AnswersPublished20 sources
How Claude Found ART, a CRISPR-Like Enzyme System in Phage DNA
Anthropic says about 950 Claude agents spent 21 hours searching DNA sequence data and identified ART: a jumbo phage reverse transcriptase linked to a repeat array and accessory gene. ART stands out because the reverse transcriptase itself was known, while its defining genomic context—non coding repeated DNA plus a n...
Published byEdited with GPT-5.6 TerraImages generated with GPT Image 2
Research answer
Key takeaways
- Anthropic says about 950 Claude agents spent 21 hours searching DNA sequence data and identified ART: a jumbo phage reverse transcriptase linked to a repeat array and accessory gene.
- ART stands out because the reverse transcriptase itself was known, while its defining genomic context—non coding repeated DNA plus a neighboring protein gene—had not been recognized as a coherent system.
- Human scientists set the goal, reviewed the computational result, and conducted the laboratory follow up; ART’s function, mechanism, and applications remain unknown.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: How did Anthropic’s Claude autonomously identify the previously uncharacterized array-associated reverse transcriptase (ART) system by deplo. Article summary: Claude did not infer ART from a single automated database query: Anthropic gave it a high-level task to find unusual reverse-transcriptase systems, then a coordinated agent system searched sequence data and escalated ano. Topic tags: general, general web, user generated, documentation. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks,
Anthropic’s claim is not that Claude independently completed a biological discovery from start to finish. Instead, the company used a large, coordinated group of Claude agents to search DNA-sequence data for unusual reverse-transcriptase systems. That search highlighted an overlooked pattern in bacteriophage genomes: a reverse transcriptase paired with a long repeat array and an adjacent gene. Anthropic calls the proposed system array-associated reverse transcriptases, or ART. 18
The discovery was a multi-agent search, not a single query
Anthropic reports that roughly 950 agents worked for about 21 hours, using around 210 million tokens. The campaign gathered more than 200,000 reverse transcriptases, narrowed them to roughly 3,500 candidate systems, and produced 20 reports for closer review. 5
6
The crucial step was looking beyond an enzyme sequence in isolation. Agents examined reverse-transcriptase families alongside their genomic neighborhoods—the DNA features and genes near each candidate. One candidate contained an unusual repeating DNA pattern beside a reverse-transcriptase gene in bacteriophage DNA. 10
18
That context mattered because the underlying reverse transcriptase in a jumbo phage had already appeared in earlier work. According to Anthropic, the newly recognized feature was the surrounding architecture: an array of non-coding DNA repeats and an additional accessory protein gene of unknown function. 18
What makes ART CRISPR-like
ART is described as having three linked components:
- a reverse transcriptase (RT), an enzyme class that copies RNA into DNA;
- a long array of evenly spaced, non-coding DNA repeats; and
- a nearby accessory-protein gene whose role is not yet known.
18
12
The repeat array is the reason for comparisons with CRISPR. CRISPR systems are also notable for organized arrays of repeated DNA and sequence segments. But “CRISPR-like” is a structural comparison, not proof that ART performs CRISPR-style targeting, cutting, editing, or immune defense. 18
8
The evidence for ART as a distinct candidate system is therefore the combination of features, rather than the presence of a phage reverse transcriptase alone. Its genomic arrangement suggests a relationship among the RT, repeated sequences, and partner protein—but the mechanism remains a hypothesis. 18
9
What people did—and what Claude did
Anthropic characterizes the computational identification phase as autonomous: people supplied the high-level research objective, while the agent system explored candidates and generated analyses. Human scientists then assessed the candidate findings and performed laboratory work. 2
3
That division is important. The reported result is best understood as AI-assisted scientific discovery with human experimental validation, not as a replacement for molecular biology. The laboratory follow-up and further analysis supported ART as a coherent phage-associated system rather than a simple sequence-annotation anomaly. 2
7
Why researchers find it promising
CRISPR pioneer Feng Zhang called the RNA-repeat arrays associated with reverse transcriptases “genuinely intriguing” and said they merit further investigation. 5
The interest has two dimensions. Biologically, unusual repeat arrays can point to unrecognized nucleic-acid systems. Methodologically, the work suggests that agents can search enormous sequence collections, compare local genomic contexts, and elevate patterns that may be missed by pipelines designed around fixed annotations or known system types. 10
18
Neither point establishes that ART will become useful biotechnology. It does make ART a worthwhile research lead: a candidate system with an unusual architecture that can now be tested directly.
What remains unknown about ART
The central unanswered question is what ART does in nature. It is not yet known whether the system participates in phage defense, RNA processing, genome modification, or another function. 2
3
There is also no demonstrated ART application in gene editing, RNA editing, diagnostics, or therapeutics. Comparisons to CRISPR should be read as an invitation to investigate the system, not as evidence that ART is a CRISPR replacement. 5
7
Early results reportedly indicate that ART-associated arrays are expressed as discrete RNA units during Staphylococcus phage infection. That supports the idea that the arrays may produce multiple RNAs, but it does not show what those RNAs target, whether the reverse transcriptase acts on them, or what outcome they create. 2
An early, open research result
Anthropic released the finding alongside a technical preprint, so it should be treated as early evidence awaiting independent replication and a clearer mechanistic account. 5
6
The company also says its new Bay Area life-sciences laboratory operates at BSL-1 and BSL-2 and does not handle pathogens capable of infecting humans. 3
12 For now, ART’s most concrete significance is not a finished biological tool. It is a test case for whether multi-agent AI systems can reliably surface experimentally meaningful patterns from vast, complex biological datasets.
18
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "How Claude Found ART, a CRISPR-Like Enzyme System in Phage DNA"?
Anthropic says about 950 Claude agents spent 21 hours searching DNA sequence data and identified ART: a jumbo phage reverse transcriptase linked to a repeat array and accessory gene.
What are the key points to validate first?
Anthropic says about 950 Claude agents spent 21 hours searching DNA sequence data and identified ART: a jumbo phage reverse transcriptase linked to a repeat array and accessory gene. ART stands out because the reverse transcriptase itself was known, while its defining genomic context—non coding repeated DNA plus a neighboring protein gene—had not been recognized as a coherent system.
What should I do next in practice?
Human scientists set the goal, reviewed the computational result, and conducted the laboratory follow up; ART’s function, mechanism, and applications remain unknown.