Genome-wide association and selection studies for pod dehiscence resistance in the USDA soybean germplasm collection

JaeBuhm Chun; Sadal Hwang

doi:10.1371/journal.pone.0318815

Abstract

As a domestication trait, pod dehiscence has a pleiotropic effect on agronomic traits and significantly contributes to yield loss in soybean. Population studies are still required to comprehend the genetic basis of dehiscence and to develop pod dehiscence-resistant cultivars with the optimal haplotype, thereby improving soybean production. We collected data for one wild (Glycine soja) (G. soja) and four cultivated (Glycine max) (G. max) populations from the USDA database. The G. max populations were evaluated in multi-environment conditions and used for genome-wide association study (GWAS) and selection. GWAS captured 86 quantitative trait loci (QTLs). Seventy-four new QTLs were colocalized in two different G. max populations, and 12 QTLs were closely mapped with previously reported QTLs. Eight out of 86 QTLs were associated with the domestication of pod dehiscence. We implemented marker-assisted selection (MAS) and genomic selection (GS) approaches to select pod dehiscence-resistant accessions with the best haplotype and lowest genomic breeding value (GBV), respectively. While our findings could be utilized for biology, genetics, and plant breeding, selecting pod dehiscence-resistant cultivars with the optimal haplotype will need further studies to confirm additional QTLs and assess advanced GS models.

Citation: Chun J, Hwang S (2025) Genome-wide association and selection studies for pod dehiscence resistance in the USDA soybean germplasm collection. PLoS ONE 20(3): e0318815. https://doi.org/10.1371/journal.pone.0318815

Editor: Santosh Gudi, North Dakota State University, UNITED STATES OF AMERICA

Received: October 25, 2024; Accepted: January 21, 2025; Published: March 28, 2025

This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors declare no competing interests.

Introduction

Pod dehiscence is one of the key propagation steps in wild plant species [1]. As the pod matures and dries, a dehiscence zone forms in the suture between the lignified pod wall and replum, and dehiscence occurs due to pectin degradation and cell wall breakdown [2]. Pod dehiscence is a natural process that results in significant yield loss in crop production and impacts food security [3]. The regulation of pod dehiscence has evolved through artificial selection in many crop species, along with other domestication syndromes [4]. As one of the major crops in the USA, cultivated soybean (G. max L. Merr.) has been domesticated to be more resilient to pod dehiscence than its wild species (G. soja Sieb. and Zucc.) over several decades [5].

It is widely recognized that pod dehiscence in soybean is highly heritable. Some population studies measured pod dehiscence using the oven-dried method. The broad- and narrow-sense heritabilities for pod dehiscence ranged from 0.62 to 0.97 and 0.70 to 0.84, respectively, when pod dehiscence was assessed 7–20 days after maturity [6–7]. Other studies reported that the broad-sense heritability for pod dehiscence was over 0.8 when field evaluation was conducted 3–4 weeks after maturity [8–10]. These studies indicated that pod dehiscence variation was mainly attributed to genetic factor, regardless of pre- or post-harvest pod dehiscence [11].

Soybean has adapted to diverse climates and geographic conditions [12]. Some climate factors, such as precipitation and temperature, can significantly affect pod dehiscence. High temperature, rapid temperature change, high and continuous rainfall, and frequent transition between wet and dry conditions increased pod dehiscence [9]. The hot and humid weather in subtropical and tropical regions of Southeast Asia has posed challenges for soybean production due to pod dehiscence, as well as disease or insect infestation [13].

We now better understand the genetic loci controlling pod dehiscence in soybean. One restriction fragment length polymorphism (RFLP) marker, B122_1, was found on chromosome 16, and it explained 44% phenotypic variation in the Young × PI416937 recombinant inbred line (RIL) population [10]. One QTL was observed on chromosome 16 in the Toymusume × Hayahikari RIL population [14]. The QTL was located between two flanking simple sequence repeat (SSR) markers, Sat_366 and Sat_093. Four SSRs were identified on chromosomes 15 and 16 in the Tokei 780 × Hidaka 4 interspecific RIL population [15]. Eight SSRs were detected on chromosomes 2, 5, 9, 14, 16, 17, 18, 19, and 20 in the Keunolkong × Sinpaldalkong and Keunolkong × Iksan 10 RIL populations [16]. Fine mapping and genetics studies (e.g., gene expression analysis, immunocytochemistry, and complementation test) have confirmed two major QTLs, SHAT1-5 [17] and qPDH1 [18–19], on chromosome 16. From both QTLs, two genes, SHAT1-5 [17] and Pdh1 [20], were characterized and associated with the domestication of pod dehiscence. Both genes were involved in the lignification in the fiber cap cells of the ventral suture (SHAT1-5) and in the inner sclerenchyma of the pod wall (Pdh1). Excessive lignification enhanced pod dehiscence by reducing fiber in the suture and increasing twisting force on the pod wall. NST1A was identified on chromosome 7 [21]. It is orthologous to NST1 in Arabidopsis thaliana (A. thaliana) and homologous to SHAT1-5 in soybean. As a NAC domain transcription factor (TF), NST1A thickens the secondary cell wall.

Pod dehiscence is associated with major agronomic traits [9,22,23] and causes yield loss in soybean [3,24–26]. Developing pod dehiscence-resistant cultivars with the optimal haplotype through population studies is crucial for enhancing soybean production. Hence, our study aimed to 1) identify QTLs for pod dehiscence by GWAS, 2) detect QTLs for the domestication of pod dehiscence, and 3) select pod dehiscence-resistant accessions with the best haplotype and lowest GBV using MAS and GS.

Materials and methods

Public trait data

The pod dehiscence data of twenty-four G. max populations was collected in the germplasm resources information network (GRIN) (https://npgsweb.ars-grin.gov/gringlobal/descriptordetail?id=51012) (S1 Table). According to the GRIN, all field experiments were conducted using a 2- or 4-row plot over one or two years. At harvest, the two rows from the 2-row plot or the two center rows from the 4-row plot were assessed for pod dehiscence using a scale of 1 to 5 based on the percentage of open pods (e.g., 1 = 0%, 2 = 12.5%, 3 = 25%, 4 = 37.5%, and 5 ≥ 50%) [21,27–29]. Average values for pod dehiscence were only available for all accessions in 24 G. max populations, combining two replicates for one or two years. Using all trait data on a scale of 1 to 5 from 24 G. max populations, the 95% confidence interval (CI) of population standard deviation (δ) was estimated as follows: (n: total number of accessions of 24 G. max populations as a sample size, s²: sample variance of n accessions, χ²_α/2 and χ²_1-α/2: critical values with n – 1 degrees of freedom (df) in the chi-square (χ²) distribution, and α: 0.05 as type I error).

Four G. max populations, SOYBEAN.EVALUATION.1IL64 (1IL64), SOYBEAN.EVALUATION.1IL66 (1IL66), SOYBEAN.EVALUATION.MN0102 (MN0102), and SOYBEAN.EVALUATION.MS967 (MS967), were selected based on the upper limit value of the 95% CI of δ. Before subsequent data analyses, some accessions were excluded from each selected G. max population, considering the adaptation zones for soybean maturity groups (MGs) in the USA [30]. USDA reports described detailed information on field design, experiment, and evaluation in 1IL64 [27], 1IL66 [27], MN0102 [28], and MS967 [29]. 1IL64 and 1IL66 were planted in Illinois. MN0102 and MS967 were planted in Minnesota and Mississippi, respectively. Plots were 2 rows × 2.4 m long with 100 cm row spacing in 1IL64 and 1IL66. MN0102 and MS967 had 4 rows × 3.6 m long with 76 cm row spacing. Except for the accessions of 1IL64 (2 replicates per year), the accessions of 1IL66, MN0102, and MS967 were replicated once per year. Percentages of open pods in each G. max population were converted into arcsine-transformed values using the following equation: Y(x) = sin^-1(square root of x) (Y(x): a transformed value and x: a percentage value / 100) [18,31]. The transformed average values were used as trait data for performing GWAS and GS.

Public genotype data

The 50K single nucleotide polymorphism (SNP) data from the SoySNP50K iSelect BeadChip [32] has been developed to genotype the USDA soybean accessions. A high-quality dataset of 42,509 SNPs in HapMap format was obtained from SoyBase [33] (https://www.soybase.org/tools/snp50k/). Because 429 unanchored SNPs were excluded from the Williams 82 reference genome assembly V 2.0 (Wm82.a2.v1) [34], 42,080 SNPs remained for one G. soja and four G. max populations (1IL64, 1IL66, MN0102, and MS967). The G. soja population comprised 1,179 wild accessions (S2 Table) and was designated WS1179 in this study. Due to the lack of pod dehiscence data for wild accessions in the GRIN, WS1179 was used only for genotype data analyses.

After converting heterozygous genotypes into missing ones, SNPs with a high ratio of missing genotypes and monomorphism were processed in each population. The percentage of SNPs with missing genotypes exceeding 10% was 0.18%, 0.13%, 0.18%, 0.64%, and 4.58% in 1IL64, 1IL66, MN0102, MS967, and WS1179, respectively. These SNPs were removed to avoid potential biases in further data analyses. Additionally, 433 monomorphic SNPs with identical alleles across all five populations were discarded. A total of 39,509 SNPs were commonly present in all five populations. The linkage disequilibrium (LD) k-nearest neighbors algorithm [35] was employed to impute missing genotypes in TASSEL (V. 5.2.70) [36] with default parameter settings (number of SNPs in high LD = 30, number of neighbors (k) to use in imputation = 10, and maximum physical map distance between SNPs to search for LD = 10e⁶).

Climate data

From the National Centers for Environmental Information (NCEI) (https://www.ncei.noaa.gov/access/monitoring/products/), we collected statewide average temperature and precipitation data from September to November, covering the years from 1895 to 2020 in the USA. The fall season from September to November was selected because soybean accessions with a wide range of MGs were harvested during this period in the USA [30]. Climate data was used to compare the Midwest (12 states) and the South (16 states) as the major divisions for soybean production.

Putative domestication-related SNPs

Domestication-related SNPs were detected with three confirmation steps: 1) Tajima’s D [37], 2) χ² goodness of fit test [38], and 3) Fixation index (F_ST) [39].

1) Tajima’s D: Nucleotide diversity (π) [40], number of segregating sites (θ) [40], and Tajima’s D values in 1IL64, 1IL66, MN0102, and MS967 were estimated using TASSEL (V. 5.2.70). A Tajima’s |D| value greater than 2 in a G. max population strongly indicated that the G. max population evolved under a non-neutral process and had domestication regions [41].
2) χ² goodness of fit test: The test was performed using 39,509 common SNPs across all five populations. The observed frequencies of an allele from a SNP were calculated in WS1179, 1IL64, 1IL66, MN0102, and MS967. The expected frequencies of the allele from the SNP were calculated by dividing the sum of observed allele frequencies from WS1179 and a G. max population by 2. The χ² goodness of fit test (df = 1) evaluated the difference in allele frequencies at a SNP between WS1179 and a G. max population using the following equation: χ² = ∑ (observed allele frequency – expected allele frequency)² / expected allele frequency. By employing the Benjamini–Hochberg procedure [42], the false discovery rate-adjusted p-value (q-value) was used as a criterion to find significant SNPs in one pair of G. soja-G. max dataset. Four q-values were generated in four pairs of G. soja-G. max datasets. SNPs with significance at one of the four q-values were considered domestication regions.
3) F_ST: Like the χ² goodness of fit test, the test was conducted using 39,509 common SNPs across all five populations. The F_ST value for a SNP was calculated using the following equation: F_ST = s² / (T × (1 – T)) (s²: sample variance in the frequencies of an allele between two different populations and T: average frequency of the allele in the total population). As a measure of population differentiation due to genetic structure, the range of F_ST value was defined to indicate moderate differentiation (0.05–0.15), high differentiation (0.15–0.25), and substantial differentiation ( > 0.25) [43]. SNPs with F_ST values greater than 0.05 between WS1179 and a G. max population were considered domestication regions.

Gene ontology analysis

Domestication-related SNPs were used for the gene ontology (GO) analysis. SNPs located within soybean genes with the Glyma2.0 gene model were selected. SoyBase [33] (https://www.soybase.org/tools/analysis/go.html) and the GO database (http://www.geneontology.org/) [44] were used to search for GO terms.

Population structure, principal component, and linkage disequilibrium analyses

The K values of WS1179, 1IL64, 1IL66, MN0102, and MS967 were inferred by Structure (V. 2.3.4) [45]. In each population, 10K SNPs were used to reduce computation. The top 500 SNPs with the highest minor allele frequency (MAF) were selected and well-distributed on each chromosome (500 SNPs × 20 chromosomes = 10K SNPs). The genotype data of each population was converted into numerical data (AA = 1, TT = 2, GG = 3, and CC = 4). The length of the Burnin period was 10,000, and the number of Markov Chain Monte Carlo replications after the Burnin was 10,000. Correlated allele frequency [45] and admixture [46] were used as variables in the statistical model without prior population knowledge. It was assumed that the value of K ranged from 1 to 30. The statistical test was iterated 20 times for each K value. When a K value was input, all hyperparameters, allele frequency (λ), degree of admixture (α), and F_ST, were estimated by Bayesian estimation in each statistical test. Twenty log-likelihood values were obtained, and their sample standard deviation (s) was calculated. The estimation of ∆ K [47] was as follows: ∆K = L″(λ, α, F_ST|K) / s L(λ, α, F_ST|K) (L″(λ, α, F_ST|K): second-order rate of change in average log-likelihood values between K – 1 and K and s L(λ, α, F_ST|K): s of 20 log-likelihood values at K). Average log-likelihood value [48] and ΔK were used to determine the best K for each population. Additionally, the Q matrix of each G. max population was created for GWAS. A Q matrix provided information about how accessions were clustered in K.

Principal component analysis (PCA) was conducted by TASSEL (V. 5.2.70) using 39,509 common SNPs across all five populations. Before performing PCA, the genotype data of each population was converted into numerical data (homozygous major allele = 1 and homozygous minor allele = 0).

