Amazon Web Services said it is launching an AI-powered drug discovery tool to help researchers design and test novel drugs. The new agentic AI application, called Amazon Bio Discovery, aims to speed up preclinical drug discovery and get new antibody treatments to patients faster.
With Amazon Bio Discovery, scientists will gain access to a broad suite of biological foundation models (bioFMs), which are specialized AI models trained on large biological datasets. bioFMs can evaluate and accelerate the early-stage development of new clinical candidates.
The application also lets scientists use natural language to work with an AI agent that selects the right models, optimizes inputs and evaluates candidates. They can train models on existing experimental data for more accurate predictions and send candidates to physical labs for synthesis and testing -- with results feeding back into the application for rapid, continuous iteration. This creates what AWS calls a "lab-in-the-loop."
Advances in GenAI have led to a surge of models that can predict protein structures and evaluate drug candidates, but using them requires coding skills and infrastructure management. With dozens of models to choose from and limited benchmarking, bench scientists have struggled to adopt them, while computational biologists, who are experts in specialized AI skills, remain in short supply.
Amazon's new platform seeks to address this challenge by providing access to a system that combines computational design with wet-lab validation, according to AWS.
"AI agents make powerful scientific capabilities accessible to all drug researchers, not just those with computational expertise," vice president of AWS healthcare AI and life sciences Rajiv Chopra said. "These AI systems can help scientists design drug molecules, coordinate testing, learn from results and get smarter with each experiment. This combination of cutting-edge AI and the robust, secure infrastructure AWS has built for regulated industries allows scientists to accelerate antibody discovery in ways that weren't possible before."
One of the most promising applications of AI in drug discovery is its ability to narrow vast chemical and biological search spaces, helping teams identify the most promising drug candidates earlier in preclinical development.
Nai-Kong Cheung, M.D., Ph.D., Enid A. Haupt Chair in Pediatric Oncology at Memorial Sloan Kettering Cancer Center (MSK), knows firsthand how slow traditional antibody drug design can be. Working with the Amazon Bio Discovery team, Cheung used the platform's AI agent to design nearly 300,000 novel antibody molecules and then sent the top 100,000 to Twist Bioscience for testing. A process that typically takes up to a year was completed in just weeks.
"We're glad to be able to join forces with Amazon Bio Discovery to develop the next generation of antibodies that will potentially speed up the process to help patients worldwide," said Cheung. "As researchers, we spent 20 years just to prove that the first generation of antibody worked, and then we spent another 13 years getting it into the human form before getting FDA approval. This path has been very inefficient."
According to AWS, 19 of the top 20 global drugmakers use Amazon's cloud platform to power their research workloads. Early adopters of Amazon Bio Discovery include Bayer, the Broad Institute, Fred Hutch Cancer Center and Voyager Therapeutics, AWS said.
Other drugmakers, including weight-loss leaders Novo Nordisk and Eli Lilly, have also hopped on the AI bandwagon. Just this week, Novo Nordisk teamed up with OpenAI to develop new candidates for its obesity and diabetes pipelines.
In January, Lilly partnered with AI giant Nvidia to invest $1 billion to support an AI co-innovation lab in San Francisco. The pharma giant also signed a deal last month worth up to $2.75 billion with AI drug developer Insilico Medicine to gain the global rights to develop and market a suite of preclinical oral therapies.
Alivia Kaylor is a scientist and the senior site editor of Pharma Life Sciences.
