Biohub announced that the new partnership within the Virtual Biology Initiative brings the total resource commitment to 1.8 billion dollars. Following artificial intelligence studies that model life molecules, Google DeepMind, Isomorphic Labs and Meta together provide $300 million to the initiative focusing on biological data production. However, the total amount covers a broader contribution package than the new cash investment transferred all at once.
The announcement dated October 7 adds data, computing resources and new measurement technologies to the total commitment, in addition to financing. The US Department of Energy plans to allocate over $500 million over five years. The US National Institutes of Health, on the other hand, associates the relevant data sets and information infrastructure it has created with a federal investment of over $500 million in the past with the initiative.
Therefore, the NIH contribution should not be read as an entirely new $500 million cash payment. Biohub will work with NIH to bring existing datasets into common standards for AI training. The partnership aims to facilitate the use of different resources in the same research infrastructure.
The initiative’s goal is to create an open data foundation for models that can predict how cells will respond to interventions. Such a model could help researchers test some biological questions digitally. On the other hand, organizations have not yet announced a completed system or approved treatment that perfectly simulates all human cells.
Biohub creates open database for cell models
Biohub had initially committed $500 million for the five-year program, which it launched in April. $400 million of this goes to large-scale data generation through measurement, imaging and biological engineering technologies. The remaining $100 million supports research outside the organization and international data production efforts.
The program aims to jointly examine the molecular structure of the cell, its location in the tissue, and its changing behavior over time. Biohub plans to observe details inside the cell with cryo-electron tomography and large numbers of cells in living tissues with new microscopy methods. Additionally, molecular and cellular engineering tools will help measure the impact of interventions.
The organization maintains its goal of making the data it produces open and free to researchers around the world. With this data, scientists will be able to train new models and investigate hypotheses about cell behavior. NVIDIA supports the initiative with accelerated computing infrastructure, domain-specific software and technical expertise.
In addition to predicting living systems, NIH’s Bio Genesis Mission program focuses on accelerating drug discovery and investigating the causes of chronic diseases. The institution also includes infrastructure and governance efforts that protect sensitive biomedical data. This framework reveals that commonalizing data is not just about buying more computing power.