The 100-pod fresh weight (PFW), 100-seed fresh weight (SFW), 100-seed dry weight (SDW) and moisture content of fresh seeds (MCFS) at the R6 stage are crucial factors for vegetable soybean yield. However, the genetic basis of yield at the R6 stage remains largely ambiguous in soybean. To better understand the molecular mechanism underlying yield, we investigated four yield-related traits of 133 soybean landraces in two consecutive years and conducted a genome-wide association study (GWAS) using 82,187 single nucleotide polymorphisms (SNPs). The GWAS results revealed a total of 14, 15, 63 and 48 SNPs for PFW, SFW, SDW and MCFS, respectively. Among these markers, 35 SNPs were repeatedly identified in all evaluated environments (2015, 2016, and the average across the two years), and most co-localized with yield-related QTLs identified in previous studies. AX-90496773 and AX-90460290 were large-effect markers for PFW and MCFS, respectively. The two markers were stably identified in all environments and tagged to linkage disequilibrium (LD) blocks. Six potential candidate genes were predicted in LD blocks; five of them showed significantly different expression levels between the extreme materials with large PFW or MCFS variation at the seed development stage. Therefore, the five genes Glyma.16g018200, Glyma.16g018300, Glyma.05g243400, Glyma.05g244100 and Glyma.05g245300 were regarded as candidate genes associated with PFW and MCFS. These results provide useful information for the development of functional markers and exploration of candidate genes in vegetable soybean high-yield breeding programs.

Grain size is a quantitative trait controlled by multiple genes. Though many genes regulating grain size have been reported in rice, their superior haplotypes or haplotype combinations for developing elite varieties remain elusive. In this study, we found that haplotypes of 35 previously functionally characterized genes governing grain yield significantly varied in the 3000 rice genome project (3K RGP) panel. Twenty-four of these genes had 2–6 haplotypes based on non-synonymous single nucleotide polymorphisms (SNPs), while the remaining 11 genes had only one haplotype. A total of 75 haplotypes in the 24 genes were identified. Also, we conducted an association analysis between the 75 haplotypes and 100-grain weight, and found that haplotype SG1c had the highest grain weight, followed by GW8e and DEP1e, while GW8f and GLW7d had the lowest. Furthermore, OsBAK1, D2, TGW6, BG1, SRS3, GL7, GW8 and GW6a haplotypes could distinguish japonica and indica, and thus were associated with rice domestication. Haplotype combination analysis showed that four haplotypes including MAPK6b, BAK1b, GL7d and SRS3b were the superior haplotype combination, and explained a possible genetic basis for accession superiority with the highest grain weight. Haplotype identification of genes controlling grain size variation will provide the theoretical basis for developing elite rice varieties with superior haplotypes or haplotype combinations of target genes and provide valuable molecular breeding targets.