Anthropic’s Claude AI autonomously discovers novel CRISPR-like enzyme system in phage DNA
Anthropic announced that its Claude AI model autonomously discovered a previously unknown enzyme system in bacteriophage DNA, named Array-associated Reverse Transcriptase (ART). Using approximately 950 Claude agents running for 21 hours and consuming about 210 million tokens, the AI screened nearly 200,000 reverse transcriptases, narrowing to 20 candidates. One agent identified regularly spaced repetitive DNA sequences adjacent to an unusual reverse transcriptase, resembling CRISPR. Human lab verification confirmed the array produces short RNAs. The system’s function and potential as a gene-editing tool remain unconfirmed.
IllustrationEditorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page reads the event directly, while its address stays stable when the title changes.
- Summary covers the current reports
Cross-source coverage
Reporting timeline
AI Agent Claude Autonomously Discovers Novel Enzyme System in Phage DNA
Anthropic announced on February 23 that its life sciences research group, using the AI model Claude, autonomously discovered a previously uncharacterized enzyme system in bacteriophages, named Array-associated Reverse Transcriptase (ART). Starting from a single high-level prompt to find novel reverse transcriptases in large-scale DNA sequences, approximately 950 Claude agents worked in parallel for 21 hours, consuming about 210 million tokens. They screened nearly 200,000 reverse transcriptases down to about 3,500 candidate systems, then focused on 20 key targets. One agent identified regularly spaced repetitive DNA sequences adjacent to an unusual reverse transcriptase, compared them to known families, and submitted the system for human review. ART consists of a reverse transcriptase, a companion protein gene of unknown function, and a CRISPR-like repetitive DNA array. Human teams confirmed the array is expressed and produces short RNAs. CRISPR pioneer Feng Zhang called it an exciting example of AI contributing to biological discovery. Anthropic emphasized that AI handled data retrieval and hypothesis organization, while wet-lab experiments were performed by humans. The actual biological function and potential as a gene-editing tool remain to be experimentally validated.
AI Agent Claude Autonomously Discovers Novel Enzyme System in Phage DNA
Anthropic announced on February 23 that its life sciences research group, using the AI agent Claude, autonomously discovered a previously uncharacterized enzyme system in bacteriophages, named array-associated reverse transcriptase (ART). The discovery was made after a single high-level prompt asking Claude to find novel reverse transcriptases in large-scale DNA sequences. Approximately 950 Claude agents worked in parallel for 21 hours, consuming about 210 million tokens, to screen nearly 200,000 reverse transcriptases down to 20 candidates. One agent identified a phage with an unusual reverse transcriptase adjacent to regularly spaced repeated DNA sequences, which it cross-referenced against known families and literature before submitting for human review. ART consists of a reverse transcriptase, a companion protein gene of unknown function, and a CRISPR-like repetitive DNA array. Human lab verification confirmed the array expresses and produces short RNAs. CRISPR pioneer Feng Zhang reviewed the findings, calling it an exciting example of AI contributing to biological discovery. Anthropic emphasized that AI handled only data retrieval and hypothesis sorting, while wet-lab experiments were done by humans. The practical biological function and potential as a gene-editing tool remain unconfirmed and require further experimental validation.
Anthropic's Claude AI Discovers New Enzyme System, Triggering Gene-Editing Stock Selloff
On September 23, Anthropic CEO Dario Amodei announced that the company's AI model, Claude, autonomously discovered a novel enzyme system named ART (Array-associated Reverse Transcriptase) hidden in phage DNA. The discovery was made by approximately 950 AI agents running for 21.5 hours, analyzing 1.9 billion protein clusters. The finding, published as a preprint, has not yet been peer-reviewed, but it caused immediate market reactions: gene-editing stocks including Prime Medicine, Editas Medicine, CRISPR Therapeutics, and Beam all fell sharply. MIT professor Feng Zhang called it 'an exciting example of AI agents participating in biological discovery.' Anthropic emphasized that ART is a new biological system, not a gene-editing tool. The company noted that the result could not be replicated in 10 subsequent runs, attributing this to the scale of the search and the unpredictability of agent behavior. The article also notes that AI companies are accelerating their push into life sciences, with Anthropic acquiring Coefficient Bio in April and OpenAI launching GPT-Rosalind for life science research.
