AI-Designed CRISPR Scissors Enter Rice Cells: Three Gene-Editing Modes Pass Testing

Rice gene editing has long relied on CRISPR proteins discovered in naturally occurring microorganisms. Researchers are now testing another approach: using artificial intelligence to design new molecular scissors based on patterns in protein sequences. A team at India’s Central Rice Research Institute adapted the AI-generated OpenCRISPR-1 into POC1, a platform suitable for plants, and demonstrated that it can function in rice cells.

Here, AI does not independently choose which rice traits to modify; rather, it contributes to the design of the nuclease. OpenCRISPR-1 is an RNA-guided protein 1,380 amino acids long, differing from typical SpCas9 by 403 amino acids. The researchers then adjusted its expression system to meet the needs of rice cells, creating Plant OpenCRISPR-1.

The team first tested the system’s ability to create DNA double-strand breaks at four gene sites in rice protoplasts. Data compiled by Plantae, a publication of the American Society of Plant Biologists, showed editing efficiencies ranging from 10.0% to 16.8%, with no significant difference compared with conventional SpCas9. The researchers also regenerated stable transgenic rice plants carrying the targeted deletions, showing that the effect was not limited to individual cells in short-term culture.

POC1 is not limited to cutting genes. The team separately coupled it with adenine and cytosine base-editing components, as well as prime-editing components, enabling the system to replace individual DNA letters or rewrite specified sequences according to an RNA template. Nature India reported that the cytosine base editor achieved higher efficiency than the corresponding Cas9 system at several target sites. However, performance varied across sites, so the findings cannot be used to conclude that POC1 is comprehensively superior to existing tools.

Another significance of the findings concerns access to the technology. Established CRISPR platforms involve complex intellectual-property rights and licensing arrangements. OpenCRISPR-1 was released as open source, and the research materials are planned to be made available through a plasmid repository, with a separate pathway for commercial licensing. This could make it easier for academic teams to test new editors, but “open source” does not mean that all commercial uses are exempt from licensing and regulatory review.

At this stage, the research demonstrates that an editing tool can operate at a limited number of rice gene sites; it has not yet produced a new variety validated in field trials with higher yield, disease resistance, or stress-tolerance traits. Efficiency across different rice lines, unintended genome-wide edits, trait stability, and inheritance over multiple generations still require systematic evaluation. Even if the technology matures, the classification, review, and market authorization of gene-edited crops will vary by country. POC1 is therefore more like a new tool that has passed laboratory cutting tests than a rice variety approaching the market.