Genome‐wide association studies reveal environment‐dependent genetic associations for plant architecture in regional soybean germplasm

Abstract Regional soybean ( Glycine max (L.) Merr.) breeding has produced germplasm adapted to different growing seasons, but the transferability of genetic associations associated with plant architecture across populations remains unclear. We performed genome‐wide association studies (GWASs) for plant height and main stem node number in 520 soybean accessions from two Chinese germplasm pools: a short‐season, cool‐climate Bashang population (260 accessions from the northeast region) and a long‐season, warm‐climate Baxia population (260 accessions from the Huang‐Huai‐Hai region). Each was evaluated in its environment over two seasons. Population analysis showed strong differentiation (fixation index [ F ST ] = 0.29) and less diversity in Bashang ( H e = 0.0808) than Baxia ( H e = 0.1538), reflecting different breeding histories. Phenotypically, Baxia plants were taller (+94.7%) and had more nodes (+88.2%) than Bashang, and the differences were highly heritable ( H 2 = 0.77–0.97). GWAS using multiple methods identified non‐overlapping loci: Bashang for height involved three loci on chromosomes 4, 7, and 12 (logarithm of the odds [LOD] = 7.44–9.88, effects +1.68 to +2.06 cm), with the chromosome 12 locus showing high differentiation ( F ST = 0.498). Baxia loci included a major height locus on chromosome 4 (+6.65 cm, LOD = 7.90, 38 Mb from Bashang's Chr4 locus) and a stem node number locus on chromosome 19 (+1.30 nodes, LOD = 7.18), near maturity genes E3 / Dt1 (where Dt1 is determinate 1 gene). Functional analysis revealed that Bashang genes are involved in DNA metabolism, whereas Baxia genes are associated with cell development and hormone regulation, indicating distinct developmental pathways. Differing differentiation metrics across loci suggest various evolutionary processes, such as selection and gene‐environment interactions. These findings highlight that genetic links to plant structure are population‐specific, impacting marker transferability and the need for tailored breeding strategies for regional environments.