LD was assessed in each G. max population. Monomorphic SNPs were excluded from each G. max population. TASSEL (V. 5.2.70) calculated the squared correlation coefficient (r²) [49] values between any two SNPs within a 30kb window. A r² value greater than 0.8 indicated a strong association between alleles, signifying high LD.

Genome-wide association study

GWAS was performed to identify QTLs for pod dehiscence in each G. max population. The k-nearest neighbors algorithm [50] was employed to impute a few missing trait values in TASSEL (V. 5.2.70) with default parameter settings (number of neighbors (k) to use in imputation = 5 and distance measures for computing nearest neighbors = Euclidean method). SNPs with MAF less than 5% and monomorphism were filtered out. An identical by descent (IBD)-based kinship (k) matrix [51] was created in TASSEL (V. 5.2.70).

Trait, genotype, and kinship data files were uploaded to GAPIT (V. 3.0) [52] in R (V. 4.2.1) [53]. In the GAPIT() function, the HapMap genotype data of each G. max population was converted into numerical data (homozygous major allele = 2 and homozygous minor allele = 0). Using the “Model.selection = TRUE” argument, forward model selection with the Bayesian information criterion determined the optimal number of PCs and created a PC matrix. In addition, another type of marker-based k matrix [54] was generated. Six statistical GWAS models, general linear model [55], mixed linear model [56], SUPER [57], MLMM [58], FarmCPU [59], and BLINK [60], were run simultaneously. Except for BLINK, Q, PC, and k matrices were utilized as cofactors in five GWAS models. A Q matrix was used interchangeably with a PC matrix. Likewise, two IBD- and marker-based k matrices were also interchangeable. A q-value was used to declare the rejection of the null hypothesis (H₀: No QTL for a tested SNP).

QTL data mining and confirmation

We examined previously reported QTLs that were tightly linked to those identified in this study. The relationship between the Consensus 4.0 genetic map [61] and the physical map [34] indicated that 5 cM on the Consensus 4.0 genetic map was equivalent to 2.2 Mb on the physical map. A distance of 5 cM or less was defined as a region of tight linkage. Therefore, QTLs for pod dehiscence identified in former studies were collected within a 2.2 Mb window of the QTLs discovered in this study.

We gathered data on the verified gene locations from QTLs and the CIs of QTL positions. In this study, the p-values for SNPs tested in each GWAS model were converted into logarithm of odds (LOD) scores using the following equation: LOD score = log₁₀(p-value^-1). The 95% CIs of QTL positions were calculated using the 1.5-LOD support interval [62–63]. On the other hand, the gene locations of major QTLs were acquired from previous gene characterization studies. Since the QTLs obtained from previous RIL population studies provided limited information about the CIs of QTL positions, we collected information about QTL marker positions, R² values, and population sizes from these studies. Then, the 95% CIs of QTL positions were estimated using the following equation: 95% CI for a RIL population = 163 / (N × R²) (R²: phenotypic variance explained by a QTL and N: population size) [64]. The positions and 95% CIs of all QTLs were projected onto the Consensus 4.0 genetic map for QTL confirmation [65]. Due to the framework consisting of various markers (e.g., SNP, SSR, RFLP, and morphological markers), the Consensus 4.0 genetic map has been utilized to confirm the CIs of QTL positions across different population studies.

Candidate gene and sequence analyses

The candidate genes for QTLs were screened within the 95% CIs of QTL positions using the 1.5-LOD support interval [62–63]. The sequences of soybean candidate genes with the Glyma2.0 gene model in the Wm82.a2.v1 were available from SoyBase [33] (https://www.soybase.org/tools/browsers/). The Arabidopsis information resource (TAIR) BLASTN (V. 2.9.0+) (https://www.arabidopsis.org/tools/blast/) was used to identify A. thaliana genes homologous to soybean candidate genes using two criteria, Bit score ( > 50) and E-value ( < 0.05).

Marker-assisted selection

MAS was performed using 86 QTLs that were identified by GWAS. The favorable and unfavorable alleles of a SNP in each G. max population were determined by comparing the average trait values of two accession groups based on the genotypes of alleles. The allele from the accession group with a lower average trait value was defined as a favorable allele. The percentage of 86 favorable alleles was calculated in an accession group with a trait value of j (j = 1, 2, 3, 4, and 5 on a scale of 1 to 5). The selection efficiency of 86 SNPs in each G. max population was defined as follows: Selection efficiency of 86 SNPs = (the average percentage of 86 favorable alleles in the accession group with a trait value of 1 + the average percentage of 86 unfavorable alleles in the accession group with a trait value of 5) / 2 / 100. The upper extreme trait value was 4 in MS967.

At a given SNP in each G. max population, the percentage of a favorable allele in an accession group with a trait value of j (j = 1, 2, 3, 4, and 5 on a scale of 1 to 5) was calculated as follows: The percentage of a favorable allele in an accession group with a trait value of j = (number of accessions with a favorable allele in an accession group with a trait value of j) / (total number of accessions in an accession group with a trait value of j) x 100. The selection efficiency of a SNP in each G. max population was defined as follows: Selection efficiency of a SNP = (the percentage of a favorable allele in the accession group with a trait value of 1 + the percentage of an unfavorable allele in the accession group with a trait value of 5) / 2 / 100.

Genomic selection

The trait, genotype, and PC matrix data for GWAS were also used for GS in each G. max population. Three basic models for the best linear unbiased prediction (BLUP), genomic best linear unbiased prediction (gBLUP) [66], compressed gBLUP (cBLUP) [67], and SUPER gBLUP (sBLUP) [67], were implemented for GS in GAPIT (V. 3.0). To investigate genomic prediction accuracy, K-fold cross-validation with replacement was employed at different folds (2, 5, 10, and 20) in gBLUP. The compressed accession groups were assigned as reference or inference panels. As the indicator of genomic prediction accuracy, the average correlation coefficient (r) [68] values between trait value and genomic prediction value at different folds were estimated with 1,000 times repetitions. The average r value ( ≥ 0.8) across the different folds was utilized as the criterion for selecting a G. max population with high genomic prediction accuracy. The r value between trait value and GBV in the selected G. max population was estimated in each BLUP model to measure selection efficiency.

Statistical analysis

In SAS 9.4 (SAS Institute Inc., Cary, NC), PROC TTEST, PROC FREQ, PROC GLM, PROC CORR, and PROC UNIVARIATE statements were used for t, χ², analysis of variance (ANOVA), correlation, and normality tests, respectively. The PROC TTEST statement was used to analyze climate data. The Welch’s t-test [69] was applied as an independent t-test, assuming two division groups had different sample sizes and unequal variances. The Welch-Satterthwaite equation [69] was employed to estimate the df from a linear combination of the two group variances. As part of the analysis for domestication-related SNPs, the PROC FREQ statement was utilized to calculate the observed frequencies of an allele at a SNP between G. soja and G. max populations. The CHISQ option estimated the expected frequencies of the allele. Additionally, the CHISQ option was used to conduct the χ² goodness of fit test [38] using the observed and expected allele frequencies. The PROC GLM statement was employed for ANOVA in MAS and GS analyses. In one- or two-way ANOVA, all independent variables had fixed effects. As a post hoc test, the Waller-Duncan k-ratio t-test [70] was conducted for multiple comparison tests (MCTs) if one-way ANOVA yielded a significant result. In the PROC CORR statement, the Pearson r [68] was used by default to measure the relationship between two variables. In the PROC UNIVARIATE statement, the NORMAL option was used to assess whether datasets, including a transformed pod dehiscence dataset, followed a normal distribution in the linear models of t-test, ANOVA, and GWAS. The Shapiro-Wilk test was employed to determine if a dataset was normally distributed. A one-tailed test with a 5% alpha level was applied for all statistical significance tests in this study.

Results

Selection of G. max populations

Pod dehiscence was evaluated in twenty-four G. max populations (S1 Table). Using all trait data from 24 G. max populations, the 95% CI for δ ranged from 0.587 to 0.698 in the χ² distribution. Four G. max populations, 1IL64, 1IL66, MN0102, and MS967, were selected (Table 1). The sample standard deviations of the selected G. max populations exceeded the upper limit of the 95% CI of δ, implying a high likelihood of identifying more QTLs due to greater variation. The MGs of 1IL64, 1IL66, and MN0102 ranged from 0 to IV. The MGs of MS967 ranged from V to VIII.

Download:

Table 1. Information about four G. max populations for pod dehiscence study.

https://doi.org/10.1371/journal.pone.0318815.t001

Evaluation of two climate factors

The soybean maturity zone in the USA indicated that three G. max populations (1IL64, 1IL66, and MN0102) and MS967 have been adapted in the Midwest and South, respectively (S1 Fig). Two major soybean production divisions, the Midwest and South, were compared for differences in temperature and precipitation (S1 Fig). According to the t-tests, the South showed higher temperature and precipitation than the Midwest (p-values < 1e-03). The average temperature and precipitation differences between the Midwest and South were 33.4 °F and 45.7 mm, respectively. The average temperature and precipitation in the South were 108.8 °F and 92.3 mm, respectively.

Putative domestication-related SNPs

The values of π, θ, and Tajima’s D were calculated for each G. max population. The values of π and θ ranged from 0.279 to 0.294 and 0.121 to 0.139 per SNP, respectively. Tajima’s D values ranged from 3.31 to 4.06, implying that all G. max populations lacked rare alleles and evolved under a non-random process.

As the criteria for the χ² goodness of fit test between WS1179 and a G. max population, four q-values, 0.0038, 0.0035, 0.0038, and 0.0029, were used to find significant SNPs in four pairs of datasets, WS1179-1IL64, WS1179-1IL66, WS1179-MN0102, and WS1179-MS967, respectively (S3 Table and S2 Fig). A total of 4,280 SNPs were significant at one of the four q-values. In 4,280 SNPs, 2,204 SNPs were significant in all G. max populations. The rest of the 2,076 SNPs were population-specifically significant in 1IL64 (284 SNPs), 1IL66 (173 SNPs), MN0102 (292 SNPs), and MS967 (1,327 SNPs).

The F_ST values for 4,280 SNPs were estimated between WS1179 and a G. max population (S3 Table and Fig 1). SNPs with significant q-values in the χ² goodness of fit test had F_ST values greater than 0.05. The average F_ST values were 0.38, 0.36, 0.38, and 0.33 in four pairs of datasets, WS1179-1IL64, WS1179-1IL66, WS1179-MN0102, and WS1179-MS967, respectively. These average F_ST values demonstrated that the genetic structure between G. soja and G. max populations was substantially different. The average F_ST values were estimated among G. max populations. The average F_ST values among 1IL64, 1IL66, and MN0102 ranged from 0.01 to 0.02, indicating high genetic similarity among the populations. The average F_ST values between MS967 and the other three G. max populations ranged from 0.04 to 0.06, showing that MS967 was moderately differentiated from 1IL64, 1IL66, and MN0102. Considering Tajima’s D, χ² goodness of fit test, and F_ST, 4,280 SNPs were detected as putative domestication-related SNPs.

Download:

Fig 1. F_ST values of putative domestication-related SNPs between WS1179 and a G. max population.

The F_ST values for 4,280 SNPs were estimated in four pairs of datasets, WS1179-1IL64, WS1179-1IL66, WS1179-MN0102, and WS1179-MS967.

https://doi.org/10.1371/journal.pone.0318815.g001

Gene ontology

A total of 1,515 out of 4,280 SNPs were located within soybean genes with the Glyma2.0 gene model (S3 Table). Therefore, 1,515 putative domestication-related genes were used for GO analysis. Based on the biological process GO term, 29 GO terms were identified (S3 Fig). Except for the GO term, uncategorized (no GO ID), the top 5 GO terms, signal transduction (GO:0007165) (12.03%), transport (GO:0006810) (9.49%), multicellular organismal development (GO:0007275) (7.91%), flower development (GO:0009908) (6.33%), and cell differentiation (GO:0030154) (5.85%), were annotated in the largest number of genes.

Population structure and linkage disequilibrium

The accessions of WS1179 originated from seven countries: China, Japan, South Korea, the Philippines, Russia, Taiwan, and the USA (S2 Table). WS1179 had eight subpopulations (K = 8). The K values of 1IL64, 1IL66, MN0102, and MS967 were 25, 23, 21, and 22, respectively (S4 Fig). PCA was performed to investigate the clustering of five populations in a two-dimensional coordinate system using principal component 1 (PC1) and principal component 2 (PC2) (Fig 2). The eigenvalues for PC1 and PC2 were 998.04 and 398.91. The percentage of variance explained by PC1 was 72%. The average coordinates were (21.5, ‒10.2), (0.1, ‒13.8), (-3.5, ‒13.3), (‒0.3, ‒13.5), and (0.5, 3.2) in WS1179, 1IL64, 1IL66, MN0102, and MS967, respectively. WS1179 was clustered separately from all G. max populations. MS967 showed a slightly different grouping compared to the other three G. max populations. In LD, the average r² values of 1IL64, 1IL66, MN0102, and MS967 were 0.35, 0.31, 0.33, and 0.26, respectively.

Download:

Fig 2. PCA using 39,509 common SNPs across all five populations.

Light blue, orange, gray, green, and black circles represented the accessions of WS1179, 1IL64, 1IL66, MN0102, and MS967, respectively.

https://doi.org/10.1371/journal.pone.0318815.g002

QTL analysis

Most QTLs for pod dehiscence were initially identified in each G. max population using PC and marker-based k matrices in GWAS models (S5 Fig). When we employed different combinations of cofactors in the GWAS models, additional QTLs were also discovered. Nevertheless, the overall QTL findings for each combination of cofactors remained largely consistent.

