Scientists are using artificial intelligence to unlock hidden genetic traits in wheat collections that could transform crop breeding and protect global food supplies. Researchers from the UK-CGIAR Centre have developed a new approach to make wheat genetic databases accessible to AI systems, enabling faster discovery of valuable traits like disease resistance. The wheat genome is nearly six times larger than the human genome, and while sequencing technology has advanced dramatically, the real challenge is analyzing the massive amounts of genetic data now being generated faster than researchers can act on it.

Why Is Wheat Genetic Data So Hard for AI to Use?

The problem sounds simple but has stymied researchers for years: wheat genetic information exists in databases, but it is not optimized for machine learning models to analyze. Genetic sequences are stored in formats that are difficult to compare across different wheat varieties, especially when looking at wild relatives or heritage plants that are genetically distant from modern crops. The Watkins Collection, a heritage wheat repository containing over 800 lines of wheat plants, holds a treasure trove of genetic diversity and unknown traits that could improve modern varieties, but extracting useful information from this data has been nearly impossible until now.

The research team from the International Maize and Wheat Improvement Center (CIMMYT), the John Innes Centre (JIC), and other institutions developed a solution using a technique called k-mers, which are short substrings of DNA sequences. Unlike traditional approaches that rely on comparing all sequences to a single reference genome, k-mers allow direct comparisons across genetically diverse plants without needing a reference point. This makes the system more scalable, requires less computing power, and works better with wild relatives and heritage varieties.

How Are Researchers Using This New Approach?

• Disease Resistance Discovery: The k-mer approach has already identified novel yellow rust resistance genes in the Watkins Collection that do not exist in CIMMYT's modern wheat varieties, providing new breeding targets for farmers worldwide.
• Landrace Core Collection: Researchers used k-mers to develop a new landrace core collection enriched to identify valuable genetics in wild relatives, initially evaluated for resistance against fungal diseases like wheat yellow rust.
• Automated Breeding Workflows: The AI-readiness framework is being applied to other wheat breeding targets beyond disease resistance, with the potential to accelerate trait discovery and bring new varieties to farmers faster.

Yellow rust, caused by the fungus Puccinia striiformis, is a major threat to wheat harvests globally. Like new flu strains, the rust fungus continues to evolve and overcome previously resistant varieties. Growing genetically resistant wheat is the primary defense farmers have against the disease, making the discovery of new resistance genes critical for protecting future harvests.

"Wheat research is advancing at an incredible pace and AI offers further potential here. New AI-driven automated workflows will help further streamline the translation of complex sequence information into precise, field-ready breeding decisions," said Dr. Susanne Dreisigacker, a researcher at the International Maize and Wheat Improvement Center.

Dr. Susanne Dreisigacker, International Maize and Wheat Improvement Center (CIMMYT)

What Are the Practical Implications for Wheat Breeders?

The new AI-readiness approach has immediate practical benefits for wheat breeders and researchers. By combining the latest laboratory techniques with AI, researchers can now identify valuable genetics in large sequenced collections much faster than before. The k-mer system makes genetic data more accessible and useful for researchers and breeders working to develop better varieties. This same pipeline is already being applied to other wheat breeding targets beyond disease resistance, suggesting the approach could accelerate improvements in yield, nutrition, resilience, and sustainability.

"Identifying these new rust resistance genes is just one small example of the huge potential within sequenced collections. By combining the latest lab approaches with AI, we will see many more such discoveries and a much faster way to bring these to the field," noted Fernando Galvan from CIMMYT.

Fernando Galvan, International Maize and Wheat Improvement Center (CIMMYT)

The challenge now is not generating genetic data, but storing, transferring, and analyzing it effectively. Genomic information is now cheap and easy to generate for wheat across large genebank collections, but computer constraints have made analysis difficult. AI tools could leverage this vast amount of data for better predictions to accelerate genetic gains and breeder applications, but this requires investment in diverse use cases and computer infrastructure.

"Genomic information is now easy and cheap to generate, including for wheat across large genebank collections. The big challenge, however, is to store, transfer and analyse this data due to computer constraints. AI tools could leverage this vast amount of data for better predictions to accelerate genetic gains and breeder applications. We now need to invest in a diversity of use cases of deployment and computer infrastructure to take full advantage of the potential of AI-enhanced crop breeding," explained Dr. Jesus Quiroz-Chavez of JIC-CIMMYT.

Dr. Jesus Quiroz-Chavez, John Innes Centre-CIMMYT

This work represents a significant step forward in using AI to address long-standing bottlenecks in gene discovery and functional characterization. By making wheat genetic databases AI-ready, researchers can now identify key traits and accelerate their incorporation into improved wheat varieties. The implications extend beyond wheat to other crops and could help ensure food security in an era of climate change and evolving crop diseases.