Amazon launches Bio Discovery

Amazon Web Services launched Amazon Bio Discovery on April 14, a new software tool that lets scientists run early drug-discovery workflows without writing code. (reuters.com) Drug discovery starts with finding and testing large numbers of possible molecules, then narrowing them to a few candidates worth making in a lab. Amazon says Bio Discovery gives researchers access to biological foundation models, a type of artificial intelligence trained on large biology datasets, to generate and evaluate those candidates. (aws.amazon.com) The product is aimed first at antibody research, where scientists try to design proteins that bind to disease targets. Amazon says researchers can upload a target structure, set design goals, generate thousands of ranked candidates, and send selected ones to lab partners for synthesis and testing from the same application. (aws.amazon.com) Amazon says the software includes more than 40 biology models, plus agents that help users choose models, tune inputs, and compare results. The company also says teams can upload proprietary models or licensed third-party models and turn finished workflows into reusable templates. (aws.amazon.com) The pitch is less about replacing scientists than removing the handoffs between computational biology and bench work. Amazon says many lab teams still rely on specialists to code pipelines, manage computing infrastructure, and coordinate outside testing, which slows each design-test cycle. (aboutamazon.com) That bottleneck has become more visible as drugmakers and technology companies push artificial intelligence deeper into pharmaceutical research. Reuters reported that Amazon is positioning the tool as a way to make complex computational workflows accessible to more scientists on Amazon Web Services. (reuters.com) Amazon says the system creates a “lab-in-the-loop” process, meaning software predictions are checked with real-world lab results and then fed back into the next round of design. In its launch materials, the company said Memorial Sloan Kettering Cancer Center used the approach to cut antibody design work for potential pediatric cancer therapies from months to weeks. (aboutamazon.com) The launch also puts Amazon into a more crowded contest over artificial intelligence tools for drug development. Axios reported that the move intensifies competition among companies promising faster and cheaper early-stage research, an area where many tools still need to prove they can turn better predictions into approved medicines. (axios.com) For Amazon, the immediate goal is to make its cloud platform part of how research teams design, test, and refine drug candidates. For scientists, the test comes next: whether a no-code system can shorten the path from a computer-generated idea to a molecule that works in the lab. (aws.amazon.com)