We chose QTLs that met one of two criteria: 1) QTLs colocalized in different G. max populations, or 2) QTLs tightly linked to previously reported QTLs. Eighty-six QTLs (S4 Table and Fig 3) were ultimately determined. Seventy-four new QTLs were colocalized in two different G. max populations. Twelve QTLs identified in a single G. max population were located near previously reported QTLs within a 2.2 Mb window. NST1A [21], SHAT1-5 [17], and Pdh1 [20] collected from previous gene characterization studies were closely located to 3 QTLs, ss715598070, ss715623567, and ss715624201, respectively. NST1A was located 93 kb upstream from ss715598070 on chromosome 7. SHAT1-5 was located 283 bp upstream from ss715623567 on chromosome 16. Pdh1 was located 42 kb downstream from ss715624201 on chromosome 16. Pdh1 was located between ss715624199 and ss715624201. ss715624199 was polymorphic in all G. max populations and located 20 kb downstream from Pdh1. Both ss715624199 and ss715624201 showed high LD across the G. max populations (average r² = 0.85). One RFLP [10] and eight SSRs [15–16] collected from previous RIL population studies (S5 Table) were located within a range of 8 kb to 1.8 Mb from 9 QTLs (ss715581242, ss715615326, ss715620013, ss715623192, ss715624318, ss715628052, ss715633661, ss715634837, and ss715636770) on chromosomes 2, 13, 14, 15, 16, 17, 19, and 20.

Download:

Fig 3. QTLs identified by GWAS in four G. max populations.

The black inverted triangle indicated the centromere of each chromosome. The green vertical lines displayed 74 QTLs identified in two different G. max populations. The blue vertical lines showed 12 QTLs located near previously reported QTLs within a 2.2 Mb window.

https://doi.org/10.1371/journal.pone.0318815.g003

The r² values of 9 pairs of QTLs (ss715579431-ss715579443, ss715580084-ss715580088, ss715610565-ss715610559, ss715610541-ss715610538, ss715621953-ss715621956, ss715620316-ss715620318, ss715624103-ss715624106, ss715625940-ss715625948, and ss715633661-ss715633692) on chromosomes 1, 11, 15, 16, 17, and 19 were greater than 0.8 in all G. max populations. Eight QTL, ss715579431, ss715583442, ss715588460, ss715598070, ss715605375, ss715607935, ss715610541, and ss715624201, were observed as domestication-related SNPs on chromosomes 1, 2, 4, 7, 9, 10, 11, and 16, respectively.

According to the 95% CIs of 86 QTL positions (S4 Table), soybean candidate genes from 86 QTLs and their homologous A. thaliana genes were described in S6 Table. Seventy-four new QTLs were confirmed in two different G. max populations over multiple years. In four pairs of QTLs (ss715580083-ss715580084, ss715588460-ss715588468, ss715610541-ss715610538, and ss715621953- ss715621956) on chromosomes 1, 4, 11, and 15, the 95% CIs of each pair of QTL positions overlapped. It was presumed that each pair of QTLs represented the same QTL. Because 12 out of 86 QTLs were identified in a single G. max population, we needed confirmation. As two major soybean genes related to pod dehiscence, Glyma.16G019400 (SHAT1-5) and Glyma.16G141500 (Pdh1) on chromosome 16 were located within the 95% CIs of two QTL positions for ss715623567 and ss715624201, respectively (S6 Table). Glyma.07G050600 (NST1A) on chromosome 7 was homologous to Glyma.16G019400 (SHAT1-5) in soybean but was not located within the 95% CI of the QTL position for ss715598070. The 95% CIs of QTL positions for one RFLP and eight SSRs (S5 Table) overlapped with the 95% CIs of 9 QTL positions (S4 Table).

Marker-assisted selection

The favorable alleles of 86 QTLs were identified in each G. max population (S7 Table). The percentage of 86 favorable alleles was calculated in accession groups based on trait values in each G. max population (Fig 4). Two-way ANOVA indicated that accession group, population, and their interaction effects were highly significant (p-values < 1e-03) across the G. max populations. In each G. max population, one-way ANOVA showed that the percentage of 86 favorable alleles among accession groups was significantly different (p-values < 1e-03).

Download:

Fig 4. Percentage of 86 favorable alleles in accession groups based on trait values in each G. max population.

Red, orange, gray, golden yellow, and blue rectangular bars represented accession groups with trait values of 1, 2, 3, 4, and 5, respectively. The upper extreme trait value was 4 in MS967. The Waller-Duncan k-ratio t-test was conducted for MCTs in each G. max population. The critical values for MCTs were 1.73, 1.74, 1.75, and 1.76 in 1IL64, 1IL66, MN0102, and MS967, respectively.

https://doi.org/10.1371/journal.pone.0318815.g004

Using two accession groups with extreme trait values, the selection efficiencies of 86 SNPs were 0.65, 0.59, 0.57, and 0.50 in 1IL64, 1IL66, MN0102, and MS967, respectively. The average selection efficiency of 86 SNPs was 0.57 across the G. max populations. The accession with the best combination of 86 favorable alleles was identified in each G. max population. Four accessions, PI89059 (72.1% / 62 QTLs), PI153292 (66.3% / 57 QTLs), PI592962B (76.7% / 66 QTLs), and PI416873B (66.3% / 57 QTLs), were found in 1IL64, 1IL66, MN0102, and MS967, respectively. The trait values of the four accessions were 1, indicating resistance to pod dehiscence.

The selection efficiency of each SNP was estimated. Table 2 displayed the top 5 SNPs with excellent selection efficiency in each G. max population. The SNP, ss715624201, was ranked first or second in 1IL64, 1IL66, and MN0102. The selection efficiency of ss715624201 ranged from 0.78 to 0.97. Two SNPs, ss715591071 and ss715606969, were ranked first in 1IL66 and MS967. Two SNPs, ss715623567 and ss715616944, were commonly observed in 1IL64 and 1IL66. The selection efficiencies of ss715623567 were 0.90 and 0.72 in 1IL64 and 1IL66. In MN0102, the selection efficiency of ss715598070 was 0.71. As previously reported QTLs (S4 Table), three SNPs, ss715598070, ss715623567, and ss715624201, were closely located to NST1A, SHAT1-5, and Pdh1, respectively.

Download:

Table 2. Top 5 SNPs with the highest selection efficiency in each G. max population.

https://doi.org/10.1371/journal.pone.0318815.t002

Genomic selection

The genomic prediction accuracies of four G. max populations were evaluated at different fold levels (Fig 5). Two-way ANOVA indicated that population, fold level, and their interaction effects were highly significant (p-values < 1e-03) across the fold levels. At each fold level, one-way ANOVA showed that the genomic prediction accuracies of four G. max populations were significantly different (p-values < 1e-03). The average genomic prediction accuracies across the fold levels were 0.76, 0.63, 0.59, and 0.56 in 1IL64, 1IL66, MN0102, and MS967, respectively.

Download:

Fig 5. Genomic prediction accuracies of four G. max populations at different fold levels in gBLUP.

Blue, orange, gray, and golden yellow rectangular bars represented 1IL64, 1IL66, MN0102, and MS967, respectively. The Waller-Duncan k-ratio t-test was conducted for MCTs at each fold level. The critical values for MCTs were 1.72144, 1.72146, 1.72144, and 1.72143 at 2-, 5-, 10-, and 20-fold levels, respectively.

https://doi.org/10.1371/journal.pone.0318815.g005

Because the average genomic prediction accuracy of 1IL64 was approximately 0.8, 1IL64 was selected for GS. The r values between trait value and GBV in 1IL64 were 0.148 (p-value < 1e-04), 0.153 (p-value < 1e-04), and 0.403 (p-value < 1e-04) in gBLUP, cBLUP, and sBLUP, respectively. 1IL64 had the largest r value in sBLUP, indicating the highest selection efficiency.

sBLUP was employed to estimate the GBV and prediction error variance (PEV) for accessions in 1IL64 (S8 Table and Fig 6). The GBV and PEV ranged from -33.06 to 31.03 and 214.23 to 582.37, respectively. The top 10 pod dehiscence-resistant accessions, PI84666-1, PI171421, PI88797, PI73780, PI153315, PI92470, PI142491, PI68696, PI79761, and PI88805-4, showed the lowest GBV. Except for PI84666-1 and PI153315, the trait values of eight accessions were 1. On the other hand, PI181570, PI90575, PI548329, PI84896, PI196158, PI181537, PI81033, PI189962, PI229354, and PI181548 were selected as the top 10 pod dehiscence-susceptible accessions with the highest GBV. The trait values of only two accessions, PI181570 and PI548329, were 5.

Download:

Fig 6. Distribution of the GBV and PEV for accessions in 1IL64 when sBLUP was applied.

In the bottom and top 5% of the GBV distribution, solid red, blue, and black circles represented accessions with trait values of 1, 5, and others, respectively. Solid green and orange circles corresponded to two accessions, PI89059 and PI181532, respectively.

https://doi.org/10.1371/journal.pone.0318815.g006

Two groups of 36 accessions were evaluated in the bottom and top 5% of the GBV distribution (Fig 6). The trait values of 30 (83.3%) and 13 (36.1%) accessions were 1 and 5 in the bottom and top 5%. In MAS, PI89059 showed the highest percentage (72.1%) of 86 favorable alleles with a trait value of 1. In contrast, PI181532 had the highest percentage (80.2%) of 86 unfavorable alleles with a trait value of 5. The GBV (19.7) of PI181532 was near the top 5%, while the GBV (-4.23) of PI89059 was in the middle of the distribution.

Discussion

Soybean candidate genes from 86 QTLs and their homologous A. thaliana genes were listed in S6 Table. Based on previous A. thaliana studies, we highlighted some notable A. thaliana genes involved in the lignification, breakdown, and biosynthesis of the cell wall, serving as critical processes for pod dehiscence.

We first explored A. thaliana genes related to cell wall components. The mature sclerenchyma of the pod wall consists of dead cells. AT5G01660.1 [71] functioned as a protein with the development and cell death (DCD) domain and was homologous to Glyma.01G126500 on chromosome 1. Laccase synthesizes lignin as the multicopper oxidase of phenols [72]. As one of the 17 laccase family genes, AT5G03260.1 (laccase 11) [73] played a role in lignin biosynthesis and was homologous to Glyma.01G173600 on chromosome 1. As an auxin response factor (ARF), AT5G62000.1 (ARF2) [74] mutation study indicated that arf2 allele had a pleiotropic effect on multiple traits, including delayed silique dehiscence. AT5G62000.1 was homologous to Glyma.04G200600 on chromosome 4. Hydrolase and cellulose synthase (CESA) are essential enzymes forming the cell wall. AT1G32860.1 [75] hydrolyzed the glycosidic bond between two or more carbohydrates, modifying the cell wall during silique dehiscence. AT5G17420.1 (CESA7) [76] mutant study showed that cesa7 affected cell growth, cell wall integrity, and cellulose level. AT1G32860.1 and AT5G17420.1 were homologous to Glyma.10G172300 and Glyma.17G072200 on chromosomes 10 and 17, respectively. Pectin is a complex polysaccharide of the primary cell wall. AT2G45220.1 [77] catalyzed the homogalacturonan backbone, remodeling the cell wall and modifying the physicochemical properties of pectin. AT1G05675.1 [78] was involved in pectin biosynthesis as a glycosyltransferase. AT2G45220.1 and AT1G05675.1 were homologous to Glyma.03G029000 and Glyma.19G025100 on chromosomes 3 and 19, respectively. Xylan is a polysaccharide mainly found in the secondary cell wall (SCW). AT1G27930.1 [79] modified xylan in the SCW by catalyzing the 4-O-methylation of glucuronoxylan. AT1G27930.1 was homologous to Glyma.11G251000 on chromosome 11.

Some A. thaliana genes were associated with signal transduction. It is well-recognized that SCW-associated NAC domain proteins are TFs and control SCW biosynthesis as central genes in a regulatory gene network [80–81]. AT4G29230.1 [82] regulated SCW biosynthesis by activating SCW NAC binding elements and initiating the transcription of target genes. AT4G29230.1 was homologous to Glyma.05G108700 on chromosome 5. Four MYB TF paralogs, AT2G38300.1, AT2G32460.1, AT3G09370.1, and AT5G15310.1, affected the transcriptional regulation of SCW biosynthesis [83]. AT2G38300.1, AT2G32460.1, AT3G09370.1, and AT5G15310.1 were homologous to Glyma.02G282900, Glyma.12G193300, Glyma.16G137000, and Glyma.19G024700 on chromosomes 2, 12, 16, and 19, respectively. A leucine-rich repeat receptor-like kinase (LRR-RLK) is a receptor on the plant cell membrane, receiving external signals for intracellular signal transduction. Four LRR-RLK families, AT1G49490.1, AT3G19020.1, AT2G33170.1, and AT2G36570.1, functioned as regulators of SCW formation [84]. AT1G49490.1, AT3G19020.1, AT2G33170.1, and AT2G36570.1 were homologous to Glyma.07G164300, Glyma.10G090800, Glyma.13G227800, and Glyma.19G190200 on chromosomes 7, 10, 13, and 19, respectively.

According to QTL information from previous soybean population studies, fifteen out of 86 QTLs for pod dehiscence exhibited pleiotropic effects on various agronomic traits, including the first flower, pod maturity, seed composition, seed weight, and yield [85–102] (S9 Table). Eight out of 15 QTLs for pod dehiscence were associated with seed weight or yield, implying that pod dehiscence significantly impacted soybean yield (S9 Table). Many studies have reported the relationship among agronomic traits in soybean. As major agronomic traits, the first flower, pod maturity, seed weight, and yield were coinherited with pod dehiscence [9,22,23]. It was well-known that seed weight positively correlated with yield [103–104]. The genotypic correlation between yield and oil was positive, and the genotypic correlation between oil (and yield) and protein was negative [105–107].

From the perspective of soybean production, we need to consider two key points to reduce pod dehiscence. First, monitoring environmental conditions during the harvest period is important. After R8 (full maturity stage), soybean harvesting typically requires 5 to 10 days of dry weather to reduce the seed moisture level to 13% [11]. Pod dehiscence-susceptible soybean cultivars resulted in yield losses of 40–100% under severe dry weather conditions or harvesting delays caused by frequent rainfall [25–26]. Climate data analysis showed that the Southern region of the USA has experienced significantly higher temperature and precipitation than the Midwest over the past 125 years (S1 Fig). The subtropical climate of the Southern USA has created suitable conditions for accelerating pod dehiscence [9,13,25]. In the Southeastern USA, pod dehiscence contributed 37% (53–310 kg ha^-1) of total soybean field loss, and the early harvesting system mitigated this loss [24]. Therefore, effective field management plans will be required to prevent huge yield loss from pod dehiscence.

