Dissecting adaptive traits with nested association mapping: Genetic architecture of inflorescence morphology in sorghum
In the cereal crop sorghum (Sorghum bicolor) inflorescence morphology variation underlies yield variation and confers adaptation across precipitation gradients, but its genetic basis is poorly understood. Here we characterized the genetic architecture of sorghum inflorescence morphology using a global nested association mapping (NAM) population (2200 recombinant inbred lines) and 198,000 phenotypic observations from multi-environment trials for four inflorescence morphology traits (upper branch length, lower branch length, rachis length, and rachis diameter). Trait correlations suggest that lower and upper branch length are under largely independent genetic control, while lower branch length and rachis diameter are pleiotropic. Joint linkage and genome-wide association mapping revealed an oligogenic architecture with 1-22 QTL per trait, each explaining 0.1%-5.0% of variation across the entire NAM population. Overall, there is a significant enrichment (2.4-fold) of QTL colocalizing with homologs of grass inflorescence genes, notably with orthologs of maize (Ramosa2) and rice (Aberrant Panicle Organization1, TAWAWA1) inflorescence regulators. In global georeferenced germplasm, allelic variation at the major inflorescence morphology QTL is significantly associated with precipitation gradients, consistent with a role for these QTL in adaptation to agroclimatic zones. The findings suggest that global inflorescence diversity in sorghum is largely controlled by oligogenic, epistatic, and pleiotropic variation in ancestral regulatory networks. This genotype-phenotype trait dissection in global germplasm provides a basis for genomics-enabled breeding of locally-adapted inflorescence morphology.