Anthropic has announced that its Claude model aided in the discovery of a novel enzyme system with properties reminiscent of mechanisms in gene-editing technology CRISPR, marking the first result from artificial intelligence firm’s biology research efforts. The findings come days after Reuters reported Anthropic had quietly developed a wet lab in the San Francisco Bay Area, as the firm expands its ambitions in life sciences and drug-related research beyond purely computer-based work.
Anthropic also stated that Claude analysed large DNA databases and witnessed an unusual system built around a reverse transcriptase (RT), an enzyme that copies RNA into DNA. As of now, Anthropic has named the system array-associated reverse transcriptases, or ART. ART consists of repeating DNA sequences that resemble patterns seen in CRISPR. The firm also said that while the underlying RT has been identified in past studies, Claude appeared to be the first to recognize key features of the broader system, comprising an array of non-coding DNA sequences and an additional protein of unknown function.
This single step by Anthropic could be the founding stone of AI moving from research to discovery. According to Anthropic, 950 Claude agents searched more than 200,000 reverse transcriptases in a span of 21 hours, processing around 210 million tokens. They generated about 3,500 candidates and narrowed down the search to 20 for highly detailed investigation. If we go by the old-school methods, genome mining needs scientists to manually search huge biological databases, identify unusual patterns and formulate hypotheses. AI can potentially compress that process on a considerable level.
And the biggest upside of the same is that ART can become the new gene-editing platform. It has reverse transcriptase bundled with a long array of repeated DNA sequences. This arrangement is identical to CRISPR-like systems, and Anthropic’s early experiment is a hit that the repeat array produces multiple short RNAs as well. The application of the same could be seen in future in cellular engineering, drug discovery, molecular diagnostics, synthetic biology, genetic disease research, and a lot more.