Second, it is critical to develop pod dehiscence-resistant cultivars with the optimal haplotype using MAS and GS. We identified accessions with the best haplotype and the top 5 SNPs with the highest selection efficiency in MAS. Because of differences in population structure among G. max populations (S4 Fig), it was challenging to determine the consensus favorable allele of each QTL across the G. max populations (S7 Table). Nevertheless, information about 86 QTLs (S4 Table) and the top 5 SNPs (Table 2) would be valuable for MAS in further population studies.

Previously reported QTLs were used to confirm 12 out of 86 QTLs (S4 Table). NST1A was not located within 95% CI of the QTL position for ss715598070. However, NST1A was closely linked to ss715598144 within a 4 kb range, and the LD between ss715598070 and ss715598144 was considerably high across the G. max populations (average r² = 0.71). ss715598070 and NST1A are likely to be inherited together. One RFLP [10] and eight SSRs [15–16] (S5 Table) were identified in three RIL populations. The RIL populations exhibited a small population size ( < 200). The Beavis effect [108] may decrease R² values and overestimate 95% CIs in the function [64]. Additionally, a small number of markers were used to construct genetic maps of the RIL populations, implying that a lower marker density may decrease R² values and impair QTL detection power [109]. Therefore, except for ss715623567 and ss715624201, the rest of the 10 QTLs will require validation in future studies.

In the comparison between MAS and GS in 1IL64, MAS outperformed GS, showing higher selection efficiency. PI89059 was the accession with the best haplotype in MAS. However, the accession was positioned in the middle of the GBV distribution in sBLUP (Fig 6). These different results between MAS and GS appeared to be due to the GS model. sBLUP utilized the additive genetic effects of all polymorphic SNPs across the genome [67]. The accessions selected by MAS and GS in this study would be useful sources for determining the optimal haplotype of pod dehiscence-resistant cultivars. It will be essential to acquire more population data for MAS by confirming additional QTLs, assessing QTL stability, and testing QTL gene action in future population studies. On the other hand, assessing other advanced GS models will be necessary to find the best-fit model.

Conclusion

Pod dehiscence is a well-known trait with high heritability in soybean, although some climate factors, such as temperature and precipitation, significantly impact pod dehiscence. As a domestication trait, pod dehiscence is a primary cause of yield loss in cultivated soybean. We obtained data for one G. soja and four G. max populations from the USDA database. GWAS identified 86 QTLs using four G. max populations. Eight out of 86 QTLs were related to the domestication of pod dehiscence. MAS and GS were implemented to select pod dehiscence-resistant accessions with the best haplotype and lowest GBV. The selected accessions would assist in identifying the optimal haplotype for developing pod dehiscence-resistant cultivars. However, further studies will be required to obtain additional QTLs and evaluate advanced GS models.

Supporting information

S1 Fig. Statewide average temperature and precipitation from September to November, spanning the years 1895-2020 in the USA.

Red dots showed the field locations of four G. max populations. The adaptation zones for soybean MGs and four divisions in the USA were displayed on the left and right climate maps. Both climate maps were obtained from the NCEI.

https://doi.org/10.1371/journal.pone.0318815.s001

(PPTX)

S2 Fig.

Putative domestication-related SNPs with significantly different allele frequencies between WS1179 and a G. max population. A total of 2,204 SNPs (magenta vertical lines) were significant in all G. max populations. From the rest of the 2,076 SNPs, 284 SNPs (dark orange vertical lines), 173 SNPs (golden brown vertical lines), 292 SNPs (green vertical lines), and 1,327 (blue vertical lines) were significant in 1IL64, 1IL66, MN0102, and MS967, respectively.

https://doi.org/10.1371/journal.pone.0318815.s002

(PPTX)

S3 Fig. GO analysis of 1,515 putative domesticated-related genes.

The pie chart displayed the percentage of annotated gene numbers for each GO term. The biological process GO terms were presented on the right.

https://doi.org/10.1371/journal.pone.0318815.s003

(PPTX)

S4 Fig. Population structure in WS1179, 1IL64, 1IL66, MN0102, and MS967.

In the individual population structure image, a single vertical line corresponded to an accession. The single vertical line was divided into K-colored segments, and the segment length was proportional to each K-inferred cluster. The numbers below the individual population structure image indicated the number of subpopulations, K.

https://doi.org/10.1371/journal.pone.0318815.s004

(PPTX)

S5 Fig. QTL detected by multiple GWAS models in four G. max populations.

The circular (top) and quantile-quantile (bottom) plots of 1IL64, 1IL66, MN0102, and MS967 were arranged from left to right. Only the circular and Q-Q plots for a combination of PC and marker-based k matrices were presented. In the circular plot, a red asterisk showed the position of a QTL, and a LOD value was used instead of a q-value (q-value = 10^-LOD). A red dotted circle in the circular plot indicated the threshold for QTL detection in each GWAS model.

https://doi.org/10.1371/journal.pone.0318815.s005

(PPTX)

S1 Table. Information about twenty-four G. max populations for pod dehiscence study in the GRIN.

https://doi.org/10.1371/journal.pone.0318815.s006

(XLSX)

S2 Table. Information about 1,179 wild accessions in WS1179.

https://doi.org/10.1371/journal.pone.0318815.s007

(XLSX)

S3 Table. P-values for χ ² goodness of fit tests and F_ST values in 4,280 SNPs between WS1179 and a G. max population.

https://doi.org/10.1371/journal.pone.0318815.s008

(XLSX)

S4 Table. Information about 86 QTLs identified in four G. max populations.

https://doi.org/10.1371/journal.pone.0318815.s009

(XLSX)

S5 Table. QTL information of one RFLP and eight SSRs from three RIL population studies.

All RIL population studies provided limited information about the CIs of QTL positions. The 95% CIs of QTL positions were estimated using the following equation: 95% CI for a RIL population = 163 / (N × R²) (R²: phenotypic variance explained by a QTL and N: population size) [64]. The positions and 95% CIs of all QTLs were projected onto the Consensus 4.0 genetic map for QTL confirmation. Therefore, left and right flanking makers covering 95% CIs were determined based on the Consensus 4.0 genetic map. NA stands for “not available.”

https://doi.org/10.1371/journal.pone.0318815.s010

(XLSX)

S6 Table. Soybean candidate genes from 86 QTLs for pod dehiscence.

A. thaliana genes homologous to soybean candidate genes were included in the table. No soybean candidate genes for 8 QTLs on chromosomes 4, 7, 8, 13, 17, and 20 were found. NA stands for “not available.”

https://doi.org/10.1371/journal.pone.0318815.s011

(XLSX)

S7 Table. Favorable alleles of 86 QTLs in each G. max population.

https://doi.org/10.1371/journal.pone.0318815.s012

(XLSX)

S8 Table. GBV and PEV for each accession in 1IL64 using sBLUP.

https://doi.org/10.1371/journal.pone.0318815.s013

(XLSX)

S9 Table. Different agronomic traits associated with pod dehiscence.

From 86 QTLs for pod dehiscence, a 10 kb window of each QTL was used to find QTLs for different agronomic traits based on previous soybean population studies. Favorable alleles contributed to the decrease in the number of open pods at harvest.

https://doi.org/10.1371/journal.pone.0318815.s014

(XLSX)

Acknowledgments

Authors gratefully acknowledge public trait and genotype data from the USDA-ARS. We thank the technical support of SoyBase curators and staff.

