AI analysis of enzymes reveals CRISPR-like system in phages
Anthropic's AI model Claude analyzed over 200,000 enzyme sequences and identified a previously unknown enzyme system in bacteriophages that resembles CRISPR. This discovery could lead to new gene-editing tools.
In a striking convergence of artificial intelligence and molecular biology, a large language model was tasked with sifting through more than 200,000 enzyme sequences. The analysis surfaced a distinct enzyme system within bacteriophages—viruses that infect bacteria—bearing structural and functional similarities to CRISPR, the renowned bacterial defense mechanism that underpins modern gene editing.
The finding is notable because it emerged from computational pattern recognition rather than traditional laboratory screening, underscoring AI's growing role in biological discovery. While the work is preliminary, the identification of a CRISPR-like system in phages suggests that nature may harbor additional, unexplored molecular machinery. Such systems, if validated, could expand the toolkit available for precise genetic manipulation, offering researchers new avenues beyond established CRISPR-based methods.
This discovery could reshape how scientists approach gene editing, potentially providing alternative tools that may complement or refine existing CRISPR techniques. Researchers and biotechnology firms may benefit from a broader molecular palette, while patients could eventually see advances in therapies for genetic disorders. However, the path from computational prediction to clinical application is long, and the practical impact will depend on experimental validation. Society may also see accelerated interest in AI-driven scientific discovery, raising questions about how such tools are integrated into research pipelines.