Anthropic's Claude designs 35,000 antibodies targeting food allergies in one month

Anthropic's Claude generated approximately 35,000 antibody designs in about a month, with preliminary lab tests identifying promising drug candidates that may be used to treat food allergies. These results stem from a collaboration with Divergence Labs, one of six drug development organizations working with Anthropic to explore how artificial intelligence can accelerate drug discovery and development. In an announcement on October 27, Anthropic detailed several outcomes from these collaborations, including a tenfold increase in cancer-targeting receptor yields, improved antibody stability, and an 89% reduction in the time required for pharmaceutical document reviews.

The company is also conducting its own drug discovery research, including experimental treatments for rare pediatric muscular dystrophies. Most of this work remains in the preclinical stage, meaning potential therapies have not yet been tested in humans.

Antibody Design Targeting IgE

Divergence Labs is developing a treatment targeting immunoglobulin E (IgE), an antibody involved in allergic reactions. Unlike existing therapies that primarily neutralize IgE in the bloodstream, the company aims to isolate IgE from immune cells. Achieving this goal requires a molecule composed of two components: one that tightly binds to IgE and another that aids in removing it from immune cells. Claude transformed the company's research objectives into design plans and generated approximately 35,000 antibody designs primarily on its own within a month.

Preliminary lab tests identified three designs capable of binding to IgE, as well as 20 designs able to isolate IgE from immune cells. According to Anthropic, three of the latter appear to remove more IgE than standard drugs. However, researchers have yet to combine these two functions into a single drug. Divergence Labs plans to develop a molecule that combines both components.

Improving CAR T Cell Therapy

Another collaboration with Generate Biomedicines focuses on improving CAR T cell therapy, which uses genetically modified immune cells to identify and attack cancer. Six researchers used Claude Code to redesign portions of the genetic instructions that control these cells. One of the designs is approximately 70% shorter than the standard version. In lab tests, the modified cells produced about ten times the number of cancer-targeting receptors, killed tumor cells more effectively, and exhibited longer-lasting activity, Anthropic reported.

This approach helps adapt CAR T cell therapy for solid tumors, which are currently primarily treated for blood cancers. However, these findings have not yet established effectiveness in patients.

Efficiency Boost in Pharmaceutical Record Reviews

In 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 over ten times the target binding strength, while one cancer-targeting candidate remained soluble at temperatures 18 degrees higher than the original version. Another design aimed to extend the drug's duration in the body, showing a more than 100-fold reduction in aggregation tendencies.

Claude also assisted Enveda, a drug discovery company, in reviewing pharmaceutical manufacturing records. In a 1,061-page document, the AI identified all 11 issues previously found by human reviewers and flagged seven additional potential problems. The company reported that human review time decreased by 89%, from over 25 hours to less than three hours.

Anthropic is also collaborating with Bristol Myers Squibb on clinical trial planning and has acquired Mithril Medicine, which is developing experimental treatments for two rare muscular dystrophies. One Mithril project achieved promising results in mice within 18 months. Anthropic stated that if development proceeds smoothly, a candidate drug could enter human trials as early as the end of 2028. The company acknowledged that its research is still in the early stages, and subsequent tests will be needed to determine whether these experimental treatments are safe and effective in humans.

Prospects for AI in Drug Development

Claude's antibody design and drug discovery capabilities demonstrate the potential of artificial intelligence in the medical field, particularly in accelerating the development of new therapies. Through collaborations with multiple drug development organizations, Claude is not only able to generate a large number of antibody designs but also enhance the efficacy and safety of therapies. This advancement not only helps address challenging health issues like food allergies but may also facilitate treatments for other diseases, highlighting the important role of artificial intelligence in drug research and development.