References

1. Fuller DQ. Contrasting patterns in crop domestication and domestication rates: recent archaeobotanical insights from the Old World. Ann Bot. 2007;100(5):903–24. pmid:17495986
- View Article
- PubMed/NCBI
- Google Scholar
2. Petersen M, Sander L, Child R, van Onckelen H, Ulvskov P, Borkhardt B. Isolation and characterisation of a pod dehiscence zone-specific polygalacturonase from Brassica napus. Plant Mol Biol. 1996;31(3):517–27. pmid:8790285
- View Article
- PubMed/NCBI
- Google Scholar
3. Meyer RS, Purugganan MD. Evolution of crop species: genetics of domestication and diversification. Nat Rev Genet. 2013;14(12):840–52. pmid:24240513
- View Article
- PubMed/NCBI
- Google Scholar
4. Di Vittori V, Gioia T, Rodriguez M, Bellucci E, Bitocchi E, Nanni L, et al. Convergent evolution of the seed shattering trait. Genes (Basel). 2019;10(1):68. pmid:30669440
- View Article
- PubMed/NCBI
- Google Scholar
5. Anderson E, Ali M, Beavis W, Chen P, Clemente T, Diers B. Soybean [Glycine max (L.) Merr.] Breeding: History, Improvement, Production and Future Opportunities. Adv Plant Breed Strateg Legumes. 2019;431–516.
- View Article
- Google Scholar
6. Tukamuhabwa P, Rubaihayo P, Dashiell KE. Euphytica. 2002;125(1):29–34.
- View Article
- Google Scholar
7. Bara N, Khare D, Shrivastava AN. Studies on the factors affecting pod shattering in soybean. Ind Jrnl Gen Plnt Bree. 2013;73(3):270.
- View Article
- Google Scholar
8. Caviness CE. Heritability of pod dehiscence and its association with some agronomic characters in soybean. Crop Sci. 1969;9(2):207–9.
- View Article
- Google Scholar
9. Tsuchiya T. Physiological and genetic analysis of pod shattering in soybean. Jpn Agric Res Q. 1987;21(3):166–75.
- View Article
- Google Scholar
10. Bailey M, Mian M, Carter Jr T, Ashley D, Boerma H. Pod dehiscence of soybean: identification of quantitative trait loci. J Hered. 1997;88(2):152–4.
- View Article
- Google Scholar
11. Fehr WR, Caviness CE, Burmood DT, Pennington JS. Stage of Development Descriptions for Soybeans, Glycine Max (L.) Merrill1. Crop Sci. 1971;11(6):929–31.
- View Article
- Google Scholar
12. Lu S, Zhao X, Hu Y, Liu S, Nan H, Li X, et al. Natural variation at the soybean J locus improves adaptation to the tropics and enhances yield. Nat Genet. 2017;49(5):773–9. pmid:28319089
- View Article
- PubMed/NCBI
- Google Scholar
13. Franca-Neto J, Henning A, Krzyzanowski F. Tropical soybean: improvement and production. Seed Production and Technology for the Tropics. FAO Plant Production and Protection Series. 1994;217–40.
- View Article
- Google Scholar
14. Funatsuki H, Ishimoto M, Tsuji H, Kawaguchi K, Hajika M, Fujino K. Simple sequence repeat markers linked to a major QTL controlling pod shattering in soybean. Plant Breed. 2006;125(2):195–7.
- View Article
- Google Scholar
15. Liu B, Fujita T, Yan Z-H, Sakamoto S, Xu D, Abe J. QTL mapping of domestication-related traits in soybean (Glycine max). Ann Bot. 2007;100(5):1027–38. pmid:17684023
- View Article
- PubMed/NCBI
- Google Scholar
16. Kang S, Kwak M, Kim H, Choung M, Han W, Baek I. Population-specific QTLs and their different epistatic interactions for pod dehiscence in soybean [Glycine max (L.) Merr.]. Euphytica. 2009;166:15–24.
- View Article
- Google Scholar
17. Dong Y, Yang X, Liu J, Wang B-H, Liu B-L, Wang Y-Z. Pod shattering resistance associated with domestication is mediated by a NAC gene in soybean. Nat Commun. 2014;5:3352. pmid:24549030
- View Article
- PubMed/NCBI
- Google Scholar
18. Funatsuki H, Hajika M, Hagihara S, Yamada T, Tanaka Y, Tsuji H. Confirmation of the location and the effects of a major QTL controlling pod dehiscence, qPDH1, in soybean. Breeding Sci. 2008;58(1):63–9.
- View Article
- Google Scholar
19. Yamada T, Funatsuki H, Hagihara S, Fujita S, Tanaka Y, Tsuji H. A major QTL, qPDH1, is commonly involved in shattering resistance of soybean cultivars. Breeding Sci. 2009;59(4):435–40.
- View Article
- Google Scholar
20. Funatsuki H, Suzuki M, Hirose A, Inaba H, Yamada T, Hajika M, et al. Molecular basis of a shattering resistance boosting global dissemination of soybean. Proc Natl Acad Sci U S A. 2014;111(50):17797–802. pmid:25468966
- View Article
- PubMed/NCBI
- Google Scholar
21. Zhang J, Singh AK. Genetic control and geo-climate adaptation of pod dehiscence provide novel insights into soybean domestication. G3 (Bethesda). 2020;10(2):545–54. pmid:31836621
- View Article
- PubMed/NCBI
- Google Scholar
22. Thakare DS, Chimote VP, Rokade RC, Deshmukh MP. Correlation of yield and associated with other characters in pod shattering soybean (Glycine max (L.) Merrill). Int J Trop Agric. 2016;34(6):1915–23.
- View Article
- Google Scholar
23. Kataliko R, Kimani P, Muthomi J, Wanderi W, Olubayo F, Nzuve F. Resistance and correlation of pod shattering and selected agronomic traits in soybeans. J lant Stud. 2019;8(2):39–48.
- View Article
- Google Scholar
24. Philbrook BD, Oplinger ES. Soybean field losses as influenced by harvest delays. Agron J. 1989;81(2):251–8.
- View Article
- Google Scholar
25. Tiwari S, Bhatnagar P. Pod shattering as related to other agronomic attributes in soybean. Trop Agric. 1991;68(1):102–3.
- View Article
- Google Scholar
26. Bhor T, Chimote V, Deshmukh M. Inheritance of pod shattering in soybean [Glycine max (L.) Merrill]. Electron J Plant Breed. 2014;5(4):671–6.
- View Article
- Google Scholar
27. Bernard R, Cremeens C, Cooper R, Collins F, Krober O, Athow K, et al. Evaluation of the USDA soybean germplasm collection: maturity groups 000-IV (FC 01.547-PI 266.807). Tech Bull USDA. 1998;1844:1–169.
- View Article
- Google Scholar
28. Hill J, Peregrine E, Sprau G, Cremeens C, Nelson R, Orf J. Evaluation of the USDA soybean germplasm collection: maturity groups 000-IV (PI 578371-PI 612761). Technical Bulletin USDA. 20081–151.
- View Article
- Google Scholar
29. Peregrine EK, Sprau GL, Cremeens CR, Handly P, Kilen TC, Smith JR. Evaluation of the USDA soybean germplasm collection: maturity group V (FC 30265-PI 612614) and maturity groups VI-VIII (PI 416758-PI 606432B). Techn Bull USDA. 1920;1–369.
- View Article
- Google Scholar
30. Zhang LX, Kyei-Boahen S, Zhang J, Zhang MH, Freeland TB, Watson CE Jr, et al. Modifications of Optimum Adaptation Zones for Soybean Maturity Groups in the USA. Crop Manag. 2007;6(1):1–11.
- View Article
- Google Scholar
31. Freeman MF, Tukey JW. Transformations Related to the Angular and the Square Root. Ann Math Statist. 1950;21(4):607–11.
- View Article
- Google Scholar
32. Song Q, Hyten DL, Jia G, Quigley CV, Fickus EW, Nelson RL, et al. Fingerprinting soybean germplasm and its utility in genomic research. G3 (Bethesda). 2015;5(10):1999–2006. pmid:26224783
- View Article
- PubMed/NCBI
- Google Scholar
33. Grant D, Nelson RT, Cannon SB, Shoemaker RC. SoyBase, the USDA-ARS soybean genetics and genomics database. Nucleic Acids Res. 2010;38(Database issue):D843-6. pmid:20008513
- View Article
- PubMed/NCBI
- Google Scholar
34. Schmutz J, Cannon SB, Schlueter J, Ma J, Mitros T, Nelson W, et al. Genome sequence of the palaeopolyploid soybean. Nature. 2010;463(7278):178–83. pmid:20075913
- View Article
- PubMed/NCBI
- Google Scholar
35. Money D, Gardner K, Migicovsky Z, Schwaninger H, Zhong G-Y, Myles S. LinkImpute: fast and accurate genotype imputation for nonmodel organisms. G3 (Bethesda). 2015;5(11):2383–90. pmid:26377960
- View Article
- PubMed/NCBI
- Google Scholar
36. Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 2007;23(19):2633–5. pmid:17586829
- View Article
- PubMed/NCBI
- Google Scholar
37. Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 1989;123(3):585–95. pmid:2513255
- View Article
- PubMed/NCBI
- Google Scholar
38. Kallenberg W, Oosterhoff J, Schriever B. The number of classes in chi-squared goodness-of-fit tests. J Am Stat Assoc. 1985;80(392):959–68.
- View Article
- Google Scholar
39. Falconer D, Mackay T. Introduction to Quantitative Genetics. 4th ed. Harlow, UK: Longman Scientific & Technical;1996;57–64.
40. Xu X, Liu X, Ge S, Jensen JD, Hu F, Li X, et al. Resequencing 50 accessions of cultivated and wild rice yields markers for identifying agronomically important genes. Nat Biotechnol. 2011;30(1):105–11. pmid:22158310
- View Article
- PubMed/NCBI
- Google Scholar
41. Varshney RK, Song C, Saxena RK, Azam S, Yu S, Sharpe AG, et al. Draft genome sequence of chickpea (Cicer arietinum) provides a resource for trait improvement. Nat Biotechnol. 2013;31(3):240–6. pmid:23354103
- View Article
- PubMed/NCBI
- Google Scholar
42. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Stat Method. 1995;57(1):289–300.
- View Article
- Google Scholar
43. Wright S. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution. 1965;19(3):395–420.
- View Article
- Google Scholar
44. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. the gene ontology consortium. Nat Genet. 2000;25(1):25–9. pmid:10802651
- View Article
- PubMed/NCBI
- Google Scholar
45. Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164(4):1567–87. pmid:12930761
- View Article
- PubMed/NCBI
- Google Scholar
46. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155(2):945–59. pmid:10835412
- View Article
- PubMed/NCBI
- Google Scholar
47. Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14(8):2611–20. pmid:15969739
- View Article
- PubMed/NCBI
- Google Scholar
48. Rosenberg NA, Burke T, Elo K, Feldman MW, Freidlin PJ, Groenen MA, et al. Empirical evaluation of genetic clustering methods using multilocus genotypes from 20 chicken breeds. Genetics. 2001;159(2):699–713. pmid:11606545
- View Article
- PubMed/NCBI
- Google Scholar
49. Lewontin RC. The interaction of selection and linkage. i. general considerations; heterotic models. Genetics. 1964;49(1):49–67. pmid:17248194
- View Article
- PubMed/NCBI
- Google Scholar
50. Troyanskaya O, Cantor M, Sherlock G, Brown P, Hastie T, Tibshirani R, et al. Missing value estimation methods for DNA microarrays. Bioinformatics. 2001;17(6):520–5. pmid:11395428
- View Article
- PubMed/NCBI
- Google Scholar
51. Endelman JB, Jannink J-L. Shrinkage estimation of the realized relationship matrix. G3 (Bethesda). 2012;2(11):1405–13. pmid:23173092
- View Article
- PubMed/NCBI
- Google Scholar
52. Lipka AE, Tian F, Wang Q, Peiffer J, Li M, Bradbury PJ, et al. GAPIT: genome association and prediction integrated tool. Bioinformatics. 2012;28(18):2397–9. pmid:22796960
- View Article
- PubMed/NCBI
- Google Scholar
53. R Core Team. R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria. 2013. Available from: http://www.R-project.org/.
- View Article
- Google Scholar
54. VanRaden PM. Efficient methods to compute genomic predictions. J Dairy Sci. 2008;91(11):4414–23. pmid:18946147
- View Article
- PubMed/NCBI
- Google Scholar
55. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9. pmid:16862161
- View Article
- PubMed/NCBI
- Google Scholar
56. Yu J, Pressoir G, Briggs W, Vroh BI, Yamasaki M, Doebley J. A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet. 2006;38(2):203–8.
- View Article
- Google Scholar
57. Wang Q, Tian F, Pan Y, Buckler ES, Zhang Z. A SUPER powerful method for genome wide association study. PLoS One. 2014;9(9):e107684. pmid:25247812
- View Article
- PubMed/NCBI
- Google Scholar
58. Segura V, Vilhjálmsson BJ, Platt A, Korte A, Seren Ü, Long Q, et al. An efficient multi-locus mixed-model approach for genome-wide association studies in structured populations. Nat Genet. 2012;44(7):825–30. pmid:22706313
- View Article
- PubMed/NCBI
- Google Scholar
59. Liu X, Huang M, Fan B, Buckler ES, Zhang Z. Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLoS Genet. 2016;12(2):e1005767. pmid:26828793
- View Article
- PubMed/NCBI
- Google Scholar
60. Huang M, Liu X, Zhou Y, Summers RM, Zhang Z. BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. Gigascience. 2019;8(2):giy154. pmid:30535326
- View Article
- PubMed/NCBI
- Google Scholar
61. Hyten DL, Choi I, Song Q, Specht JE, Carter TE, Shoemaker RC, et al. A High density integrated genetic linkage map of soybean and the development of a 1536 universal soy linkage panel for quantitative trait locus mapping. Crop Sci. 2010;50(3):960–8.
- View Article
- Google Scholar
62. Dupuis J, Siegmund D. Statistical methods for mapping quantitative trait loci from a dense set of markers. Genetics. 1999;151(1):373–86. pmid:9872974
- View Article
- PubMed/NCBI
- Google Scholar
63. Li H. A quick method to calculate QTL confidence interval. J Genet. 2011;90(2):355–60. pmid:21869489
- View Article
- PubMed/NCBI
- Google Scholar
64. Guo B, Sleper D, Lu P, Shannon J, Nguyen H, Arelli P. QTLs associated with resistance to soybean cyst nematode in soybean: Meta‐analysis of QTL locations. Crop Sci. 2006;46(2):595–602.
- View Article
- Google Scholar
65. Hwang S, King CA, Ray JD, Cregan PB, Chen P, Carter TE Jr, et al. Confirmation of delayed canopy wilting QTLs from multiple soybean mapping populations. Theor Appl Genet. 2015;128(10):2047–65. pmid:26163767
- View Article
- PubMed/NCBI
- Google Scholar
66. Zhang Z, Todhunter RJ, Buckler ES, Van Vleck LD. Technical note: Use of marker-based relationships with multiple-trait derivative-free restricted maximal likelihood. J Anim Sci. 2007;85(4):881–5. pmid:17085728
- View Article
- PubMed/NCBI
- Google Scholar
67. Wang J, Zhou Z, Zhang Z, Li H, Liu D, Zhang Q, et al. Expanding the BLUP alphabet for genomic prediction adaptable to the genetic architectures of complex traits. Heredity (Edinb). 2018;121(6):648–62. pmid:29765161
- View Article
- PubMed/NCBI
- Google Scholar
68. Sedgwick P. Pearson’s correlation coefficient. Br Med J. 2012;345:e4483.
- View Article
- Google Scholar
69. Ruxton GD. The unequal variance t-test is an underused alternative to Student’s t-test and the Mann–Whitney U test. Behavioral Ecology. 2006;17(4):688–90.
- View Article
- Google Scholar
70. Duncan DB, Godbold JH. Approximate k-Ratio t Tests for Differences between Unequally Replicated Treatments. Biometrics. 1979;35(4):749.
- View Article
- Google Scholar
71. Tenhaken R, Doerks T, Bork P. DCD - a novel plant specific domain in proteins involved in development and programmed cell death. BMC Bioinformatics. 2005;6:169. pmid:16008837
- View Article
- PubMed/NCBI
- Google Scholar
72. Wang P, Guo L, Morgan J, Dudareva N, Chapple C. Transcript and metabolite network perturbations in lignin biosynthetic mutants of Arabidopsis. Plant Physiol. 2022;190(4):2828–46. pmid:35880844
- View Article
- PubMed/NCBI
- Google Scholar
73. Turlapati PV, Kim K-W, Davin LB, Lewis NG. The laccase multigene family in Arabidopsis thaliana: towards addressing the mystery of their gene function(s). Planta. 2011;233(3):439–70. pmid:21063888
- View Article
- PubMed/NCBI
- Google Scholar
74. Okushima Y, Mitina I, Quach HL, Theologis A. AUXIN RESPONSE FACTOR 2 (ARF2): a pleiotropic developmental regulator. Plant J. 2005;43(1):29–46. pmid:15960614
- View Article
- PubMed/NCBI
- Google Scholar
75. He H, Bai M, Tong P, Hu Y, Yang M, Wu H. CELLULASE6 and MANNANASE7 affect cell differentiation and silique dehiscence. Plant Physiol. 2018;176(3):2186–201. pmid:29348141
- View Article
- PubMed/NCBI
- Google Scholar
76. Kumar M, Mishra L, Carr P, Pilling M, Gardner P, Mansfield SD, et al. Exploiting CELLULOSE SYNTHASE (CESA) class specificity to probe cellulose microfibril biosynthesis. Plant Physiol. 2018;177(1):151–67. pmid:29523715
- View Article
- PubMed/NCBI
- Google Scholar
77. Lionetti V, Fabri E, De Caroli M, Hansen AR, Willats WGT, Piro G, et al. Three pectin methylesterase inhibitors protect cell wall integrity for arabidopsis immunity to botrytis. Plant Physiol. 2017;173(3):1844–63. pmid:28082716
- View Article
- PubMed/NCBI
- Google Scholar
78. Voxeur A, André A, Breton C, Lerouge P. Identification of putative rhamnogalacturonan-II specific glycosyltransferases in Arabidopsis using a combination of bioinformatics approaches. PLoS One. 2012;7(12):e51129. pmid:23272088
- View Article
- PubMed/NCBI
- Google Scholar
79. Yuan Y, Teng Q, Lee C, Zhong R, Ye Z-H. Modification of the degree of 4-O-methylation of secondary wall glucuronoxylan. Plant Sci. 2014;219–220:42–50. pmid:24576763
- View Article
- PubMed/NCBI
- Google Scholar
80. Jaradat MR, Ruegger M, Bowling A, Butler H, Cutler AJ. A comprehensive transcriptome analysis of silique development and dehiscence in Arabidopsis and Brassica integrating genotypic, interspecies and developmental comparisons. GM Crops Food. 2014;5(4):302–20. pmid:25523176
- View Article
- PubMed/NCBI
- Google Scholar
81. Taylor-Teeples M, Lin L, de Lucas M, Turco G, Toal TW, Gaudinier A, et al. An Arabidopsis gene regulatory network for secondary cell wall synthesis. Nature. 2015;517(7536):571–5. pmid:25533953
- View Article
- PubMed/NCBI
- Google Scholar
82. Zhong R, Lee C, Haghighat M, Ye Z-H. Xylem vessel-specific SND5 and its homologs regulate secondary wall biosynthesis through activating secondary wall NAC binding elements. New Phytol. 2021;231(4):1496–509. pmid:33908063
- View Article
- PubMed/NCBI
- Google Scholar
83. Kim W-C, Kim J-Y, Ko J-H, Kang H, Han K-H. Identification of direct targets of transcription factor MYB46 provides insights into the transcriptional regulation of secondary wall biosynthesis. Plant Mol Biol. 2014;85(6):589–99. pmid:24879533
- View Article
- PubMed/NCBI
- Google Scholar
84. Wang J, Kucukoglu M, Zhang L, Chen P, Decker D, Nilsson O, et al. The Arabidopsis LRR-RLK, PXC1, is a regulator of secondary wall formation correlated with the TDIF-PXY/TDR-WOX4 signaling pathway. BMC Plant Biol. 2013;13:94. pmid:23815750
- View Article
- PubMed/NCBI
- Google Scholar
85. Khan NA, Githiri SM, Benitez ER, Abe J, Kawasaki S, Hayashi T, et al. QTL analysis of cleistogamy in soybean. Theor Appl Genet. 2008;117(4):479–87. pmid:18506418
- View Article
- PubMed/NCBI
- Google Scholar
86. Han Y, Teng W, Wang Y, Zhao X, Wu L, Li D, et al. Unconditional and conditional QTL underlying the genetic interrelationships between soybean seed isoflavone, and protein or oil contents. Plant Breeding. 2015;134(3):300–9.
- View Article
- Google Scholar
87. Sun Y-N, Pan J-B, Shi X-L, Du X-Y, Wu Q, Qi Z-M, et al. Multi-environment mapping and meta-analysis of 100-seed weight in soybean. Mol Biol Rep. 2012;39(10):9435–43. pmid:22740134
- View Article
- PubMed/NCBI
- Google Scholar
88. Qi Z, Wu Q, Han X, Sun Y, Du X, Liu C, et al. Soybean oil content QTL mapping and integrating with meta-analysis method for mining genes. Euphytica. 2011;179(3):499–514.
- View Article
- Google Scholar
89. Bachlava E, Dewey RE, Burton JW, Cardinal AJ. Mapping and Comparison of Quantitative Trait Loci for Oleic Acid Seed Content in Two Segregating Soybean Populations. Crop Sci. 2009;49(2):433–42.
- View Article
- Google Scholar
90. Zhang W-K, Wang Y-J, Luo G-Z, Zhang J-S, He C-Y, Wu X-L, et al. QTL mapping of ten agronomic traits on the soybean ( Glycine max L. Merr.) genetic map and their association with EST markers. Theor Appl Genet. 2004;108(6):1131–9. pmid:15067400
- View Article
- PubMed/NCBI
- Google Scholar
91. Han Y, Li D, Zhu D, Li H, Li X, Teng W, et al. QTL analysis of soybean seed weight across multi-genetic backgrounds and environments. Theor Appl Genet. 2012;125(4):671–83. pmid:22481120
- View Article
- PubMed/NCBI
- Google Scholar
92. Li X, Han Y, Teng W, Zhang S, Yu K, Poysa V. Pyramided QTL underlying tolerance to Phytophthora root rot in mega-environments from soybean cultivars ‘Conrad’ and ‘Hefeng 25’. Theoret Appl Genet. 2010;121:651–8.
- View Article
- Google Scholar
93. Mao T, Jiang Z, Han Y, Teng W, Zhao X, Li W. Identification of quantitative trait loci underlying seed protein and oil contents of soybean across multi-genetic backgrounds and environments. Plant Breed. 2013;132(6):630–41.
- View Article
- Google Scholar
94. Shibata M, Takayama K, Ujiie A, Yamada T, Abe J, Kitamura K. Genetic relationship between lipid content and linolenic acid concentration in soybean seeds. Breed Sci. 2008;58(4):361–6.
- View Article
- Google Scholar
95. Yao D, Liu ZZ, Zhang J, Liu SY, Qu J, Guan SY, et al. Analysis of quantitative trait loci for main plant traits in soybean. Genet Mol Res. 2015;14(2):6101–9. pmid:26125811
- View Article
- PubMed/NCBI
- Google Scholar
96. Brummer EC, Graef GL, Orf J, Wilcox JR, Shoemaker RC. Mapping QTL for seed protein and oil content in eight soybean populations. Crop Sci. 1997;37(2):370–8.
- View Article
- Google Scholar
97. Wang D, Graef GL, Procopiuk AM, Diers BW. Identification of putative QTL that underlie yield in interspecific soybean backcross populations. Theor Appl Genet. 2004;108(3):458–67. pmid:14504749
- View Article
- PubMed/NCBI
- Google Scholar
98. Tasma IM, Lorenzen LL, Green DE, Shoemaker RC. Mapping genetic loci for flowering time, maturity, and photoperiod insensitivity in soybean. Mol Breed. 2001;8:25–35.
- View Article
- Google Scholar
99. Wang X, Jiang G-L, Green M, Scott RA, Song Q, Hyten DL, et al. Identification and validation of quantitative trait loci for seed yield, oil and protein contents in two recombinant inbred line populations of soybean. Mol Genet Genomics. 2014;289(5):935–49. pmid:24861102
- View Article
- PubMed/NCBI
- Google Scholar
100. Ning H, Yuan J, Dong Q, Li W, Xue H, Wang Y, et al. Identification of QTLs related to the vertical distribution and seed-set of pod number in soybean [Glycine max (L.) Merri]. PLoS One. 2018;13(4):e0195830. pmid:29664958
- View Article
- PubMed/NCBI
- Google Scholar
101. Eskandari M, Cober ER, Rajcan I. Genetic control of soybean seed oil: I. QTL and genes associated with seed oil concentration in RIL populations derived from crossing moderately high-oil parents. Theor Appl Genet. 2013;126(2):483–95. pmid:23192670
- View Article
- PubMed/NCBI
- Google Scholar
102. Komatsu K, Okuda S, Takahashi M, Matsunaga R, Nakazawa Y. Quantitative trait loci mapping of pubescence density and flowering time of insect-resistant soybean (Glycine max L. Merr.). Genet Mol Biol. 2007;30:635–9.
- View Article
- Google Scholar
103. Byth DE, Caldwell BE, Weber CR. Specific and non-specific index selection in soybeans, Glycine max L. (Merrill). Crop Sci. 1969;9(6):702–5.
- View Article
- Google Scholar
104. Simpson AM, Wilcox JR. Genetic and phenotypic associations of agronomic characteristics in four high protein soybean populations1. Crop Sci. 1983;23(6):1077–81.
- View Article
- Google Scholar
105. Wilcox JR, Cavins JF. Backcrossing high seed protein to a soybean cultivar. Crop Sci. 1995;35(4):1036–41.
- View Article
- Google Scholar
106. Cober ER, D Voldeng H. Developing High‐Protein, High‐Yield Soybean Populations and Lines. Crop Sci. 2000;40(1):39–42.
- View Article
- Google Scholar
107. Chung J, Babka H, Graef G, Staswick P, Lee D, Cregan P. The seed protein, oil, and yield QTL on soybean linkage group I. Crop Sci. 2003;43(3):1053–67.
- View Article
- Google Scholar
108. Xu S. Theoretical basis of the Beavis effect. Genetics. 2003;165(4):2259–68. pmid:14704201
- View Article
- PubMed/NCBI
- Google Scholar
109. Xu Z, Zou F, Vision TJ. Improving quantitative trait loci mapping resolution in experimental crosses by the use of genotypically selected samples. Genetics. 2005;170(1):401–8. pmid:15781710
- View Article
- PubMed/NCBI
- Google Scholar

