Executive Conversations: Using AI and ML to accelerate joint learning in pharmaceuticals with consortium members of AION Labs

Like businesses in any other industry, pharmaceutical companies compete with each other. Despite their individual races to the finish line, the mission is the same: develop and produce lifesaving and life-changing treatments that move medicine, and humanity, forward. The recognition of this mutual goal, and the potential of breaking down siloes to share information, has led to the formation of AION Labs.

A first-of-its-kind global alliance, AION Labs is comprised of leading pharmaceutical companies, technology leaders, and investors that are pioneering the adoption of cloud-powered artificial intelligence (AI) and computational technology for drug discovery and development.

In this session of Executive Conversations, Dan Sheeran, General Manager of Healthcare and Life Sciences (HCLS) Industry at Amazon Web Services (AWS), sits down with three consortium members: Mati Gill, CEO of AION Labs; Jim Weatherall, Vice President of Data Science and AI, Research and Development at AstraZeneca; and Eran Harary, Global Head of Innovative Medicines & Clinical Development at Teva Pharmaceuticals. Together, they discuss how pre-competitive consortiums can use innovative technologies to accelerate collaboration, joint learning, and knowledge sharing—solving the biggest challenges of pharmaceutical research and development.

____

Dan Sheeran: To kick off our conversation, can you share with us the mission of AION Labs and what was the driving force behind starting it?

Mati Gill: We are an alliance bringing together brilliant inventors, scientists and technologists to solve the biggest R&D challenges of today’s pharma industry, with the support and guidance of years of accumulated know-how, data, and the experience of our partners. Our mission is to drive the creation and adoption of gateway technologies that will transform the process of drug discovery and development, while providing tremendous advancements in our understanding of diseases and treatments—benefiting patients and healthcare providers around the world.

We have two key goals that drive the mission. Our technological goal is to drive digitalized drug discovery and development by leveraging the capabilities of computational sciences and AI. This will allow us to discover new therapies, develop new insights into disease, and improve the time and cost efficiency of drug development. Our business goal is to create an average of five new startups annually that are applying AI and computational technology to drug discovery and development―ensuring their time at AION Labs sets them up for success.

We have been able to do this, in part, thanks to the support of the Israeli government, which presented an attractive offer for initiating our ambitious, multi-disciplinary partnership. We thank the Israel Innovation Authority for being a catalyst behind establishing the lab at the center of this consortium.

DS: Bringing together investors, pharmaceutical companies, and technology experts is a unique approach. What was the thought process behind creating this alliance with such a diverse stakeholder base?

MG: Everyone understands the promise of bringing in AI and computational technologies in drug discovery. But, its potential is yet to be fully realized because it is still such a challenge scientifically, operationally, and technologically. Bringing our partners together has created strength in numbers, which enables us to fully leverage the capabilities of how we discover and develop drugs using cloud-based technology.

Each partner brings in different advantages, and has ownership over specific aspects of the value chain as a part of the consortium. For example, AWS as a consortium partner, is bringing in its expertise in cloud technologies and AI for pharma. Similarly, we have a venture capital fund to help bring our startups to a place where they can be attractive for continuous funding with a good viable business model, not just great science. BioMed X helps us with their successful model to attract top scientific founders and seed their companies for success. So, bringing all these partners together gives us the strength in numbers to do something transformative.

Jim Weatherall: At AstraZeneca, our partnership with AION Labs aligns really well with our long-term goals of bringing differentiated medicines to improve patient outcomes. It was clear to us that we’d need to make sense of a multitude of scientific, biomedical, and healthcare data to find patterns and generate scientific hypotheses to drive the future of precision medicine. And, while we understood a lot about biology, we realized it made sense to include the incredibly smart, creative, and entrepreneurial startups that are looking at problems from different angles—using AI/ML [machine learning] and other cloud-based computational technologies. Which is why we are incredibly excited about what the consortium aims to accomplish.

Eran Harary: Although Teva has its own highly advanced innovative capabilities and technologies, some of which are cutting edge, having four pharmaceutical companies working together on the consortium is also helping us find the blind spots and solve pressing industry challenges, by doing away with information silos. We aren’t just doing away with the silos inside us, but also amongst us—encouraging conversation between organizations. AION Labs requires participants to go through an exercise where we all workshop exciting ideas, which are then shared among the organizations and screened for further pursuit. This has facilitated discussions amongst different parts of our company, as well as with teams outside of our organization. It’s triggering thought provoking questions, and fueling collaboration between companies to tackle broader industry challenges holistically.

