Building Foundational AI Infrastructure for Holistic View of Biology with Jean-Philippe Vert Bioptimus

Jean-Philippe Vert, the Co-Founder and CEO of Bioptimus, is building a foundational AI model for biology to solve the problem of siloed biomedical research. Key goals are to bridge the translational gaps between drug discovery and development, and between clinical research and real-world patient outcomes, and to redesign clinical trials for greater efficiency and improved results. Creating digital twins of patients is a way to simulate treatment outcomes and create synthetic control arms for clinical trials, ultimately lowering the risk and cost of drug development and enabling the creation of new medicines for a broader range of conditions, including rare diseases. Jean-Philippe explains, "So at its core, what we try to build at Bioptimus is the foundational AI infrastructure for biology. The problem we're trying to solve is that biology is complex and operates across different scales, from genes and proteins to cells, organs, patients, etc. And historically, lots of research, lots of biological, biomedical research has been very siloed, has been focusing on specific aspects of biology, like studying only genes or studying only cells. What we are building at Bioptimus is an AI-intelligent system that can see across all these layers, all of these cases, to get a holistic picture of biology. And it's not only a scientific endeavor, but the reason why it's hard to make a drug today, why so many diseases remain untreated, is that the siloed nature of biomedical research has created difficulties in how we move from research in discovery, like understanding a disease, to making a treatment for the patients."   "So we have indeed a model called H-Optimus, which is a foundation model for one type of modality, which is one thing you see in an image. It's for histopathology slides. This is when someone has, for example, a cancer, you take a biopsy and then typically a pathologist looks at the biopsy under the microscope to characterize the disease, to see if there are cancer cells, to see the shape, to see the organization, and so to pose a diagnostic and suggest a treatment. We have trained an AI system that helps pathologists be better because our systems have been trained by looking at billions of such images, and so have a very detailed understanding of the subtle variations that can be observed in images."  #Bioptimus #ArtificialIntelligence #DrugDiscovery #Biotechnology #PrecisionMedicine #FoundationModels #BiologyAI #ClinicalTrials #CancerResearch #RareDiseases #DigitalHealth #Innovation bioptimus.com Download the transcript here