[ref1] 1. Fuller DQ. Contrasting patterns in crop domestication and domestication rates: recent archaeobotanical insights from the Old World. Ann Bot. 2007;100(5):903–24. pmid:17495986
View Article
PubMed/NCBI
Google Scholar

[2] View Article

[3] PubMed/NCBI

[4] Google Scholar

[ref2] 2. Petersen M, Sander L, Child R, van Onckelen H, Ulvskov P, Borkhardt B. Isolation and characterisation of a pod dehiscence zone-specific polygalacturonase from Brassica napus. Plant Mol Biol. 1996;31(3):517–27. pmid:8790285
View Article
PubMed/NCBI
Google Scholar

[6] View Article

[7] PubMed/NCBI

[8] Google Scholar

[ref3] 3. Meyer RS, Purugganan MD. Evolution of crop species: genetics of domestication and diversification. Nat Rev Genet. 2013;14(12):840–52. pmid:24240513
View Article
PubMed/NCBI
Google Scholar

[10] View Article

[11] PubMed/NCBI

[12] Google Scholar

[ref4] 4. Di Vittori V, Gioia T, Rodriguez M, Bellucci E, Bitocchi E, Nanni L, et al. Convergent evolution of the seed shattering trait. Genes (Basel). 2019;10(1):68. pmid:30669440
View Article
PubMed/NCBI
Google Scholar

[14] View Article

[15] PubMed/NCBI

[16] Google Scholar

[ref5] 5. Anderson E, Ali M, Beavis W, Chen P, Clemente T, Diers B. Soybean [Glycine max (L.) Merr.] Breeding: History, Improvement, Production and Future Opportunities. Adv Plant Breed Strateg Legumes. 2019;431–516.
View Article
Google Scholar

[18] View Article

[19] Google Scholar

[ref6] 6. Tukamuhabwa P, Rubaihayo P, Dashiell KE. Euphytica. 2002;125(1):29–34.
View Article
Google Scholar

[21] View Article

[22] Google Scholar

[ref7] 7. Bara N, Khare D, Shrivastava AN. Studies on the factors affecting pod shattering in soybean. Ind Jrnl Gen Plnt Bree. 2013;73(3):270.
View Article
Google Scholar

[24] View Article

[25] Google Scholar

[ref8] 8. Caviness CE. Heritability of pod dehiscence and its association with some agronomic characters in soybean. Crop Sci. 1969;9(2):207–9.
View Article
Google Scholar

[27] View Article

[28] Google Scholar

[ref9] 9. Tsuchiya T. Physiological and genetic analysis of pod shattering in soybean. Jpn Agric Res Q. 1987;21(3):166–75.
View Article
Google Scholar

[30] View Article

[31] Google Scholar

[ref10] 10. Bailey M, Mian M, Carter Jr T, Ashley D, Boerma H. Pod dehiscence of soybean: identification of quantitative trait loci. J Hered. 1997;88(2):152–4.
View Article
Google Scholar

[33] View Article

[34] Google Scholar

[ref11] 11. Fehr WR, Caviness CE, Burmood DT, Pennington JS. Stage of Development Descriptions for Soybeans, Glycine Max (L.) Merrill1. Crop Sci. 1971;11(6):929–31.
View Article
Google Scholar

[36] View Article

[37] Google Scholar

[ref12] 12. Lu S, Zhao X, Hu Y, Liu S, Nan H, Li X, et al. Natural variation at the soybean J locus improves adaptation to the tropics and enhances yield. Nat Genet. 2017;49(5):773–9. pmid:28319089
View Article
PubMed/NCBI
Google Scholar

[39] View Article

[40] PubMed/NCBI

[41] Google Scholar

[ref13] 13. Franca-Neto J, Henning A, Krzyzanowski F. Tropical soybean: improvement and production. Seed Production and Technology for the Tropics. FAO Plant Production and Protection Series. 1994;217–40.
View Article
Google Scholar

[43] View Article

[44] Google Scholar

[ref14] 14. Funatsuki H, Ishimoto M, Tsuji H, Kawaguchi K, Hajika M, Fujino K. Simple sequence repeat markers linked to a major QTL controlling pod shattering in soybean. Plant Breed. 2006;125(2):195–7.
View Article
Google Scholar

[46] View Article

[47] Google Scholar

[ref15] 15. Liu B, Fujita T, Yan Z-H, Sakamoto S, Xu D, Abe J. QTL mapping of domestication-related traits in soybean (Glycine max). Ann Bot. 2007;100(5):1027–38. pmid:17684023
View Article
PubMed/NCBI
Google Scholar

[49] View Article

[50] PubMed/NCBI

[51] Google Scholar

[ref16] 16. Kang S, Kwak M, Kim H, Choung M, Han W, Baek I. Population-specific QTLs and their different epistatic interactions for pod dehiscence in soybean [Glycine max (L.) Merr.]. Euphytica. 2009;166:15–24.
View Article
Google Scholar

[53] View Article

[54] Google Scholar

[ref17] 17. Dong Y, Yang X, Liu J, Wang B-H, Liu B-L, Wang Y-Z. Pod shattering resistance associated with domestication is mediated by a NAC gene in soybean. Nat Commun. 2014;5:3352. pmid:24549030
View Article
PubMed/NCBI
Google Scholar

[56] View Article

[57] PubMed/NCBI

[58] Google Scholar

[ref18] 18. Funatsuki H, Hajika M, Hagihara S, Yamada T, Tanaka Y, Tsuji H. Confirmation of the location and the effects of a major QTL controlling pod dehiscence, qPDH1, in soybean. Breeding Sci. 2008;58(1):63–9.
View Article
Google Scholar

[60] View Article

[61] Google Scholar

[ref19] 19. Yamada T, Funatsuki H, Hagihara S, Fujita S, Tanaka Y, Tsuji H. A major QTL, qPDH1, is commonly involved in shattering resistance of soybean cultivars. Breeding Sci. 2009;59(4):435–40.
View Article
Google Scholar

[63] View Article

[64] Google Scholar

[ref20] 20. Funatsuki H, Suzuki M, Hirose A, Inaba H, Yamada T, Hajika M, et al. Molecular basis of a shattering resistance boosting global dissemination of soybean. Proc Natl Acad Sci U S A. 2014;111(50):17797–802. pmid:25468966
View Article
PubMed/NCBI
Google Scholar

[66] View Article

[67] PubMed/NCBI

[68] Google Scholar

[ref21] 21. Zhang J, Singh AK. Genetic control and geo-climate adaptation of pod dehiscence provide novel insights into soybean domestication. G3 (Bethesda). 2020;10(2):545–54. pmid:31836621
View Article
PubMed/NCBI
Google Scholar

[70] View Article

[71] PubMed/NCBI

[72] Google Scholar

[ref22] 22. Thakare DS, Chimote VP, Rokade RC, Deshmukh MP. Correlation of yield and associated with other characters in pod shattering soybean (Glycine max (L.) Merrill). Int J Trop Agric. 2016;34(6):1915–23.
View Article
Google Scholar

[74] View Article

[75] Google Scholar

[ref23] 23. Kataliko R, Kimani P, Muthomi J, Wanderi W, Olubayo F, Nzuve F. Resistance and correlation of pod shattering and selected agronomic traits in soybeans. J lant Stud. 2019;8(2):39–48.
View Article
Google Scholar

[77] View Article

[78] Google Scholar

[ref24] 24. Philbrook BD, Oplinger ES. Soybean field losses as influenced by harvest delays. Agron J. 1989;81(2):251–8.
View Article
Google Scholar

[80] View Article

[81] Google Scholar

[ref25] 25. Tiwari S, Bhatnagar P. Pod shattering as related to other agronomic attributes in soybean. Trop Agric. 1991;68(1):102–3.
View Article
Google Scholar

[83] View Article

[84] Google Scholar

[ref26] 26. Bhor T, Chimote V, Deshmukh M. Inheritance of pod shattering in soybean [Glycine max (L.) Merrill]. Electron J Plant Breed. 2014;5(4):671–6.
View Article
Google Scholar

