In a pioneering project aimed at revolutionizing the design of antibodies and antibody-like molecules through the power of AI, Lawrence Livermore National Laboratory (LLNL) is working to transform medical countermeasure development and biologics discovery. The project is part of an interagency agreement between the Advanced Research Projects Agency for Health and the Department of Energy.
Under the three-year DeNOVO project, LLNL and other institutions will apply high-performance computing and AI to push the boundaries of antibody design. These efforts could overcome significant challenges in antibody design, including the sparseness of publicly available datasets and the substantial computing resources required for model training. DeNOVO will develop experimental methods for ultra-high-throughput antibody characterization to enable computational prediction and, ultimately, generative design of antibody-antigen interactions.
"AI-based design of antibody-based medicines will revolutionize drug discovery and dramatically accelerate our responsiveness to pandemic pathogens and other biothreats," said Matthew Jemielita, who leads the DeNOVO initiative at LLNL. "This program provides an incredible opportunity to bring together the resources, expertise and unique capabilities needed to deliver on the promise of in silico biodesign."
The initiative leverages LLNL's unmatched computational capabilities and foster interdisciplinary collaboration with world-class research partners, including the University of California, San Francisco, the Institute for Protein Design at the University of Washington, Los Alamos National Laboratory and A-Alpha Bio. This collaboration is crucial for the success of the project, requiring active interplay between computer scientists and experimentalists to collect training data, validate model predictions and develop new methods for data generation.
The potential impact of this initiative is profound: by enabling the proactive design of potent antibodies, it could significantly accelerate the discovery process and streamline downstream development, ensuring that new drugs reach patients with unprecedented speed. This advancement could rapidly accelerate the pace of drug discovery, potentially allowing for the identification of potent neutralizing antibodies against high-value targets in hours instead of months or years, according to the researchers.
