Anthropic AI Discovered New Enzyme System
The Claude AI model identified a novel DNA enzyme system while analyzing over a billion protein clusters.
Updated on Sept. 23, 2026 in Biotech

Live Poll
Do you trust AI agents to conduct original scientific research without direct human oversight?
Anthropic researchers used the Claude Mythos 5 model to discover an undescribed enzyme system dubbed array-associated reverse transcriptases (ART) in bacteriophage DNA. The AI identified these systems within a database containing 1.9 billion protein clusters.
Why it matters
This discovery demonstrates the utility of large language models in accelerating complex biological research by screening massive datasets in a fraction of the time required by traditional methods. Anthropic is now conducting laboratory experiments to confirm the biological function of these newly identified arrays.
The AI campaign processed 215.6 million tokens over 21.5 hours across 949 agent sessions. The discovered ART arrays contain 3 to 21 copies of a short DNA repeat and lack cas genes.
The players
Anthropic
Anthropic is an AI safety and research company that develops large language models like Claude.
Claude Mythos 5
Claude Mythos 5 is an artificial intelligence model developed by Anthropic for complex research and data analysis tasks.
Coefficient Bio
Coefficient Bio is a biotech firm that Anthropic acquired for $400 million in April 2026.
OpenAI
OpenAI is an AI research organization that competes in the life sciences model sector with products like GPT-Rosalind.
The details
Anthropic utilized its Claude Mythos 5 model to scan 1.9 billion protein clusters, eventually flagging 19 reports for expert review. The system is comprised of an enzyme, a partner gene, and an array of DNA repeats found in bacteriophage DNA.
Timeline
Spring 2026: Anthropic established its molecular biology group.
April 2026: Anthropic acquired Coefficient Bio and OpenAI debuted GPT-Rosalind.
June 2026: Anthropic released Claude Code and Claude Science.
September 23, 2026: Anthropic publicly announced the enzyme discovery.
The Big Picture
The discovery of ART systems was reviewed by researchers at the Broad Institute, marking a collaborative bridge between AI-native biotech startups and established genomic research powerhouses. This integration of silicon-based screening and wet-lab validation signals a paradigm shift in how novel genomic elements are cataloged.
This research suggests that AI tools may eventually lower the cost and time required for drug discovery and genetic engineering, potentially accelerating the development of new medical treatments. Users interested in biotechnology may see faster iteration cycles in life sciences products as AI models become standard tools in laboratory settings.
The takeaway
The use of generative AI to identify novel biological structures represents a significant departure from traditional, manual sequence analysis. Readers can expect AI to play an increasingly central role in classifying undiscovered biological data as the technology continues to mature.
What happens next
Anthropic is currently soliciting research proposals from external scientists to further investigate the function of the identified ART system.
Further reading
Learn more about the intersection of AI and biology in our Biotech section.
Live Poll
Do you trust AI agents to conduct original scientific research without direct human oversight?










