Claude finds a new enzyme system with CRISPR-like DNA repeats
3-line summary
- Anthropic introduced a life sciences research group and lab, and shared early results in which Claude found a new enzyme system reminiscent of CRISPR.
- About 950 Claude agents searched DNA databases for 21 hours, gathering more than 200,000 reverse transcriptases and narrowing down candidates.
- The system, named ART, has no known function yet; the team shared the findings early in a preprint.
Anthropic introduced a life sciences research group and lab that use Claude for fundamental biology, and shared its first results. With only high-level direction from scientists, Claude independently found a new enzyme system with features reminiscent of CRISPR gene editing.
The team simply prompted Claude to search a massive DNA sequence database for interesting reverse transcriptases (enzymes that copy RNA into DNA). Over 21 hours, about 950 agents using 210 million tokens gathered more than 200,000 reverse transcriptases, picked out 3,500 new candidate systems and wrote human-readable reports on the 20 most compelling. Anthropic says that kind of analysis takes an expert weeks to months.
Along the way, one agent spotted evenly spaced, repeating DNA sequences next to an unusual reverse transcriptase gene. Found mainly in bacteriophages (viruses that infect bacteria), the system has three parts — the reverse transcriptase, a partner gene and a long array of repeats — and the team named it ART, for array-associated reverse transcriptases. The enzyme itself had been reported before, but Claude appears to be the first to notice the repeat array and accessory protein that define the system.
First experiments show the array is expressed as a set of distinct short RNAs, but what ART actually does is still unknown. All lab work was done by human scientists in a lab limited to low biosafety levels (BSL-1 and BSL-2) that does not handle human pathogens. Feng Zhang of MIT and the Broad Institute, a CRISPR pioneer, reviewed the preprint and called the finding intriguing and worth further investigation.
Why it matters
It shows AI moving beyond summarizing papers to spotting anomalies in vast datasets and forming hypotheses on its own. But this is an early result, before the system's function is known and before peer review, so its scientific significance still needs follow-up work.
Who it's for
- Beginners
- CRISPR itself was first noticed as an odd repeat sequence in bacterial DNA before becoming the basis of gene editing — which is why this finding draws attention.
- Developers
- Hundreds of agents gathering candidates, writing reports and then critically re-checking the evidence is a good template for splitting large analysis jobs across agents.
