Skip to main content
Back to timeline
AnthropicSource publication:

Anthropic puts Claude inside six drug developers' real pipelines, and a rare muscular dystrophy antibody reaches about twice the peak force and grip strength of vehicle-treated mice

Synopsis

Anthropic's life sciences team has formed deep research partnerships with six organizations developing medicines, BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group, and has begun drug discovery of its own, with Claude handling target assessment, molecule design, manufacturing paperwork checks and trial planning across the drug development process. Its acquired one-person company Mithril Medicine used Claude to coordinate 11 external labs and took a bispecific bridging antibody for LAMA2-related congenital muscular dystrophy from idea to mouse results in under 18 months, where the tetravalent construct gave roughly twice the peak force and grip strength of vehicle-treated mice. Most of this work remains preclinical.

Source-provided article image: *Anthropic’s life sciences team is working to accelerate curing disease. Over the past several months, our scientists have engaged in deep research partnerships with six organizations developing new medicines: BigHat Biosciences, Bristol My
Anthropic

Interpretation

Claude was placed inside real programs at six drug development organizations rather than confined to proposing initial molecule designs; the collaborations span antibody design, lead optimization, manufacturing and filing documents, company creation and clinical trial planning. Drug development spreads across target identification, molecule design, scalable manufacturing, regulatory filings and trial design, with no single bottleneck, while the most commonly discussed AI application has been proposing initial designs for new molecules. Over tests on 1,061 pages of documentation at Enveda, Claude caught all 11 issues human reviewers had spotted and flagged seven they had missed, cutting human review time from over 25 hours to less than three.

Anthropic also started drug discovery efforts of its own and acquired Mithril Medicine, a one-person company where Claude handled contracts, logistics and experiments with 11 external labs. Mithril had been developing treatments for two rare forms of muscular dystrophy with a very small team, with founder Cody Tranbarger delegating much of its operations to Claude. A possible treatment for one form of muscular dystrophy went from an idea to promising results in mice in less than 18 months, and if the candidates hold up, a drug is expected to be ready to start human trials as soon as late 2028.

Mithril's technical report gives one concrete molecular result: a bispecific bridging antibody for LAMA2-related congenital muscular dystrophy binds beta-dystroglycan on one side and the LG4-5 domains of laminin-alpha4 on the other, reconnecting the laminin-411 remaining in patient muscle to the muscle fiber membrane. The disease follows loss of laminin-alpha2, and the laminin-411 left in patient muscle neither polymerizes on its own nor binds its receptors strongly, so it does not compensate; earlier linker-protein strategies need two AAV vectors and have not entered an interventional trial. In a pilot study in laminin-alpha2-deficient dyW mice, mice treated with the tetravalent construct developed about twice the peak force and grip strength of vehicle-treated mice, and the bivalent antibody was recovered from muscle days after the last dose at concentrations exceeding those required for bridging in vitro.

Perspective

The demonstrated scope is mainly preclinical: Mithril's antibody results come from a pilot study in laminin-alpha2-deficient dyW mice, while Enveda's document checks and the Generate, Divergence and BigHat molecule results come from their own in vitro or laboratory tests. The intended reader is a researcher or engineer following how AI enters drug development workflows. The clinical trial planning work with Bristol Myers Squibb is planned to expand in 2027.

The report states that this research is still at an early stage and mostly preclinical, with the drugs not yet tested in humans. Mithril's antibody data come from a pilot study, and the strength gain with the tetravalent construct and the muscle exposure of the bivalent antibody were measured on different constructs. The Bristol Myers Squibb work is at an early stage of assessing trial plans for an early-stage drug, and the expansion is planned rather than reported.