Anthropic’s Claude generated roughly 35,000 antibody designs in about a month, with early laboratory tests identifying promising candidates for a potential food allergy treatment.The results emerged from a collaboration with Divergence Labs, one of six drug development organizations working with Anthropic to explore how artificial intelligence can accelerate the discovery and development of medicines.In an October 27 announcement, Anthropic detailed several results from these partnerships, including a 10-fold increase in cancer-targeting receptor production, improved antibody stability and an 89% reduction in pharmaceutical document review time.The company is also conducting its own drug discovery research, including experimental treatments for rare childhood forms of muscular dystrophy. Most of the work remains preclinical, meaning the potential treatments have not yet been tested in humans.Designing antibodies against allergiesDivergence Labs is developing a treatment that targets immunoglobulin E (IgE), an antibody involved in allergic reactions. Unlike existing treatments that primarily neutralize IgE circulating in the bloodstream, the company aims to detach IgE from immune cells.Achieving this requires a molecule with two components: one that binds tightly to IgE and another that helps remove it from the immune cell.Claude translated the company’s research goals into a design plan and generated approximately 35,000 antibody designs over a month, largely autonomously.Initial laboratory testing identified three designs capable of binding to IgE and 20 that could detach it from immune cells. Three of the latter appeared to remove more IgE than a standard medication, according to Anthropic.However, the researchers have not yet combined the two functions into a single drug. Divergence Labs plans to develop a molecule incorporating both components.Another collaboration, with Generate Biomedicines, focused on improving CAR T-cell therapy, which uses genetically modified immune cells to recognize and attack cancer.Six researchers used Claude Code to redesign part of the genetic instructions controlling these cells. One design was approximately 70% shorter than the standard version.In laboratory tests, the modified cells produced roughly 10 times as many cancer-recognizing receptors, killed tumor cells more efficiently, and remained active longer, Anthropic reported.The approach could help adapt CAR T-cell treatments, currently used primarily against blood cancers, for solid tumors. However, the findings do not establish effectiveness in patients.Faster testing and documentationIn another project, BigHat Biosciences connected Claude to its automated laboratory system to improve two experimental antibody fragments.Researchers manufactured and tested approximately 1,010 molecular variants over 10 weeks. The best designs showed more than 10-fold stronger target binding, while one cancer-targeting candidate remained soluble at temperatures 18 degrees Celsius higher than the original.A separate candidate designed to extend a drug’s duration in the body showed more than a 100-fold reduction in its tendency to clump together. Claude also helped Enveda, a drug discovery company, review pharmaceutical manufacturing records.Across 1,061 pages of documentation, the AI identified all 11 issues previously found by human reviewers and flagged seven additional potential problems. The company reported an 89% reduction in human review time, from more than 25 hours to less than three.Anthropic is also working with Bristol Myers Squibb on clinical trial planning and has acquired Mithril Medicine, which is developing experimental treatments for two rare forms of muscular dystrophy.One Mithril program reached promising results in mice within 18 months. Anthropic said a candidate could enter human trials as early as late 2028 if development progresses successfully.The company acknowledged that its research remains at an early stage. Further testing will be needed to establish whether the experimental treatments are safe and effective in humans.