In a major step toward securing global food supplies and advancing sustainable agriculture, a team of scientists has proposed an integrated framework that combines biotechnology and artificial intelligence (AI) to revolutionize crop breeding.
Published in Nature on July 24, the review was co-corresponding authored by Prof. GAO Caixia from the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences, Prof. LI Guotian of Huazhong Agricultural University, with contributions from additional co-authors including international collaborators.
Facing a rapidly growing global population, intensifying climate change, and shrinking arable land area, ensuring food security and sustainable agriculture is now one of the world's most pressing challenges.
This review explores the integration of multi-omics, genome editing, protein design, and high-throughput phenotyping (HTP) as part of a comprehensive vision for "AI + BT" (biotechnology) integration to genetically improve crops. The authors present a forward-looking framework for AI-assisted crop germplasm design, offering a clear roadmap for the future of sustainable agriculture.
The authors first emphasize the foundational role of modern omics technologies in creating a paradigm shift in crop breeding. They show that advances in genomics, metabolomics, and single-cell omics offer unprecedented insights into the genetic and biological mechanisms influencing crop traits, while revealing precise new targets for trait improvement. They also note that HTP technologies—which leverage drones, sensors, and automation platforms—enable rapid and accurate phenotypic assessments crucial for linking genotypes to phenotypes and identifying valuable genetic variations.
The review also spotlights powerful tools for crop improvement. For example, CRISPR-based genome editing enables efficient and precise genome modification, greatly reducing breeding cycles and enabling the rapid creation and stacking of desirable traits. Meanwhile, AI-driven protein design is emerging as a transformative technology through its capacity to design de novo functional proteins not found in nature. This approach facilitates the development of novel disease-resistance proteins, real-time biosensors for crop monitoring, and custom enzymes for environmental cleanup, thereby endowing crops with transformative traits.
The review particularly focuses on the proposal for an integrative "AI-assisted crop design" model that would use AI to analyze multimodal big data from genomes, phenotypes, environments, and agricultural practices. Breeders would define specific goals—such as increasing yield, enhancing stress tolerance, or improving nutritional quality—while AI would generate optimized, technically actionable breeding strategies through deep learning and knowledge inference. This data-driven approach marks a shift from experience-based breeding to precision design.
The authors also address the challenges ahead. High-quality, standardized data is essential for training robust AI models, and new technologies must comply with biosafety regulations. Encouragingly, global regulatory frameworks for genome-edited crops are evolving toward more scientific and streamlined approaches, paving the way for broader adoption.
This review was supported by the Biological Breeding-National Science and Technology Major Project, the National Key Research and Development Program, the National Natural Science Foundation of China, the Ministry of Agriculture and Rural Affairs of China and the New Cornerstone Science Foundation.