DS: The concept of applying AI and ML to accelerate drug discovery and development has been popular for years, but it seems to really be gaining traction now. Do you agree with that perception, and if so, why now?

JW: Absolutely, there has been a recent acceleration in uptake and realized possibilities with artificial intelligence due to a few factors. Increased computing power, the popularity of cloud-based and other architectures to drive that computing, and the ability to meet needs around storage, processor cycles, and polarizability have all contributed to increased AI implementation. Those innovations have paved the way for modern machine learning, which has enabled AI-powered architectures that weren’t easy to build or scale in previous years. Outside of the technological drivers, the industry has become more educated on the art of the possible—it’s real and there are incredible opportunities out there.

EH: Drug discovery and development is a long and complex process, and it’s become clear that there are many opportunities to use both AI and machine learning to not only accelerate timelines but also optimize those processes. The applications range from clinical studies, to using predictive AI and ML, to selecting which initiatives and opportunities to pursue. Additionally, AI can serve in optimizing the drug approval process. All these use cases can help us come up with better medications and make them accessible to patients. The combination of machine learning and deep knowledge in chemistry and biology, in concert, can really create step function advances for drug discovery and development. It’s a part of our strategy, not just a nice-to-have tool that we play with.

DS: While we are fairly early in the journey, what have we learnt thus far and how is the model playing out in reality?

MG: The results so far have been promising. We have a number of bootcamps under our belt, and we’ve been able to see some really creative thinking by our applicants. What’s pleasantly surprised me is how scalable the initiative has been. We have a good model and a systematic process for reaching out to ground-breaking scientists, bringing in the partners to work together with, and setting up new companies for success. Our first two bootcamps, which focused on predicting likely successful candidates in lead optimization and the de novo antibody design, have led to finding top-caliber scientist-founders with very diverse backgrounds.

The second highlight has been to see these partners work closely as a team. Its heartening to see competitors come together as collaborators to learn from each other and share non-proprietary knowledge—albeit with all anti-trust requirements adhered to—to propel the industry forward.

One key element to making this complex partnership work is placing focus on building a strong culture at AION Labs where all people involved are laser focused on our mission and collective success, and of the right character and value-set which we stand for. This is key.

JW: We’ve also talked about having clusters of companies work on demanding challenges together. The possibilities of using AI and deep computation to improve the success rates of drug discovery and accelerate the likelihood of developing personalized, differentiated drugs can do wonders for improving patient outcomes.

DS: What does success look like to you, and what do you expect to see in the next five years?

MG: Success would be having most of the startups that graduate from AION Labs raise their next investment rounds and grow as independent companies. We want to establish this as a forum where high-tech life sciences entrepreneurs can come partner with the best experts in the field, and develop their technologies, as well as build their new companies. A lot of high-tech entrepreneurs, who were previously hesitant to step into the waters of biotechnology, would now find the resources and guidance to enter the space. In the process, they would find it appealing to put meaning behind the technologies they’ve developed, because it’s going to directly impact humanity.

Secondly, we want to prove that competitors and different disciplinary partners can work together towards a bigger mission for improving patient outcomes. If we can do that, we can support a macro view of how AI can propel this industry forward, across the value chain. That will evolve how we understand diseases, discover and develop new drugs, and create efficiency using common data platforms that leverage the cloud.

JW: In five years I think there will be many more established use cases in the discovery and development of new medicines that will utilize AI, transforming how research and development are carried out. Once we’re able to present a few of those success stories, there will be more of these types of platforms that will help patients. Ultimately, when you successfully develop technology in healthcare and life sciences, you’re doing good for humanity almost by definition. I hope that in five years’ time we’ve established some different models for fostering AI innovation.

EH: I see AION Labs as the flagship business case that will be disruptive enough to inspire many models and platforms that will solve huge problems. Right now, we have many opportunities and potential drugs, but we need to wisely manage resources to make them all happen. Eventually, I see the AION Labs process becoming an integral part of our internal processes, making us actually collaborate with other industry innovators to solve problems—which will bring those treatments into reality more quickly, and more cost-efficiently. This will open up an appetite for more collaboration in the future, because no company can face all the challenges the industry poses on its own.

____

DS: At AWS, we see the value in learning from each other, and as fellow collaborators who will be on the journey with this consortium, we are equally excited about the concept. We agree that it takes multiple diverse capabilities and partners to do something transformative. AWS is constantly exploring new ways to help our customers innovate, which is why we support this mission and encourage other aspiring collaborators to reach out to us. To learn more, read about our involvement with AION Labs.