Artificial intelligence (AI) is rapidly transforming the state of biotechnology, enabling genome editing tool development to become faster, smarter, and more iterative in both discovery and optimization processes. By integrating structural prediction, evolutionary modeling, and experimental feedback, AI enables rational design far beyond the limits of lab-based protein engineering approaches.

In 2023, we introduced a pioneering AI-driven structural clustering method for enzyme mining, which, for the first time, introduced the concept of using parallel high-throughput AI to expedite genome editor discovery. Building on this, we established a closed-loop AI-assisted platform for enzyme optimization, accelerating the evolution of genome editing tools. These innovations are pushing the boundaries of programmable biology, with broad implications for plant biotechnology, synthetic biology, and precision medicine.

Precise control of gene expression is essential for engineering complex traits in crops. Our lab develops programmable regulatory strategies at multiple levels—transcriptional, post-transcriptional, and translational—to achieve quantitative tuning of gene expression in plants.

At the translational level, we reprogram upstream open reading frames (uORFs) to fine-tune protein translation. At the post-transcriptional level, we modulate pre-mRNA alternative splicing by targeting conserved splice sites. These tools have enabled customized trait design in lettuce, strawberry, and rice, improving nutritional content, sugar accumulation, plant architecture, and more.

Our continuing efforts seek to advance CRISPR-based large-fragment engineering, RNA splicing control, translational fine-tuning, and nuclear architecture manipulation. By combining genome editing, directed evolution, single-cell multi-omics, and machine learning, we aim to decode regulatory networks and establish a framework for multiscale, programmable gene expression in plants.

The development of efficient delivery and regeneration systems remains a critical bottleneck for broad applications of plant genome editing. Our lab develops innovative systems to overcome genotype dependence, long regeneration cycles, and species recalcitrance during these editing processes.

We have established robust delivery platforms for wheat, including the TaGRF4–TaGIF1-based co-transformation approach and BSMV-sgRNA viral vectors, enabling heritable editing across diverse wheat cultivars. In parallel, we developed a DNA-free RNP and mRNA delivery method that meets biosafety requirements. By integrating single-cell transcriptomics, metabolomics, and proteomics, we have identified key regeneration factors that enable transformation and plant regeneration in recalcitrant species such as indica rice, tetraploid wild rice, and octoploid strawberry.

Moving forward, we aim to optimize conventional transformation methods such as Agrobacterium-mediated delivery and biolistics, and to extend DNA-free editing to horticultural crops and forest trees. By combining AI-assisted regulatory network modeling, directed evolution of regeneration factors, and high-throughput screening, we are building a programmable framework for efficient and genotype-independent plant regeneration.

Biotechnology is fundamentally changing how we engineer crop traits, enabling the development of novel, high-value cultivars for food, health, and sustainability. Our lab is driving this transformation by integrating genome editing, multi-omics, and computational design into trait development and breeding pipelines.

We have pioneered approaches such as in planta saturation mutagenesis, de novo domestication of wild species, and precision engineering of complex traits—including targeted editing in polyploid crops. These approaches have enabled the generation of plants with improved disease resistance, enhanced nutrition, and consumer-desired traits such as extended shelf life and flavor profile changes.

Looking ahead, we are integrating artificial intelligence, high-throughput functional screening, and programmable editing platforms to build intelligent crop design pipelines. These efforts are bridging the gap between fundamental discoveries and real-world agricultural deployment, supporting the development of resilient, high-performance crops for future food systems.