Morning Panel: AI Partnerships: Integrating for Strategic Value

Daniel Brog, SVP at Locust Walk, led a discussion with 4 AI industry leaders Brian Alexander, SVP at Roche Genentech, David Harel, CEO & Co-Founder of CytoReason, Zach Taylor, SVP, Corporate Development & Strategy at BioNTech, and Liz Schwarzbach, CBO of BigHat Biosciences.

The conversation focused on the transformative potential of AI in drug discovery and development.  AI currently serves as a tool to improve drug development efficiency, but also can deliver proprietary value with large, complex data sets where AI platforms have an advantage over human driven analysis alone.  Emphasized was the crucial role of integrating data from various sources to enhance clinical development efforts, along with the need for a platform capable of integrating remote data. Speakers also discussed challenges such as integrating external data sets, collaborating with partners, and the importance of establishing clearing houses and deciding between building or buying data analysis systems. Speakers commented that it may be too early to value the financial impact of the AI component on drug discovery.  More time will be required to understand the impact of utilizing technology to drive programs versus internal logic alone.  Lastly, speakers stressed the importance of balancing innovation and governance in AI integration within the pharma and biotech industries, highlighting cultural sensitivity, resource allocation, and appropriate governance as key factors.

Lunch Panel: AI in Drug Discovery: Navigating Innovations and Limitations

Chandra Ramanathan, Ph.D., Head of External Innovation, Life Science Innovations Group at Danaher, led a discussion with Petrina Kamya, Ph.D., Global Head of AI Platforms at Insilco Medicine, Ramesh Durvasula, SVP R&D Information Technology at Eli Lilly, Anna Marie Wagner, SVP at Ginkgo Bioworks, and Chris Regan, Managing Director, Strategy and Transformation for Healthcare & Life Sciences at Microsoft.

The panelists started on a positive tone, each expressing enthusiasm for AI, noting the strategic importance to their organizations and referencing recent financing/transaction activity underlying their viewpoint (e.g., $1B Xaria fundraise). Beyond the general sense of optimism, two key themes emerged over the course of the conversation. Firstly, the integration of AI into the biotech industry will only accelerate and the breadth of workflows it disrupts will continue to broaden. Potential applications discussed ranged across the biotech value chain, from enabling discovery of novel therapeutic approaches and streamlining early clinical develop to optimizing patient recruitment for large clinical trials. Furthermore, there was the sense that the integration of AI into these capacities will ultimately become commonplace, with one speaker using the analogy that the distinction of ‘smartphone’ has disappeared from our lexicon as the technology has become omnipresent. For the second key theme, panelists stressed the importance of leveraging robust datasets to create multimodal learning models. Speakers noted that increasingly differentiated datasets represent a key feature in driving partnership interest for their organizations, not only due to data quality, but also the growing recognition of the importance of incorporating multiple inputs required to build more comprehensive and nuanced models.

Keynote: GenAI in the Life Sciences: Moving Hype to Reality

Delphine Nain Zurkiya, Senior Partner at McKinsey and Company, unpacked how AI has rapidly advanced in language understanding and image recognition, thanks to neural networks and machine learning, thereby democratizing access.  The ability to leverage and analyze multiple data sources at scale is starting to lead to a greater understanding of disease pathology and target identification.  However, only about 11 percent of the industry is currently utilizing AI learning platforms at scale to drive drug discovery.  We are still learning how best to deploy AI, integrate it into existing workflows and determine what multi-modal models will be most impactful.  Technology alone does not create value, but rather execution and how the industry utilizes AI to innovate faster and better is what matters.  Six elements are required for successful technology implementation: (1) strategy alignment, (2) talent, (3) partnerships, (4) monitoring, (5) continuous education and (6) risk management.  Companies must curate and connect data across their organizations and involve legal and others to monitor the system for inaccurate responses and system integrity.  Alliances with Big Tech have become more critical due to the broader expertise and compute necessary for multi-modal analysis.  Speakers deliberated on the challenges and opportunities of AI implementation in their industries, stressing the significance of comprehending AI limitations and optimal use scenarios.

AI Company Presentations

In two sessions spanning the morning and the afternoon, 25 top-tier AI/ML-centered drug discovery and development companies presented their mission across two rooms, outlining how AI/ML-based approaches enable differentiated and unique solutions to previously unaddressed unmet needs. While the development and use of proprietary AI models represented a foundational theme across these presentations, but the application and integration into the larger drug development process varied widely, as did company stage, modality, and indication.

While each of the presenting companies was individually selected by the conference selection committee and all were impressive their own regard, two companies that stood out as representative archetypes for diversity of approaches are Cellarity and Enveda. Cellarity’s AI model is built around visualizing cellular dysfunction to predict interventions and develop non-intuitive NCEs. Their lead candidate targets Sickle Cell Disease and is an oral compound that has demonstrated in vivo efficacy within the ranges of leading gene therapies. This lead compound shows a 2-3x improvement in HbF, however, it is still to be seen if this is enough differentiation to be meaningful in the space.

Enveda’s AI platform enables the annotation of structure and function of unknown chemistry at scale, in turn, optimizing the chemistry behind early-stage drug development, with a focus on natural products. Enveda’s lead candidate is a first-in-class immune cell trafficking inhibitor expected to begin human trials in 4Q 2024. Preclinical data suggests efficacy similar to JAK inhibitors, but with safety akin to Dupixent (a remarkable feat for a small molecule). As this landscape continually evolves and companies built around AI/ML in biotech advance their pipelines, the question has shifted away from an “if” AI/ML is going to influence the biopharma industry, to “how”. While the verdict is still out as to where AI will find its greatest value in the biopharma space, given the wealth of innovation and possibility displayed by the 25 presenting companies, the future of AI in biotech is bright.