In a breakthrough that could redefine crop biotechnology, scientists at the Indian Council of Agricultural Research’s Central Rice Research Institute have developed and experimentally validated the world’s first AI-designed genome-editing tool for plants, marking a major advance beyond the existing CRISPR systems that depend on naturally occurring microbial proteins.

Gene editing works like a molecular scalpel, allowing scientists to precisely cut and rewrite sections of a plant’s DNA to develop traits such as higher yield, climate resilience or disease resistance without introducing foreign genes. Until now, these tools relied on proteins borrowed from bacteria and other microbes found in nature. The Indian team’s breakthrough lies in showing that artificial intelligence can design entirely new enzymes from scratch that work efficiently inside plants. This could open the door to tailor-made editing systems that could be cheaper, more versatile and less constrained by global CRISPR patents.

Led by scientist Kutubuddin Ali Molla, the team showed that AI-designed enzymes can accurately edit plant DNA, enabling gene knockout, base editing and prime editing in crops. While a US company had earlier developed a similar AI-designed system for human cells, this is the first successful demonstration in plants.

Speaking to businessline, Molla said the research has been accepted for publication in New Phytologist, a reputed international plant science journal. “Unlike current genome-editing platforms that mostly depend on naturally occurring microbial proteins, our invention shows that computationally designed enzymes can also support robust gene knockout, base editing, and prime editing in crops. The platform may open new avenues for building customized genome-editing systems for agriculture and could also help address some intellectual property limitations associated with existing CRISPR technologies,” he said.

The new platform, named Plant-OpenCRISPR1 (POC1), builds on OpenCRISPR-1 (OC1), an AI-generated nuclease developed for human cells. Scientists say the advance is significant because genome editing has so far depended on a narrow pool of bacterial enzymes such as Cas9 and Cas12a, which often face limitations in plant systems.

AI-driven protein designing is now transforming nuclease engineering by leveraging large language models trained on the vast natural diversity of protein sequences, he added.

“We have developed and validated a suite of OC1-based genome-editing tools for plants, taking rice as a model, that support highly efficient gene knockout, base editing, and prime editing, demonstrating the versatility of AI-designed nucleases in plant systems,” Molla said.

Like OC1, POC1 will be open for academic and commercial use. Molla said AI-driven protein design, powered by large language models trained on vast protein databases, is now transforming next-generation genome engineering and could have major implications for biotechnology and food security.

The research team included Priya Das, Romio Saha, Debasmita Panda, Chandana Ghosh, S P Avinash, Sonali Panda and Mirza J Baig from CRRI, Cuttack.

Published on May 18, 2026