[86] View Article

[87] Google Scholar

[ref27] 27. Bernard R, Cremeens C, Cooper R, Collins F, Krober O, Athow K, et al. Evaluation of the USDA soybean germplasm collection: maturity groups 000-IV (FC 01.547-PI 266.807). Tech Bull USDA. 1998;1844:1–169.
View Article
Google Scholar

[89] View Article

[90] Google Scholar

[ref28] 28. Hill J, Peregrine E, Sprau G, Cremeens C, Nelson R, Orf J. Evaluation of the USDA soybean germplasm collection: maturity groups 000-IV (PI 578371-PI 612761). Technical Bulletin USDA. 20081–151.
View Article
Google Scholar

[92] View Article

[93] Google Scholar

[ref29] 29. Peregrine EK, Sprau GL, Cremeens CR, Handly P, Kilen TC, Smith JR. Evaluation of the USDA soybean germplasm collection: maturity group V (FC 30265-PI 612614) and maturity groups VI-VIII (PI 416758-PI 606432B). Techn Bull USDA. 1920;1–369.
View Article
Google Scholar

[95] View Article

[96] Google Scholar

[ref30] 30. Zhang LX, Kyei-Boahen S, Zhang J, Zhang MH, Freeland TB, Watson CE Jr, et al. Modifications of Optimum Adaptation Zones for Soybean Maturity Groups in the USA. Crop Manag. 2007;6(1):1–11.
View Article
Google Scholar

[98] View Article

[99] Google Scholar

[ref31] 31. Freeman MF, Tukey JW. Transformations Related to the Angular and the Square Root. Ann Math Statist. 1950;21(4):607–11.
View Article
Google Scholar

[101] View Article

[102] Google Scholar

[ref32] 32. Song Q, Hyten DL, Jia G, Quigley CV, Fickus EW, Nelson RL, et al. Fingerprinting soybean germplasm and its utility in genomic research. G3 (Bethesda). 2015;5(10):1999–2006. pmid:26224783
View Article
PubMed/NCBI
Google Scholar

[104] View Article

[105] PubMed/NCBI

[106] Google Scholar

[ref33] 33. Grant D, Nelson RT, Cannon SB, Shoemaker RC. SoyBase, the USDA-ARS soybean genetics and genomics database. Nucleic Acids Res. 2010;38(Database issue):D843-6. pmid:20008513
View Article
PubMed/NCBI
Google Scholar

[108] View Article

[109] PubMed/NCBI

[110] Google Scholar

[ref34] 34. Schmutz J, Cannon SB, Schlueter J, Ma J, Mitros T, Nelson W, et al. Genome sequence of the palaeopolyploid soybean. Nature. 2010;463(7278):178–83. pmid:20075913
View Article
PubMed/NCBI
Google Scholar

[112] View Article

[113] PubMed/NCBI

[114] Google Scholar

[ref35] 35. Money D, Gardner K, Migicovsky Z, Schwaninger H, Zhong G-Y, Myles S. LinkImpute: fast and accurate genotype imputation for nonmodel organisms. G3 (Bethesda). 2015;5(11):2383–90. pmid:26377960
View Article
PubMed/NCBI
Google Scholar

[116] View Article

[117] PubMed/NCBI

[118] Google Scholar

[ref36] 36. Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 2007;23(19):2633–5. pmid:17586829
View Article
PubMed/NCBI
Google Scholar

[120] View Article

[121] PubMed/NCBI

[122] Google Scholar

[ref37] 37. Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 1989;123(3):585–95. pmid:2513255
View Article
PubMed/NCBI
Google Scholar

[124] View Article

[125] PubMed/NCBI

[126] Google Scholar

[ref38] 38. Kallenberg W, Oosterhoff J, Schriever B. The number of classes in chi-squared goodness-of-fit tests. J Am Stat Assoc. 1985;80(392):959–68.
View Article
Google Scholar

[128] View Article

[129] Google Scholar

[ref39] 39. Falconer D, Mackay T. Introduction to Quantitative Genetics. 4th ed. Harlow, UK: Longman Scientific & Technical;1996;57–64.

[ref40] 40. Xu X, Liu X, Ge S, Jensen JD, Hu F, Li X, et al. Resequencing 50 accessions of cultivated and wild rice yields markers for identifying agronomically important genes. Nat Biotechnol. 2011;30(1):105–11. pmid:22158310
View Article
PubMed/NCBI
Google Scholar

[132] View Article

[133] PubMed/NCBI

[134] Google Scholar

[ref41] 41. Varshney RK, Song C, Saxena RK, Azam S, Yu S, Sharpe AG, et al. Draft genome sequence of chickpea (Cicer arietinum) provides a resource for trait improvement. Nat Biotechnol. 2013;31(3):240–6. pmid:23354103
View Article
PubMed/NCBI
Google Scholar

[136] View Article

[137] PubMed/NCBI

[138] Google Scholar

[ref42] 42. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Stat Method. 1995;57(1):289–300.
View Article
Google Scholar

[140] View Article

[141] Google Scholar

[ref43] 43. Wright S. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution. 1965;19(3):395–420.
View Article
Google Scholar

[143] View Article

[144] Google Scholar

[ref44] 44. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. the gene ontology consortium. Nat Genet. 2000;25(1):25–9. pmid:10802651
View Article
PubMed/NCBI
Google Scholar

[146] View Article

[147] PubMed/NCBI

[148] Google Scholar

[ref45] 45. Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164(4):1567–87. pmid:12930761
View Article
PubMed/NCBI
Google Scholar

[150] View Article

[151] PubMed/NCBI

[152] Google Scholar

[ref46] 46. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics. 2000;155(2):945–59. pmid:10835412
View Article
PubMed/NCBI
Google Scholar

[154] View Article

[155] PubMed/NCBI

[156] Google Scholar

[ref47] 47. Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14(8):2611–20. pmid:15969739
View Article
PubMed/NCBI
Google Scholar

[158] View Article

[159] PubMed/NCBI

[160] Google Scholar

[ref48] 48. Rosenberg NA, Burke T, Elo K, Feldman MW, Freidlin PJ, Groenen MA, et al. Empirical evaluation of genetic clustering methods using multilocus genotypes from 20 chicken breeds. Genetics. 2001;159(2):699–713. pmid:11606545
View Article
PubMed/NCBI
Google Scholar

[162] View Article

[163] PubMed/NCBI

[164] Google Scholar

[ref49] 49. Lewontin RC. The interaction of selection and linkage. i. general considerations; heterotic models. Genetics. 1964;49(1):49–67. pmid:17248194
View Article
PubMed/NCBI
Google Scholar

[166] View Article

[167] PubMed/NCBI

[168] Google Scholar

[ref50] 50. Troyanskaya O, Cantor M, Sherlock G, Brown P, Hastie T, Tibshirani R, et al. Missing value estimation methods for DNA microarrays. Bioinformatics. 2001;17(6):520–5. pmid:11395428
View Article
PubMed/NCBI
Google Scholar

[170] View Article

[171] PubMed/NCBI

[172] Google Scholar

[ref51] 51. Endelman JB, Jannink J-L. Shrinkage estimation of the realized relationship matrix. G3 (Bethesda). 2012;2(11):1405–13. pmid:23173092
View Article
PubMed/NCBI
Google Scholar

[174] View Article

[175] PubMed/NCBI

[176] Google Scholar

[ref52] 52. Lipka AE, Tian F, Wang Q, Peiffer J, Li M, Bradbury PJ, et al. GAPIT: genome association and prediction integrated tool. Bioinformatics. 2012;28(18):2397–9. pmid:22796960
View Article
PubMed/NCBI
Google Scholar

[178] View Article

[179] PubMed/NCBI

[180] Google Scholar

[ref53] 53. R Core Team. R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria. 2013. Available from: http://www.R-project.org/.
View Article
Google Scholar

[182] View Article

[183] Google Scholar

[ref54] 54. VanRaden PM. Efficient methods to compute genomic predictions. J Dairy Sci. 2008;91(11):4414–23. pmid:18946147
View Article
PubMed/NCBI
Google Scholar

[185] View Article

[186] PubMed/NCBI

[187] Google Scholar

[ref55] 55. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9. pmid:16862161
View Article
PubMed/NCBI
Google Scholar

[189] View Article

[190] PubMed/NCBI

[191] Google Scholar

[ref56] 56. Yu J, Pressoir G, Briggs W, Vroh BI, Yamasaki M, Doebley J. A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet. 2006;38(2):203–8.
View Article
Google Scholar

[193] View Article

[194] Google Scholar

[ref57] 57. Wang Q, Tian F, Pan Y, Buckler ES, Zhang Z. A SUPER powerful method for genome wide association study. PLoS One. 2014;9(9):e107684. pmid:25247812
View Article
PubMed/NCBI
Google Scholar

[196] View Article

[197] PubMed/NCBI

[198] Google Scholar

[ref58] 58. Segura V, Vilhjálmsson BJ, Platt A, Korte A, Seren Ü, Long Q, et al. An efficient multi-locus mixed-model approach for genome-wide association studies in structured populations. Nat Genet. 2012;44(7):825–30. pmid:22706313
View Article
PubMed/NCBI
Google Scholar

[200] View Article

[201] PubMed/NCBI

[202] Google Scholar

[ref59] 59. Liu X, Huang M, Fan B, Buckler ES, Zhang Z. Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLoS Genet. 2016;12(2):e1005767. pmid:26828793
View Article
PubMed/NCBI
Google Scholar

[204] View Article

[205] PubMed/NCBI

[206] Google Scholar

[ref60] 60. Huang M, Liu X, Zhou Y, Summers RM, Zhang Z. BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. Gigascience. 2019;8(2):giy154. pmid:30535326
View Article
PubMed/NCBI
Google Scholar

[208] View Article

[209] PubMed/NCBI

[210] Google Scholar

[ref61] 61. Hyten DL, Choi I, Song Q, Specht JE, Carter TE, Shoemaker RC, et al. A High density integrated genetic linkage map of soybean and the development of a 1536 universal soy linkage panel for quantitative trait locus mapping. Crop Sci. 2010;50(3):960–8.
View Article
Google Scholar

[212] View Article

[213] Google Scholar

[ref62] 62. Dupuis J, Siegmund D. Statistical methods for mapping quantitative trait loci from a dense set of markers. Genetics. 1999;151(1):373–86. pmid:9872974
View Article
PubMed/NCBI
Google Scholar

[215] View Article

[216] PubMed/NCBI

[217] Google Scholar

[ref63] 63. Li H. A quick method to calculate QTL confidence interval. J Genet. 2011;90(2):355–60. pmid:21869489
View Article
PubMed/NCBI
Google Scholar

[219] View Article

[220] PubMed/NCBI

[221] Google Scholar

[ref64] 64. Guo B, Sleper D, Lu P, Shannon J, Nguyen H, Arelli P. QTLs associated with resistance to soybean cyst nematode in soybean: Meta‐analysis of QTL locations. Crop Sci. 2006;46(2):595–602.
View Article
Google Scholar

[223] View Article

[224] Google Scholar

[ref65] 65. Hwang S, King CA, Ray JD, Cregan PB, Chen P, Carter TE Jr, et al. Confirmation of delayed canopy wilting QTLs from multiple soybean mapping populations. Theor Appl Genet. 2015;128(10):2047–65. pmid:26163767
View Article
PubMed/NCBI
Google Scholar

[226] View Article

[227] PubMed/NCBI

[228] Google Scholar

[ref66] 66. Zhang Z, Todhunter RJ, Buckler ES, Van Vleck LD. Technical note: Use of marker-based relationships with multiple-trait derivative-free restricted maximal likelihood. J Anim Sci. 2007;85(4):881–5. pmid:17085728
View Article
PubMed/NCBI
Google Scholar

[230] View Article

[231] PubMed/NCBI

[232] Google Scholar

[ref67] 67. Wang J, Zhou Z, Zhang Z, Li H, Liu D, Zhang Q, et al. Expanding the BLUP alphabet for genomic prediction adaptable to the genetic architectures of complex traits. Heredity (Edinb). 2018;121(6):648–62. pmid:29765161
View Article
PubMed/NCBI
Google Scholar

[234] View Article

[235] PubMed/NCBI

[236] Google Scholar

[ref68] 68. Sedgwick P. Pearson’s correlation coefficient. Br Med J. 2012;345:e4483.
View Article
Google Scholar

[238] View Article

[239] Google Scholar

[ref69] 69. Ruxton GD. The unequal variance t-test is an underused alternative to Student’s t-test and the Mann–Whitney U test. Behavioral Ecology. 2006;17(4):688–90.
View Article
Google Scholar

[241] View Article

[242] Google Scholar

[ref70] 70. Duncan DB, Godbold JH. Approximate k-Ratio t Tests for Differences between Unequally Replicated Treatments. Biometrics. 1979;35(4):749.
View Article
Google Scholar

[244] View Article

[245] Google Scholar

[ref71] 71. Tenhaken R, Doerks T, Bork P. DCD - a novel plant specific domain in proteins involved in development and programmed cell death. BMC Bioinformatics. 2005;6:169. pmid:16008837
View Article
PubMed/NCBI
Google Scholar

[247] View Article

[248] PubMed/NCBI

[249] Google Scholar

[ref72] 72. Wang P, Guo L, Morgan J, Dudareva N, Chapple C. Transcript and metabolite network perturbations in lignin biosynthetic mutants of Arabidopsis. Plant Physiol. 2022;190(4):2828–46. pmid:35880844
View Article
PubMed/NCBI
Google Scholar

[251] View Article

[252] PubMed/NCBI

[253] Google Scholar

[ref73] 73. Turlapati PV, Kim K-W, Davin LB, Lewis NG. The laccase multigene family in Arabidopsis thaliana: towards addressing the mystery of their gene function(s). Planta. 2011;233(3):439–70. pmid:21063888
View Article
PubMed/NCBI
Google Scholar

[255] View Article

[256] PubMed/NCBI

[257] Google Scholar

[ref74] 74. Okushima Y, Mitina I, Quach HL, Theologis A. AUXIN RESPONSE FACTOR 2 (ARF2): a pleiotropic developmental regulator. Plant J. 2005;43(1):29–46. pmid:15960614
View Article
PubMed/NCBI
Google Scholar

