping our scientists unlock the information that makes cures possible. When it comes to data, analytics, visualization, and anything around those areas, our Information Research team is there to help enable the scientists and the science.
I came to AbbVie in October of 2018 after many years in the finance community. I spent over a decade working in the high-frequency trading and proprietary trading industry, building technology systems, computer networks, IT organizational teams. I then spent about five years doing consulting work in the finance industry, most specifically as technology architect of the Common Securitization Platform, which is the joint venture Fannie Mae and Freddie Mac platform built to create an objective space for mortgage securitization after the subprime mortgage crisis.
Now that you're in the biotech space, what do you think some of the biggest challenges in pharma research are today?
Well, obviously if we start from the foundation of data, some of the biggest challenges are just breaking down silos. How do we get to the point where traditionally siloed organizations are no longer operating in isolated spheres of influence and are starting to converge their knowledge across those divisions? A lot of that from a foundational data perspective is mechanical in nature. How do we locate, how do we store, how do we harmonize and transport and transform data to be able to be used? But I think the real frontier there is how do we facilitate the evolution from data to information to knowledge to insight? And then most importantly, how do we put those insights into action? If our goal is to make a remarkable impact in the lives of patients—and, internally, for our scientists—how do we use those insights? How do we not just arrive at those insights, but then use them the right way? From the data perspective, the foundational challenges are still immense. I think that we are at the cusp of that type of harmonization, not just at AbbVie, but in most of the pharma companies that we talk with. They are at various stages of trying to de-silo all that data and information across their R&D organizations.
Tell me a little bit about what innovations are happening at AbbVie to achieve this, and what has you most excited?
I think that some of the biggest innovations right now are focused on delivering operational value. If we look at how value is delivered—especially from artificial intelligence capabilities and technologies in the space but also more generally—there's this virtuous cycle that every company hopes to engage, where tools drive users. Tools are used by users that generate data that help us create better tools, and it creates this momentum, engaging this virtuous feedback cycle. If we look at the types of value that can create, operational value is really sort of the easiest for most companies because you can deliver that with simple efficiencies.
Concepts and technologies like robotic process automation provide a really rapid way for companies to deliver some operational value right at the start and then be able to turn the benefits from that into resources they can use to move around that virtuous cycle.
From operational value, we usually see companies work backward in the cycle to analytical value. How can we provide better analytics, be they descriptive, predictive, prescriptive or otherwise, for the scientists? How can we get better about that process?
And then that eventually lets us deliver the experiential value, which is that third type of value. And that one is really transformative. The experience of your users in a company whether they're your coworkers or your colleagues—for us scientists, business folks, clinicians, patients—how do you deliver that transformative experience? Building systems and platforms with a human in the loop mindset becomes key to powering that momentum from experience.
When you start to deliver value in all three of those spaces—operational, analytical, experiential—it creates unparalleled momentum towards future capabilities driven by human experience and empathy in any space, and that is the power that AI technologies offer today.
Exciting. I understand you and AbbVie are involved in a working group called the Qupharm Alliance, organized to accelerate the use of quantum computing and drug discovery. Can you tell me a little bit about that effort, and how you see it helping?
If we use “artificial intelligence”—and I'll put quotes around that, because it is a horribly vague and ambiguous term—but if we look at where companies are with artificial intelligence technologies and things like that, quantum computing is much earlier, right? We think of it as being about where regular computers were in the late 1950s or early ‘60s in terms of the maturity of the technology, systems, and software. But we know that it has potential to really unlock our ability to solve problems that computationally we cannot solve with today's resources in today's approaches. And so from the science community especially, the idea of running simulations or optimizations at a level of capability that we can't do now is huge. When we think of the funnel that we're fighting through in bringing drugs to market, finding the right compound in a combinatorically vast space—all possible small molecules or all possible biologic molecules. It is an overwhelmingly daunting challenge.
Qupharm is this great vehicle to bring together, pre-competitively, the pharma companies to help advise industry and academia. Through this vehicle, we can give guidance on the biopharma industry point of view as to what are the challenging problems that quantum computing can help us solve. We can help them make decisions on where they should and could choose to focus their efforts with the understanding that we, as this collective of pharma companies, are there to help guide them and support them. By helping them validate use cases, validate outcomes from things that they've actually built, or providing data in appropriate spaces and with appropriate controls, we are really creating an industry-focused mechanism to push forward these possible use cases.
Then once a concept is really proven, you’ll see the pharma companies choosing to engage and invest in various different ways. But because we’re so early in the process, it's hard for any pharma company to really justify jumping into that space unless they happen to have resources laying around, which nobody does. So it's really an opportunity to engage collectively, pre-competitively in exploring this space together.
From the very beginning, I think the most important thing is that the focus of every conversation we've had is: “This needs to be about a better outcome for patients.” It's not about technology. It's not about how we run our businesses more cost-effectively. It's about better outcomes for patients faster. And that is the driving motif for all of us that are engaged in this Qupharm initiative. We know quantum computing can, in theory, and hopefully in practice, do this. So how do we collectively get it along that path faster?