Design and Implementation of a Comprehensive Experimental Teaching Platform for Intelligent Breeding in Agricultural and Forestry Higher Education

Rapid advances in genomics and artificial intelligence require breeding courses to integrate quantitative-genetic theory with genome-scale analysis and breeding decisions. We developed a Plant Intelligent Breeding Teaching Platform using Python (version 3.12.7) and PyQt5 (version 5.15.11), integrating quality control, population-structure analysis, GWAS, genomic prediction, and cross-design simulation. The demonstration used 898 individuals of black poplar (Populus nigra L.), with nine-year diameter at breast height as the phenotype and 50,537 SNPs retained after quality control and LD pruning from approximately 30× whole-genome resequencing. The platform supports adjustable QC parameters, PCA, genomic relationship analysis, GEMMA-based linear mixed-model GWAS, and six genomic-prediction models evaluated by five-fold cross-validation. Pearson correlations ranged from 0.489 to 0.574, with RF performing best. Known female/male information and additive GBLUP were used for cross prediction. In 64 Biological Sciences students, self-reported mastery increased by 26.6–71.9 percentage points across five modules (exact McNemar tests, p < 0.001). The platform provides an integrated environment for teaching the workflow from genomic data to breeding decisions.