[259] View Article

[260] PubMed/NCBI

[261] Google Scholar

[ref75] 75. He H, Bai M, Tong P, Hu Y, Yang M, Wu H. CELLULASE6 and MANNANASE7 affect cell differentiation and silique dehiscence. Plant Physiol. 2018;176(3):2186–201. pmid:29348141
View Article
PubMed/NCBI
Google Scholar

[263] View Article

[264] PubMed/NCBI

[265] Google Scholar

[ref76] 76. Kumar M, Mishra L, Carr P, Pilling M, Gardner P, Mansfield SD, et al. Exploiting CELLULOSE SYNTHASE (CESA) class specificity to probe cellulose microfibril biosynthesis. Plant Physiol. 2018;177(1):151–67. pmid:29523715
View Article
PubMed/NCBI
Google Scholar

[267] View Article

[268] PubMed/NCBI

[269] Google Scholar

[ref77] 77. Lionetti V, Fabri E, De Caroli M, Hansen AR, Willats WGT, Piro G, et al. Three pectin methylesterase inhibitors protect cell wall integrity for arabidopsis immunity to botrytis. Plant Physiol. 2017;173(3):1844–63. pmid:28082716
View Article
PubMed/NCBI
Google Scholar

[271] View Article

[272] PubMed/NCBI

[273] Google Scholar

[ref78] 78. Voxeur A, André A, Breton C, Lerouge P. Identification of putative rhamnogalacturonan-II specific glycosyltransferases in Arabidopsis using a combination of bioinformatics approaches. PLoS One. 2012;7(12):e51129. pmid:23272088
View Article
PubMed/NCBI
Google Scholar

[275] View Article

[276] PubMed/NCBI

[277] Google Scholar

[ref79] 79. Yuan Y, Teng Q, Lee C, Zhong R, Ye Z-H. Modification of the degree of 4-O-methylation of secondary wall glucuronoxylan. Plant Sci. 2014;219–220:42–50. pmid:24576763
View Article
PubMed/NCBI
Google Scholar

[279] View Article

[280] PubMed/NCBI

[281] Google Scholar

[ref80] 80. Jaradat MR, Ruegger M, Bowling A, Butler H, Cutler AJ. A comprehensive transcriptome analysis of silique development and dehiscence in Arabidopsis and Brassica integrating genotypic, interspecies and developmental comparisons. GM Crops Food. 2014;5(4):302–20. pmid:25523176
View Article
PubMed/NCBI
Google Scholar

[283] View Article

[284] PubMed/NCBI

[285] Google Scholar

[ref81] 81. Taylor-Teeples M, Lin L, de Lucas M, Turco G, Toal TW, Gaudinier A, et al. An Arabidopsis gene regulatory network for secondary cell wall synthesis. Nature. 2015;517(7536):571–5. pmid:25533953
View Article
PubMed/NCBI
Google Scholar

[287] View Article

[288] PubMed/NCBI

[289] Google Scholar

[ref82] 82. Zhong R, Lee C, Haghighat M, Ye Z-H. Xylem vessel-specific SND5 and its homologs regulate secondary wall biosynthesis through activating secondary wall NAC binding elements. New Phytol. 2021;231(4):1496–509. pmid:33908063
View Article
PubMed/NCBI
Google Scholar

[291] View Article

[292] PubMed/NCBI

[293] Google Scholar

[ref83] 83. Kim W-C, Kim J-Y, Ko J-H, Kang H, Han K-H. Identification of direct targets of transcription factor MYB46 provides insights into the transcriptional regulation of secondary wall biosynthesis. Plant Mol Biol. 2014;85(6):589–99. pmid:24879533
View Article
PubMed/NCBI
Google Scholar

[295] View Article

[296] PubMed/NCBI

[297] Google Scholar

[ref84] 84. Wang J, Kucukoglu M, Zhang L, Chen P, Decker D, Nilsson O, et al. The Arabidopsis LRR-RLK, PXC1, is a regulator of secondary wall formation correlated with the TDIF-PXY/TDR-WOX4 signaling pathway. BMC Plant Biol. 2013;13:94. pmid:23815750
View Article
PubMed/NCBI
Google Scholar

[299] View Article

[300] PubMed/NCBI

[301] Google Scholar

[ref85] 85. Khan NA, Githiri SM, Benitez ER, Abe J, Kawasaki S, Hayashi T, et al. QTL analysis of cleistogamy in soybean. Theor Appl Genet. 2008;117(4):479–87. pmid:18506418
View Article
PubMed/NCBI
Google Scholar

[303] View Article

[304] PubMed/NCBI

[305] Google Scholar

[ref86] 86. Han Y, Teng W, Wang Y, Zhao X, Wu L, Li D, et al. Unconditional and conditional QTL underlying the genetic interrelationships between soybean seed isoflavone, and protein or oil contents. Plant Breeding. 2015;134(3):300–9.
View Article
Google Scholar

[307] View Article

[308] Google Scholar

[ref87] 87. Sun Y-N, Pan J-B, Shi X-L, Du X-Y, Wu Q, Qi Z-M, et al. Multi-environment mapping and meta-analysis of 100-seed weight in soybean. Mol Biol Rep. 2012;39(10):9435–43. pmid:22740134
View Article
PubMed/NCBI
Google Scholar

[310] View Article

[311] PubMed/NCBI

[312] Google Scholar

[ref88] 88. Qi Z, Wu Q, Han X, Sun Y, Du X, Liu C, et al. Soybean oil content QTL mapping and integrating with meta-analysis method for mining genes. Euphytica. 2011;179(3):499–514.
View Article
Google Scholar

[314] View Article

[315] Google Scholar

[ref89] 89. Bachlava E, Dewey RE, Burton JW, Cardinal AJ. Mapping and Comparison of Quantitative Trait Loci for Oleic Acid Seed Content in Two Segregating Soybean Populations. Crop Sci. 2009;49(2):433–42.
View Article
Google Scholar

[317] View Article

[318] Google Scholar

[ref90] 90. Zhang W-K, Wang Y-J, Luo G-Z, Zhang J-S, He C-Y, Wu X-L, et al. QTL mapping of ten agronomic traits on the soybean ( Glycine max L. Merr.) genetic map and their association with EST markers. Theor Appl Genet. 2004;108(6):1131–9. pmid:15067400
View Article
PubMed/NCBI
Google Scholar

[320] View Article

[321] PubMed/NCBI

[322] Google Scholar

[ref91] 91. Han Y, Li D, Zhu D, Li H, Li X, Teng W, et al. QTL analysis of soybean seed weight across multi-genetic backgrounds and environments. Theor Appl Genet. 2012;125(4):671–83. pmid:22481120
View Article
PubMed/NCBI
Google Scholar

[324] View Article

[325] PubMed/NCBI

[326] Google Scholar

[ref92] 92. Li X, Han Y, Teng W, Zhang S, Yu K, Poysa V. Pyramided QTL underlying tolerance to Phytophthora root rot in mega-environments from soybean cultivars ‘Conrad’ and ‘Hefeng 25’. Theoret Appl Genet. 2010;121:651–8.
View Article
Google Scholar

[328] View Article

[329] Google Scholar

[ref93] 93. Mao T, Jiang Z, Han Y, Teng W, Zhao X, Li W. Identification of quantitative trait loci underlying seed protein and oil contents of soybean across multi-genetic backgrounds and environments. Plant Breed. 2013;132(6):630–41.
View Article
Google Scholar

[331] View Article

[332] Google Scholar

[ref94] 94. Shibata M, Takayama K, Ujiie A, Yamada T, Abe J, Kitamura K. Genetic relationship between lipid content and linolenic acid concentration in soybean seeds. Breed Sci. 2008;58(4):361–6.
View Article
Google Scholar

[334] View Article

[335] Google Scholar

[ref95] 95. Yao D, Liu ZZ, Zhang J, Liu SY, Qu J, Guan SY, et al. Analysis of quantitative trait loci for main plant traits in soybean. Genet Mol Res. 2015;14(2):6101–9. pmid:26125811
View Article
PubMed/NCBI
Google Scholar

[337] View Article

[338] PubMed/NCBI

[339] Google Scholar

[ref96] 96. Brummer EC, Graef GL, Orf J, Wilcox JR, Shoemaker RC. Mapping QTL for seed protein and oil content in eight soybean populations. Crop Sci. 1997;37(2):370–8.
View Article
Google Scholar

[341] View Article

[342] Google Scholar

[ref97] 97. Wang D, Graef GL, Procopiuk AM, Diers BW. Identification of putative QTL that underlie yield in interspecific soybean backcross populations. Theor Appl Genet. 2004;108(3):458–67. pmid:14504749
View Article
PubMed/NCBI
Google Scholar

[344] View Article

[345] PubMed/NCBI

[346] Google Scholar

[ref98] 98. Tasma IM, Lorenzen LL, Green DE, Shoemaker RC. Mapping genetic loci for flowering time, maturity, and photoperiod insensitivity in soybean. Mol Breed. 2001;8:25–35.
View Article
Google Scholar

[348] View Article

[349] Google Scholar

[ref99] 99. Wang X, Jiang G-L, Green M, Scott RA, Song Q, Hyten DL, et al. Identification and validation of quantitative trait loci for seed yield, oil and protein contents in two recombinant inbred line populations of soybean. Mol Genet Genomics. 2014;289(5):935–49. pmid:24861102
View Article
PubMed/NCBI
Google Scholar

[351] View Article

[352] PubMed/NCBI

[353] Google Scholar

[ref100] 100. Ning H, Yuan J, Dong Q, Li W, Xue H, Wang Y, et al. Identification of QTLs related to the vertical distribution and seed-set of pod number in soybean [Glycine max (L.) Merri]. PLoS One. 2018;13(4):e0195830. pmid:29664958
View Article
PubMed/NCBI
Google Scholar

[355] View Article

[356] PubMed/NCBI

[357] Google Scholar

[ref101] 101. Eskandari M, Cober ER, Rajcan I. Genetic control of soybean seed oil: I. QTL and genes associated with seed oil concentration in RIL populations derived from crossing moderately high-oil parents. Theor Appl Genet. 2013;126(2):483–95. pmid:23192670
View Article
PubMed/NCBI
Google Scholar

[359] View Article

[360] PubMed/NCBI

[361] Google Scholar

[ref102] 102. Komatsu K, Okuda S, Takahashi M, Matsunaga R, Nakazawa Y. Quantitative trait loci mapping of pubescence density and flowering time of insect-resistant soybean (Glycine max L. Merr.). Genet Mol Biol. 2007;30:635–9.
View Article
Google Scholar

[363] View Article

[364] Google Scholar

[ref103] 103. Byth DE, Caldwell BE, Weber CR. Specific and non-specific index selection in soybeans, Glycine max L. (Merrill). Crop Sci. 1969;9(6):702–5.
View Article
Google Scholar

[366] View Article

[367] Google Scholar

[ref104] 104. Simpson AM, Wilcox JR. Genetic and phenotypic associations of agronomic characteristics in four high protein soybean populations1. Crop Sci. 1983;23(6):1077–81.
View Article
Google Scholar

[369] View Article

[370] Google Scholar

[ref105] 105. Wilcox JR, Cavins JF. Backcrossing high seed protein to a soybean cultivar. Crop Sci. 1995;35(4):1036–41.
View Article
Google Scholar

[372] View Article

[373] Google Scholar

[ref106] 106. Cober ER, D Voldeng H. Developing High‐Protein, High‐Yield Soybean Populations and Lines. Crop Sci. 2000;40(1):39–42.
View Article
Google Scholar

[375] View Article

[376] Google Scholar

[ref107] 107. Chung J, Babka H, Graef G, Staswick P, Lee D, Cregan P. The seed protein, oil, and yield QTL on soybean linkage group I. Crop Sci. 2003;43(3):1053–67.
View Article
Google Scholar

[378] View Article

[379] Google Scholar

[ref108] 108. Xu S. Theoretical basis of the Beavis effect. Genetics. 2003;165(4):2259–68. pmid:14704201
View Article
PubMed/NCBI
Google Scholar

[381] View Article

[382] PubMed/NCBI

[383] Google Scholar

[ref109] 109. Xu Z, Zou F, Vision TJ. Improving quantitative trait loci mapping resolution in experimental crosses by the use of genotypically selected samples. Genetics. 2005;170(1):401–8. pmid:15781710
View Article
PubMed/NCBI
Google Scholar

[385] View Article

[386] PubMed/NCBI

[387] Google Scholar

Figures

Abstract

Introduction

Materials and methods

Public trait data

Public genotype data

Climate data

Putative domestication-related SNPs

Gene ontology analysis

Population structure, principal component, and linkage disequilibrium analyses

Genome-wide association study

QTL data mining and confirmation

Candidate gene and sequence analyses

Marker-assisted selection

Genomic selection

Statistical analysis

Results

Selection of G. max populations

Evaluation of two climate factors

Putative domestication-related SNPs

Gene ontology

Population structure and linkage disequilibrium

QTL analysis

Marker-assisted selection

Genomic selection

Discussion

Conclusion

Supporting information

S1 Fig. Statewide average temperature and precipitation from September to November, spanning the years 1895-2020 in the USA.

S2 Fig.

S3 Fig. GO analysis of 1,515 putative domesticated-related genes.

S4 Fig. Population structure in WS1179, 1IL64, 1IL66, MN0102, and MS967.

S5 Fig. QTL detected by multiple GWAS models in four G. max populations.

S1 Table. Information about twenty-four G. max populations for pod dehiscence study in the GRIN.

S2 Table. Information about 1,179 wild accessions in WS1179.

S3 Table. P-values for χ 2 goodness of fit tests and FST values in 4,280 SNPs between WS1179 and a G. max population.

S4 Table. Information about 86 QTLs identified in four G. max populations.

S5 Table. QTL information of one RFLP and eight SSRs from three RIL population studies.

S6 Table. Soybean candidate genes from 86 QTLs for pod dehiscence.

S7 Table. Favorable alleles of 86 QTLs in each G. max population.

S8 Table. GBV and PEV for each accession in 1IL64 using sBLUP.

S9 Table. Different agronomic traits associated with pod dehiscence.

Acknowledgments

References

S3 Table. P-values for χ ² goodness of fit tests and F_ST values in 4,280 SNPs between WS1179 and a G. max population.