Claude agents flag a CRISPR-like enzyme system in phage DNA

Anthropic says Claude agents flagged a previously uncharacterized enzyme system in the DNA of bacteriophages, the viruses that infect bacteria, and that its scientists then tested it in the lab. According to Anthropic's research post, the system, which the team calls array-associated reverse transcriptases (ART), has three parts: a reverse transcriptase, an enzyme that copies RNA into DNA; a partner gene of unknown function beside it; and a long array of evenly spaced repeated DNA. That repeat layout resembles a CRISPR array, the bank of RNA sequences that makes CRISPR gene-editing systems programmable. Anthropic's first experiments show the ART array is also expressed as a set of distinct short RNAs, which it says suggests something analogous may be at work, but it does not yet know what the system does.

The division of labor is the central claim. Anthropic's scientists wrote one prompt asking Claude to search a massive database of DNA sequences for interesting new reverse transcriptases, and they did the lab work; the company says Claude agents handled everything in between. Roughly 950 agents searched for 21 hours using 210 million tokens. They gathered more than 200,000 reverse transcriptases, picked out 3,500 new candidate systems and narrowed those to the 20 most compelling, each written up as a report for human review. Anthropic says that kind of analysis can take an expert weeks to months. The reverse transcriptase itself, found in a jumbo phage, had been identified in earlier studies; what Anthropic credits Claude with is noticing the repeat array and the accessory protein beside it, which it says Claude appears to be the first to have spotted.

The resemblance to CRISPR is why the find is worth chasing. Anthropic says only a handful of known systems share this combination of features, and all of them are programmable and can cut, copy or paste DNA; CRISPR became the foundation of gene-editing medicines, and several of the others are in development as tools. That is a reason to study ART, not evidence that it will become one. Feng Zhang, a CRISPR pioneer at MIT and the Broad Institute, reviewed the preprint and called the reverse transcriptase repeat arrays "genuinely intriguing" and worth further investigation. The preprint could not be retrieved for this explanation, so the account here rests on Anthropic's post. What it supports is that Claude found and documented a promising candidate with little human direction, and that the system's function and any practical use are still open questions.