Khaberni - Amazon Web Services (AWS) launched its new app "Amazon Bio Discovery" powered by artificial intelligence, in a strategic move aimed at enabling scientists to design and test new drugs quickly and reliably, overcoming the technical obstacles that have long slowed the path of biological research.
The app contributed to the design of 300,000 antibody molecules and selected 100,000 for testing within weeks instead of a year.
Artificial intelligence in the service of biology
The new app provides researchers with direct access to "biological foundation models" (bioFMs), which are specialized AI models trained on massive biological datasets. These models work on generating and evaluating potential drug molecules, especially in the field of antibody therapies, significantly shortening years of traditional research in the early stages of drug discovery.
"Amazon Bio Discovery" not only provides models, but also offers an "artificial intelligence agent" that acts as a smart assistant understanding the natural language of scientists, helping them automate complex tasks, such as selecting the right models for research objectives, optimizing inputs, and evaluating clinical trial candidates.
Breaking technical barriers
This innovation addresses the programming skill gap that many biologists face; while generative AI has made leaps in protein structure prediction, the complex computational infrastructure remained an obstacle to adopting these technologies.
From the digital lab to reality
The app features the ability to create an interactive experimental cycle, where scientists can:
- Train models: using their own experimental data to improve the accuracy of predictions while ensuring complete privacy of intellectual property.
- Actual manufacturing: send drug candidates with one click to a network of integrated laboratories including partners like Twist Bioscience and Ginkgo Bioworks.
- Feedback: automatically input physical lab results back into the app to improve the next design, closing the loop of experiments in an integrated system.
Concrete success with the "Sloan Kettering" Center
The app proved effective in a real trial with the Memorial Sloan Kettering Cancer Center (MSK) where the research team led by Dr. Nay Kong Cheong managed to design about 300,000 new antibody molecules and selected the best 100,000 of them for lab testing within a few weeks, a process that previously took a full year.
The app also relies on the reliable infrastructure of AWS, which is currently used by 19 out of the top 20 pharmaceutical companies worldwide.
The app ensures complete data isolation, giving academic and pharmaceutical institutions full ownership of their results and research, with providing massive capabilities for scalability and performance.
