Researchers use an AI-designed gene editor to make rice resist bacterial blight
The adapted OpenCRISPR-1 tool produced inherited edits and resistance to three bacterial strains. Moving the results into commercial rice remains unfinished.
Loading page…
The adapted OpenCRISPR-1 tool produced inherited edits and resistance to three bacterial strains. Moving the results into commercial rice remains unfinished.
Listen to this story
University of Missouri researchers adapted Profluent’s OpenCRISPR-1 for rice and used it to alter SWEET genes exploited by bacterial blight, producing resistant plants with inherited edits. The result shows an AI-generated editor can be engineered for a crop, and its open terms could make research and commercialization easier. Still, the reported 97% editing rate came from a research variety; commercial deployment requires further testing, regulatory steps and industry partners.
In 12-day tests, edited plants resisted three Xoo strains, but they were not commercially grown varieties.
The maximum reported editing success rate was 97% in a rice variety commonly used for research.
For rice, researchers rewrote the DNA instructions for OpenCas9 and added signals to direct the editor to the plant’s DNA.
An AI-designed gene editor has helped laboratory rice plants resist bacterial blight, with genetic changes passed to the next generation. SingularityHub’s November 18 account details how University of Missouri researchers adapted the tool for rice, achieving editing success rates of up to 97% in a research variety—not a crop ready for farmers.
The findings come from a study by Ajay Gupta and colleagues published in aBIOTECH and described by the University of Missouri on October 27. Led by professor Bing Yang, the team used OpenCRISPR-1, a gene-editing system designed by biotechnology company Profluent using artificial intelligence.
The biological target was already understood. The blight-causing bacterium known as Xoo enters rice leaves through wounds and spreads through the plant’s water-carrying vessels. It then injects molecular tools into nearby cells, increasing activity in a family of genes called SWEET. Those genes make proteins that move sugar into spaces where bacteria can feed.
Yang’s team had pursued that vulnerability before. In 2019, it used CRISPR—a tool for making targeted DNA changes—to alter switches controlling three SWEET genes in two commercially cultivated rice varieties. In confined field tests, those plants grew normally and resisted multiple bacterial strains. That earlier work established a disease-resistance strategy; the current study investigates a different editing tool.
Access helped motivate the switch. Yang said proprietary gene editors can require commercial licensing agreements. He described OpenCRISPR-1 as open source and free for research and commercialization of crop genome editing, giving researchers and companies more flexibility to pursue useful traits.
Profluent’s AI learned patterns from millions of microbial CRISPR systems to generate new editors. OpenCRISPR-1 had already worked in animal cells, but adapting it to rice required further engineering. The Missouri researchers changed the DNA instructions used to produce its cutting protein, OpenCas9, without changing the protein’s structure.
That rewrite used DNA instructions better suited to rice cells, helping them manufacture the protein more efficiently. The team also added molecular signals to move the editor into the compartment containing the plant’s DNA. These changes addressed production and delivery inside the cell, rather than choosing a different disease target.
The editor also needed guide RNA: a molecule with one part that attaches to the cutting protein and another that directs it to a DNA target. Researchers paired an AI-designed RNA backbone with guides aimed at three SWEET genes exploited by Xoo. Despite substantial differences from the conventional system, the adapted editor worked comparably well in the research rice variety.
The study also explored prime editing, a method that rewrites DNA without cutting both strands. Researchers modified OpenCas9 for that task and edited four genes in isolated rice cells. The changes included single DNA-letter substitutions and three simultaneous letter swaps. This part of the work was a cell experiment, distinct from the disease-resistance tests in plants.
In some prime-editing tests, OpenCas9 produced fewer unintended changes away from the target than a conventional prime editor. That result was limited to some comparisons; it does not establish that the adapted tool is more precise across every editing task.
For farmers, the next step remains transferring the findings from laboratory rice into commercially grown varieties. The Missouri team says additional testing, regulatory steps and industry partners are still needed before resulting varieties can reach fields.
Loading discussion...
Join the conversation
Explain what would matter most to your confidence.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.