The researchers developed a multidimensional pan-genome graph from nine genome assemblies of four related Saccharum species with different ploidy levels, including modern varieties and their founder species.[1] This graph includes 47 to 57 haplotypes and approximately 74 to 271 thousand allelic genes.[1] It captures about 82% of sugarcane genomic diversity, compared to only 34% using a single reference genome.[1] The graph enables analysis at the genome, gene, or protein level, supports variant discovery, cross-ploidy comparisons, and multiomic studies.[1] It distinguished haplotypes and allele dosage, improved variant discovery, and increased explained phenotypic variability.[1] It supports marker design, allele search and more accurate genomic predictions in polyploid crops.[1] In genome-wide association with dosage, it improved heritability estimates and identified loci associated with sugar or leaf angle, including the SaIRX10 and SaBAK5 genes.[1] The framework serves as a basis for graph genetic studies in sugarcane and other polyploid genomes.[1]