NOTICE

At the authors' request, the data files for this submission have been removed from public view due to concerns regarding compliance with data usage agreements. (June 2024)

Vegetable soybean (Glycine max L.) stands as a significant crop in East Asia, distinguished by unique phenotypic traits such as seed size, coat color, and sweetness, setting it apart from grain soybeans. These traits are likely attributed to intensive, extensive, and long-term selection processes, yet the genomic alterations underlying these variations remain inadequately elucidated. Here, we established a core collection panel comprising 52 vegetable soybean and 192 grain soybean samples, selected from a pool of 2,394 Taiwanese accessions. Through population structure, genetic diversity, and differentiation analyses, we observed reduced genetic variability in vegetable soybeans compared to grain soybeans, with significant differentiation between the two types. Our investigation identified 159 significant signatures within 67 putative regions linked to selective breeding in vegetable soybeans, many of which overlap with previously identified genomic loci, indicating selective breeding footprints. Phenotypic variation patterns suggest that improved productivity in vegetable soybeans may be achieved by pyramiding specific favorable alleles. Furthermore, we pinpointed two promising genes, GmAP2-2 and GmTFL1b (Dt1), as potential focal points for future vegetable soybean breeding, given their roles in flowering time and seed development. This comprehensive study underscores the utility of core collection panels and provides valuable insights into soybean genetics, laying the groundwork for future breeding endeavors and research in vegetable soybeans.

A total of 2,618 Taiwanese soybean accessions from the National Plant Genetic Resources Center at the Taiwan Agricultural Research Institute were classified into 2,511 grain soybeans and 107 vegetable soybeans. DNA extraction and purification from 2,618 soybean accessions utilized the DNeasy 96 Plant Kit (QIAGEN) system, ensuring an A₂₆₀/A₂₈₀ ratio of 1.8-2.0 and concentration ≥ 60 ng/μl. Genotyping employed the Axiom^® SoyaSNP180K (180,961) chip array. The criteria, including a batch’s quality (Dish quality control > 0.82) and sample call rate (> 97%), led to the exclusion of 33 redundant and 191 poor-quality samples. Filtering retained single nucleotide polymorphisms (SNPs) with missing rate < 0.10, excluding 366 unanchored scaffold SNPs and chloroplast genome SNPs. The quality-filtered SNP dataset comprised 147,560 SNPs, with a call rate of 99.78%. The core collection of Taiwanese vegetable and grain soybeans was constructed to minimize redundancy within germplasm pools. Principal coordinate analysis (PCoA) was executed on the quality-filtered 147k SNP dataset using Tassel v.5.0 software, enabling the identification of principal components (PCs) capturing 80% of the total genetic variation. Subsequently, the PowerCore program v.1.0 systematically facilitated the selection of core accessions based on phenotypic traits and the PCs derived from genotypic data, resulting in a core collection comprising 192 accessions for grain soybeans and 52 accessions for vegetable soybeans. Further SNP data refinement using the command --maf 0.05 --hardy on PLINK v.1.9 filtered out SNPs with rare variants (i.e., minor allele frequency (MAF) < 0.05) and heterozygous rate < 0.10, resulting in a final dataset of 78,189 SNPs.