Rice feeds billions of people, but it is constantly threatened by bacterial blight, one of the most destructive diseases affecting the crop. Researchers have now used a gene-editing tool designed by artificial intelligence to protect rice from the disease, according to a report from SingularityHub. The tool, called OpenCRISPR-1, is free to use, which could make advanced crop engineering available to many more scientists. CRISPR is a molecular system that can cut DNA at precise locations, letting scientists disable or modify genes.
Bacteria spread through water and enter rice leaves through wounds. Once inside, they switch on a family of plant genes called SWEET, which produce proteins that pump sugar into spaces where the bacteria feed. Turning off these genes starves the bacteria and protects the plant. Scientists knew the targets, but proprietary gene editors often require costly licensing agreements, which has slowed progress.
An editor dreamed up by AI
OpenCRISPR-1 was created by the California startup Profluent and released openly on GitHub. The company trained an AI model on millions of natural CRISPR systems found in microbes, and the model designed an entirely new gene editor unlike any seen in nature. Researchers adapted the tool for rice. They rewrote its DNA instructions using genetic vocabulary suited to the plant and added molecular hooks that guide the tool into the cell compartment where DNA is stored. They paired it with an AI-designed guide RNA, the component that steers the editor to its target, aimed at three SWEET genes.
The system edited rice with a success rate of up to 97 percent in a variety commonly used in research. The modified plants fought off three different bacterial strains during a 12-day infection, and they passed the changes to the next generation. The team also rebuilt the tool for prime editing, a gentler method that rewrites DNA without cutting both strands, and it precisely edited four genes in rice cells.
The work matters because it shows AI can produce practical gene editors that anyone can use, lowering the barrier created by expensive patents. Rice was only a proof of concept, and it remains to be seen whether the tool works in commercial varieties and real-world conditions. Still, the researchers say the approach could extend to other crops and enable more sophisticated plant engineering.