Read sourceShow 13 older updatesHide older updates
Anthropic's Claude Discovers Novel DNA Enzyme System Using 950 AI Agents and 210 Million Tokens
Anthropic announced that its AI model Claude autonomously discovered a previously unknown enzyme system in phage DNA, named ART (Array-associated Reverse Transcriptase system). The discovery involved 950 Claude agents running for 21 hours, consuming approximately 210 million tokens to scan over 200,000 reverse transcriptases from public genomic databases. The system identified 3,500 candidate systems, narrowed to 20 core candidates, and ultimately one novel enzyme system that structurally resembles CRISPR. Human scientists at Anthropic's molecular biology lab, established in spring 2026, then conducted wet-lab experiments to verify the finding. CEO Dario Amodei stated the result would be a career-defining achievement even if made during his PhD. The article speculates that ART could become a next-generation gene-editing tool, though further research is needed to understand its function. The report frames this as a paradigm shift in life sciences research, with AI acting as a 'chief exploration officer' accelerating biological discovery.
Read sourceClaude AI Discovers Novel DNA Enzyme System in 21 Hours, Anthropic Announces
Anthropic's molecular biology lab announced that its Claude AI model autonomously discovered a previously unknown enzyme system in bacteriophage DNA, named ART (Array-associated Reverse Transcriptase system). The discovery involved 950 AI agents running for 21 hours, consuming approximately 210 million tokens to scan over 200,000 reverse transcriptase sequences from public databases. The system, structurally similar to CRISPR, was identified by Claude detecting a repeating DNA array pattern that human researchers had previously dismissed as noise. Anthropic CEO Dario Amodei stated the finding would be a career-defining achievement even in a PhD program. Human scientists subsequently verified the discovery through wet-lab experiments. The article positions this as a paradigm shift in life sciences research, with AI acting as a 'chief exploration officer' to accelerate biological discovery, potentially leading to major medical breakthroughs within 5-10 years.
Read sourceAnthropic Establishes Life Sciences Research Team; Claude Discovers Novel CRISPR-Like Enzyme System
On September 23, Anthropic announced the formation of a life sciences research team and laboratory, alongside early results from Claude's involvement in biological research. Researchers reported that Claude, after analyzing over 200,000 reverse transcriptase (RT) sequences, discovered a previously undescribed enzyme system, which they named 'Array-associated Reverse Transcriptase' (ART). According to Anthropic, approximately 950 Claude agents used about 210 million tokens over 21 hours to screen 3,500 candidate systems, narrowing them down to 20. ART is primarily found in bacteriophages that infect bacteria and consists of a reverse transcriptase, adjacent auxiliary genes, and a set of regularly arranged DNA repeats that share features with CRISPR repeats. Preliminary experiments indicate these repeats are transcribed into multiple short RNAs, though their specific function requires further study. The research was experimentally validated by Anthropic scientists, and the company has released a preprint.
Anthropic CEO: Claude AI Discovers New Gene-Editing Mechanism at UN Security Council
Anthropic CEO Dario Amodei announced at the UN Security Council that the company's AI model, Claude, has discovered a new molecular machine that may constitute a novel gene-editing mechanism with potential applications in gene therapy. According to Amodei, Claude led the literature review, theorizing, and experimental design, while humans ran and verified the experiments in the lab. The finding is described as preliminary but could represent a significant advancement in gene editing technology.
Read sourceEveryone knows large language models can find patterns in text that humans might overlook. However, Claude didn’t just identify a novel biological system; it autonomously ran the discovery search itself for said system!!! Anthropic gave roughly
Anthropic deployed approximately 950 Claude agents with a single high-level prompt to autonomously search for novel biological systems over 21 hours. The agents scanned more than 200,000 reverse transcriptases, narrowing the field to 3,500 candidate systems and then 20 finalists. One agent identified a CRISPR-like repeat architecture, leading to a previously uncharacterized enzyme system that Anthropic is now calling ART. The post claims this represents a significant acceleration of scientific discovery, as the AI not only identified patterns but independently conducted the entire discovery process without human step-by-step guidance.
Read sourceDario Amodei still believes AI could “cure most diseases in 5-10 years,” and Anthropic’s latest discovery offers an early glimpse of the research he thinks could get us there. He calls that goal "barely possible", and
Anthropic CEO Dario Amodei maintains his belief that AI could 'cure most diseases in 5-10 years,' calling the goal 'barely possible' and contingent on applying AI to every stage of the research pipeline. The post highlights a recent discovery by Anthropic's Claude AI as an early glimpse of such potential. According to Anthropic, Claude identified a previously overlooked biological system in viruses that infect bacteria. After 21 hours of searching data using roughly 950 agents and 210 million tokens, an agent spotted a repeating pattern of DNA sequences next to a gene for an unusual reverse transcriptase. Initial lab work showed those repeats produce short RNA molecules, resembling CRISPR, though its function and gene-editing utility remain unknown. The post frames this as a proof of concept for AI driving biological discoveries, with the author noting that OpenAI and Anthropic were farsighted to focus on biology early.
