PubMed · 2026-05-11
Researchers have mapped out a new AI-powered framework to dramatically improve sugarcane breeding by combining genetic data, environmental data, and machine learning — potentially accelerating how quickly breeders can develop better-performing varieties.
A 'three-model' computational framework was designed to simultaneously decode sugarcane's complex polyploid genetics, map fine-scale environmental conditions, and predict how individual plant clones will perform in the field.
The proposed 'isoenvironment' design allows breeders to group field sites by environmental similarity, enabling more precise prediction of genotype-by-environment interactions that have long confounded conventional breeding models.
AI extensions to the framework are designed to exploit sugarcane's clonal propagation biology and perennial ratoon growth cycles — two biological traits that standard genomic prediction tools were not built to handle.
