Leveraging AI and integrated genomic-enviromic prediction for intelligent sugarcane breeding

Area of Science:

• Plant genetics and breeding
• Genomic prediction
• Computational biology

Background:

• Traditional sugarcane breeding relies on phenotypic selection, which is limited by the crop's complex polyploid genome and genotype-by-environment interactions (G×E).
• Existing genomic prediction models struggle to address the unique biological constraints of sugarcane, hindering genetic gain.
• The Integrated Genomic-Environmic Prediction (iGEP) framework offers a potential solution but requires adaptation for clonal crops.

Purpose of the Study:

• To provide a comprehensive roadmap for implementing the iGEP framework in sugarcane breeding.
• To address the specific challenges posed by sugarcane's polyploid genome and G×E.
• To propose a tailored computational framework integrating genetic, environmental, and phenotypic data for improved prediction.

Main Methods:

• Development of a "Three-Model" computational framework (Genetic, Environmental, Phenotypic) to decode polyploid allelic dosage and quantify environmental drivers.
• Application of artificial intelligence (AI) and iGEP models to leverage clonal propagation and optimize multi-trait selection.
• Addressing perennial ratoon dynamics in predictive models.

Main Results:

• A systematic approach to implementing iGEP in sugarcane, considering its unique biological characteristics.
• A computational framework capable of decoding complex polyploid genetics and high-resolution environmental data.
• Strategies for extending AI and iGEP to enhance clonal selection and manage crop dynamics.

Conclusions:

• This work establishes a new paradigm for accelerating genetic gain in sugarcane through advanced predictive analytics.
• The proposed iGEP framework and AI integration offer a transformative path from data digitization to synthetic crop design.
• The strategy is transferable to other complex-genome species, offering broad implications for crop improvement.