Read sourceAnthropic's wet lab and 949 Claude agents discover new ART enzyme system in 21 hours
Anthropic announced that its new wet laboratory achieved its first success after 949 Claude AI agents autonomously tracked down a new enzyme system in just 21.5 hours. The agents searched 1.94 billion protein clusters and recovered 198,290 RT clusters, using 215.6 million tokens. During the search, Claude noticed an unusual pattern of a small DNA sequence repeating next to an enzyme, resembling CRISPR. Anthropic scientists investigated and identified a new family of biological systems they named ART. The system produces many small RNA molecules, indicating the repeated DNA sections serve a functional purpose rather than being random. The discovery highlights the potential of AI-driven autonomous research in biotechnology.
Read sourceAnthropic CEO announces Claude-led discovery of potential new gene editing mechanism
Dario Amodei, CEO of Anthropic, announced that their AI model Claude led the discovery of a molecular machine that could represent a new gene editing mechanism. The work was conducted with Anthropic's life sciences team, where Claude read through literature and genome data, identified something interesting, and proposed experiments that the team carried out. Amodei notes the precise function, biotechnological utility, and significance are not yet clear, but considers it PhD-level work. He draws parallels to AI's rapid progress in mathematics, predicting a similar exponential trend in biology. Amodei argues that while biology requires experimentation, humans can collaborate with AI to perform and validate experiments quickly. He frames this as the first step toward AI-driven biological discovery, potentially accelerating the entire biomedical pipeline from fundamental discoveries to drug development, though clinical trial times would remain unchanged. The discovery builds on prior systems like CRISPR and reverse transcriptase-based systems, with a similar but independently evolved system recently described by a Stanford team.
Anthropic's wet lab discovers new enzyme using 950 Claude agents in 21 hours
Anthropic announced that its new wet lab has made its first discovery, with nearly 950 Claude AI agents independently identifying a new enzyme after 21 hours of continuous work. According to The Verge, the agents consumed 210 million tokens during the process before one agent noticed an unusual repeating DNA pattern. Subsequent analysis and wet-lab testing confirmed what Anthropic described as 'a previously uncharacterized enzyme system found in bacteriophages.' The discovery marks a milestone in AI-driven scientific research, demonstrating the ability of large-scale autonomous agent systems to conduct complex biological investigations without direct human intervention.
Read sourceClaude discovers unknown enzyme system in bacteriophage DNA with CRISPR-like structure
Anthropic's AI model Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. The enzyme's gene is accompanied by a long array of repeating DNA, a structure resembling CRISPR. The function of this system is not yet understood, but only a few known systems share its features, and all of them are capable of cutting, copying, and pasting DNA. The discovery of such programmable systems has historically revolutionized medicine, with CRISPR serving as the foundation for genetic medicines. However, further research is needed to determine what this system does and whether it can be applied similarly. The finding was announced by AnthropicAI via a social media post linking to additional details.
Anthropic Says Claude AI Found Unknown Enzyme System in Phage DNA With CRISPR-Like Array
Anthropic announced that its AI model, Claude, has discovered a previously unknown enzyme system in bacteriophage DNA, which features a CRISPR-like repetitive DNA array adjacent to the enzyme gene. The function of this system remains unclear, but Anthropic noted that the few known similar systems are capable of cutting, copying, and pasting DNA. Historically, such programmable systems have led to foundational gene-editing tools like CRISPR. The company emphasized that further research is needed to verify the system's potential applications. The announcement was made via Anthropic's official X account, providing specific characteristics of the discovery while honestly acknowledging its unknown function, allowing readers to assess its distance from CRISPR-class discoveries.
Read sourceAnthropic's wet lab produces first discovery with 1,000 Claude agents in 21 hours
Anthropic's new wet lab has achieved its first discovery, with nearly 1,000 Claude AI agents autonomously uncovering a novel enzyme system after 21 hours of continuous work. The announcement, made via a post on Polymarket, marks a milestone in AI-driven scientific research, demonstrating the ability of large-scale agentic systems to conduct laboratory work and make biological discoveries without direct human intervention. The specific details of the enzyme system and its potential applications have not been disclosed in the initial announcement. This development highlights the growing capability of AI agents to perform complex, autonomous research tasks in wet lab environments, potentially accelerating the pace of scientific discovery in fields such as biotechnology and synthetic biology.
Anthropic Launches Life Sciences Lab; Claude Agent Autonomously Discovers Novel CRISPR-Like Enzyme System
Anthropic announced the establishment of a life sciences research group and laboratory. In a milestone for AI-driven scientific discovery, the company's Claude agent autonomously identified a novel enzyme system related to DNA repeat sequences and structurally similar to CRISPR, with only high-level human guidance. The system, named array-associated reverse transcriptases (ART), was detailed in a preprint published by the company. The announcement highlights how AI agents can participate in fundamental biological discovery, with the original report providing specific process data and scale of the discovery. This marks a significant step in applying large language models to autonomous scientific research.
Read source