# Claude discovers a novel enzyme system with CRISPR-like repeats

> Anthropic's new AI-powered biology lab discovered a previously unknown enzyme system by having Claude autonomously search through massive DNA databases.
- Title: Claude discovers a novel enzyme system with CRISPR-like…
- Summary: Anthropic's new AI-powered biology lab discovered a previously unknown enzyme system by having Claude autonomously search through massive DNA databases. AI…
- Keywords: anthropic, claude, artificial intelligence, biology research, enzyme discovery, crispr, reverse transcriptase, dna analysis, bacteriophages, genome mining, ai-driven discovery, life sciences
- Source: Anthropic — https://anthropic.com/news/claude-discovers-novel-enzyme-system
- Read time: 3 min
- Topics: anthropic, claude, artificial intelligence, biology research, enzyme discovery, crispr, reverse transcriptase, dna analysis, bacteriophages, genome mining, ai-driven discovery, life sciences
## Anthropic launches biology research lab
Anthropic formed a new life sciences research group in spring 2026 to explore whether AI models can systematize and accelerate biological discoveries by analyzing DNA datasets and generating hypotheses at scale.
## History of biology's biggest breakthroughs
Restriction enzymes from bacterial immune systems launched biotechnology. Taq polymerase from Yellowstone bacteria enabled PCR diagnostics. CRISPR, first noticed as unusual DNA repeats, became the foundation for gene editing medicine.
## Claude discovers array-associated reverse transcriptases
Claude autonomously discovered a novel enzyme system called array-associated reverse transcriptases (ART) associated with DNA repeats resembling CRISPR arrays, found in bacteriophages.
## The search that found ART
Claude agents spent 21 hours searching a massive DNA database using 950 agents and 210 million tokens. One agent spotted a repeating pattern of DNA sequences next to an unusual reverse transcriptase gene.
## What Claude noticed about ART
Claude identified the system's defining features: an associated array of non-coding DNA sequences and an additional accessory protein of unknown function, alongside a reverse transcriptase from a jumbo phage.
## Expert validation from CRISPR pioneer
Feng Zhang, MIT professor and CRISPR pioneer, called the discovery exciting and said the identification of RNA-repeat arrays associated with reverse transcriptases is genuinely intriguing and merits further investigation.

> This is an exciting example of how AI agents can contribute to biological discovery.
## How Claude searched the database
Scientists gave Claude a single prompt to search for interesting reverse transcriptases. Claude agents then independently combed through the database, investigated distinct RT families, and used their own judgment to identify compelling candidates.
## Claude's workflow with scientists
Claude searches DNA databases for uncharacterized proteins, reads literature, reproduces established results, searches for family members that fit no known system, and writes reports proposing functions with supporting evidence.
## Scale of Claude's analysis
Claude agents gathered over 200,000 reverse transcriptases, picked out 3,500 new candidate systems, and narrowed those to 20 most-compelling candidates. This type of analysis typically takes expert scientists weeks to months.
## Claude's moment of discovery
While examining raw DNA sequences, Claude noticed a tandem repeat array and exclaimed it was spectacular and CRISPR-like. It then counted repeats, measured spacing, compared layouts, and searched literature before concluding it found a new system.

> [The DNA next to the RT] is spectacular: I can see by eye a tandem repeat array … that's a CRISPR-like … repeat array?!
## Structure of the ART system
ART consists of three parts: a reverse transcriptase, a partner gene beside it, and a long array of evenly spaced DNA repeat sequences. The repeat layout resembles CRISPR arrays, suggesting the system may be programmable.
## Early experiments on ART function
First experiments show that the ART array is expressed as a set of distinct short RNAs, suggesting something analogous to CRISPR may be at play. Further experiments are underway to determine how ART works.
## Anthropic's biology lab team and approach
The Bay Area lab team specializes in computational approaches to read DNA, interpret evolution, and identify biological systems. All lab work is performed by human scientists, with Claude helping interpret data and generate hypotheses.
## Future of AI-driven biological discovery
Anthropic is sharing these early findings to demonstrate Claude's capabilities and invite other scientists to explore how AI can support their research across genomics and other fields.
## Key takeaway

AI agents can systematically discover novel biological systems by analyzing vast datasets of DNA sequences, potentially accelerating scientific breakthroughs in the way restriction enzymes and CRISPR once did.