Seed protein and oil content are the two important traits determining the quality and value of soybean. Development of improved cultivars requires detailed understanding of the genetic basis underlying the trait of interest. However, it is prerequisite to have a high-density linkage map for precisely mapping genomic regions, and therefore the present study used high-density genetic map containing 2267 recombination bin markers distributed on 20 chromosomes and spanned 2453.79 cM with an average distance of 1.08 cM between markers using restriction-site-associated DNA sequencing (RAD-seq) approach. A recombinant inbred line (RIL) population of 104 lines derived from a cross between Linhefenqingdou and Meng 8206 cultivars was evaluated in six different environments to identify main- and epistatic-effect quantitative trait loci (QTLs)as well as their interaction with environments. A total of 44 main-effect QTLs for protein and oil content were found to be distributed on 17 chromosomes, and 15 novel QTL were identified for the first time. Out of these QTLs, four were major and stable QTLs, viz., qPro-7-1, qOil-8-3, qOil-10-2 and qOil-10-4, detected in at least two environments plus combined environment with R² values >10%. Within the physical intervals of these four QTLs, 111 candidate genes were screened for their direct or indirect involvement in seed protein and oil biosynthesis/metabolism processes based on gene ontology and annotation information. Based on RNA sequencing (RNA-seq) data analysis, 15 of the 111 genes were highly expressed during seed development stage and root nodules that might be considered as the potential candidate genes. Seven QTLs associated with protein and oil content exhibited significant additive and additive × environment interaction effects, and environment-independent QTLs revealed higher additive effects. Moreover, three digenic epistatic QTLs pairs were identified, and no main-effect QTLs showed epistasis. In conclusion, the use of a high-density map identified closely linked flanking markers, provided better understanding of genetic architecture and candidate gene information, and revealed the scope available for improvement of soybean quality through marker assisted selection (MAS).

Soybean is a globally important legume crop that provides a primary source of high-quality vegetable protein and oil. Seed protein and oil content are two valuable quality traits controlled by multiple genes in soybean. In this study, the restricted two-stage multi-locus genome-wide association analysis (RTM-GWAS) procedure was performed to dissect the genetic architecture of seed protein and oil content in a diverse panel of 279 soybean accessions from the Yangtze and Huaihe River Valleys in China. We identified 26 quantitative trait loci (QTLs) for seed protein content and 23 for seed oil content, including five associated with both traits. Among these, 39 QTLs corresponded to previously reported QTLs, whereas 10 loci were novel. As reported previously, the QTL on chromosome 20 was associated with both seed protein and oil content. This QTL exhibited opposing effects on these traits and contributed the most to phenotype variation. From the detected QTLs, 55 and 51 candidate genes were identified for seed protein and oil content, respectively. Among these genes, eight may be promising candidate genes for improving soybean nutritional quality. These results will facilitate marker-assisted selective breeding for soybean protein and oil content traits.

Deciphering the genetic basis underlying agronomic traits is of importance for soybean improvement. However, covariation, modulated by genetic correlations between complicated traits via hub QTLs, commonly affects the efficiency and accuracy of soybean improvement. The goals of soybean improvement have nearly all focused on agronomic traits, including yield, plant type traits, and seed-related traits especially. To decipher the hub QTLs of yield, plant type, and seed, nine pertinent traits of an F₂ population (181 plants) derived from a cross between KeXin No.03 and JiDou 17, which were different in multiple traits such as plant height, seed protein, and 100-seed weight, were investigated with a high-density genetic map covering 2708.63 cM. A highly significant negative phenotypic correlation (−0.95) was found between seed protein (Pro) and seed oil (Oil). A total of 35 final QTLs after combining the ones closely linked physically were identified for eight traits explaining from 0.10% to 24.63% of the phenotypic variance explained (PVE) using composite interval mapping (CIM) and inclusive composite interval mapping (ICIM) procedures, and 13 QTLs were novel genes. A genomic region on chromosome 14 (qPro14, qOil14.2, and qSw14) was associated with three seed-related traits based on the relationship within and among the three trait modules. In addition, four genomic regions were detected as hub QTLs which linked to the seed-related module and plant-type model, including the E loci (E1 and E2). From the QTL results, 31 candidate genes were annotated, including the verified genes E1, E2, and QNE1, and they were grouped into three categories of biological processes. These results illustrate the genetic architecture as correlations among various soybean traits, and the hub QTLs should provide insights into the genetic improvement of complex traits in soybean.