Genome-wide association study and genome prediction of tallness trait in spinach tallness phenotyping

Genome-wide association study and genome prediction of tallness trait in spinach tallness phenotyping | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Genome-wide association study and genome prediction of tallness trait in spinach tallness phenotyping Ibtisam Alatawi, Haizheng Xiong, Hanan Alkabkabi, Kenani Chiwina, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6305818/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Plant height (tallness) is a crucial agronomic trait in spinach (Spinacia oleracea L.), affecting mechanical harvesting efficiency and overall yield. This study evaluated plant height variation in a panel of 307 USDA germplasm accessions, which were phenotyped for this trait and genotyped using 15,058 single-nucleotide polymorphisms (SNPs) from whole-genome resequencing. A genome-wide association study (GWAS) using GLM, MLM, FarmCPU, and BLINK in GAPIT3 identified six SNPs significantly associated with plant height: SOVchr2_68062488 (68,062,488 bp) on chromosome 2; SOVchr4_38323167 (38,323,167 bp) and SOVchr4_188084338 (188,084,338 bp) on chromosome 4; SOVchr5_70192260 (70,192,260 bp) on chromosome 5; and SOVchr6_8139833 (8,139,833 bp) and SOVchr6_91175684 (91,175,684 bp) on chromosome 6. Additionally, genomic prediction (GP) models estimated genomic estimated breeding values (GEBVs) for plant height, achieving r-values of 0.50 using GWAS-derived SNP markers in cross-population prediction. The integration of GWAS and GP provides valuable insights into the genetic architecture of plant height in spinach, supporting marker-assisted breeding strategies to enhance crop management and economic returns. Genome-wide association study (GWAS) genomic prediction (GP) plant height single-nucleotide polymorphism (SNP) Spinacia oleracea L. spinach tallness Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Spinach ( Spinacia oleracea L.) is a highly nutritious leafy vegetable, widely cultivated in the United States and globally (Shi et al., 2016 ; Tang et al., 2015 ). Its increasing demand is driven by consumer awareness of its rich nutritional profile, including essential vitamins, minerals, antioxidants, and bioactive compounds such as carotenoids and flavonoids (Frary et al., 2010 ; Rashid et al., 2020). Among its key agronomic traits, plant height plays a crucial role in spinach production and management (Jones et al., 2019 ). Taller spinach varieties are preferred for their compatibility with mechanical harvesting, which enhances operational efficiency, reduces labor costs, and improves economic returns (Li et al., 2020). However, optimizing plant height requires a balance, as taller plants must also resist lodging—a condition where plants collapse under adverse weather, leading to yield loss (Jones et al., 2019 ). Plant height in spinach, like in other major crops such as rice and maize, is a polygenic trait governed by multiple genetic factors (Huang & Han, 2016 ). Traditional quantitative trait loci (QTL) mapping approaches have been useful in identifying large-effect loci but often fail to detect small-effect loci that collectively influence complex traits (Yu & Buckler, 2006 ). This limitation underscores the need for genome-wide approaches such as genome-wide association studies (GWAS) and genomic selection (GS), which enable the identification of multiple loci contributing to plant height and improve breeding efficiency through genome-wide marker predictions. The substantial phenotypic variation observed in spinach plant height reflects its rich genetic diversity, making it an excellent candidate for advanced genomic studies and breeding efforts (Huang & Han, 2016 ; Yu & Buckler, 2006 ). GWAS has been a powerful tool for dissecting complex traits by identifying associations between single-nucleotide polymorphisms (SNPs) and phenotypic variation. In spinach, GWAS has successfully identified genetic loci controlling plant height, downy mildew resistance, and leaf morphology (Cai et al., 2018 ). By leveraging high-density SNP markers, GWAS facilitates the discovery of key genetic regions associated with important traits, supporting marker-assisted selection (MAS) in breeding programs. For instance, previous studies have identified height-related SNPs on chromosomes 2 and 6, linked to increased plant tallness (Shi et al., 2016 ). The high resolution of GWAS enables the detection of both major and minor loci, enhancing genetic improvement strategies without compromising other critical traits such as leaf texture, flavor, or pest resistance (Jones et al., 2019 ). The integration of GWAS with traditional breeding methods can significantly improve selection efficiency, supporting the development of spinach varieties optimized for both agricultural productivity and consumer preferences. Genomic selection (GS) is an advanced breeding approach that utilizes genome-wide markers to predict genetic potential before trait expression (Goddard & Hayes, 2007 ). While GS has been successfully implemented in maize, wheat, and rice to enhance yield, disease resistance, and stress tolerance (Crossa et al., 2017 ), its application in spinach remains limited (Bhattarai & Shi, 2021 ). Nevertheless, studies in other crops highlight the potential of GS to accelerate breeding cycles and improve cultivar development (Gaynor et al., 2017 ). In spinach, GS could be particularly valuable for optimizing plant height, biomass, and leaf morphology by enabling early selection of superior genotypes, reducing reliance on time-intensive field evaluations (Heffner et al., 2009 ). Expanding the application of GS in spinach breeding holds promise for improving agricultural efficiency and developing high-performing cultivars suited to market demands. This study had two primary objectives: (1) to perform a GWAS to identify SNP markers associated with plant height in spinach, and (2) to implement genomic prediction (GP) models to assess the accuracy of these markers in predicting plant height. We utilized a dataset of 15,058 high-quality SNPs obtained from whole-genome resequencing of 307 USDA-GRIN spinach accessions, forming the basis for GWAS and GP analyses. Our findings contribute to a deeper understanding of the genetic architecture of plant height in spinach and provide valuable resources for breeding programs aimed at improving mechanical harvesting efficiency and overall crop performance. Materials and methods Plant material A total of 307 spinach accessions were obtained from the USDA-GRIN spinach germplasm repository. These accessions represented 30 countries, with the majority originating from Turkey (n = 96), the United States (n = 52), Afghanistan (n = 21), North Macedonia (n = 18), China (n = 16), Iran (n = 13), and Belgium (n = 11), collectively accounting for 74.9% of the total collection. Phenotypic assessments for plant height were conducted, and whole-genome resequencing was performed to generate genotypic data. Detailed information on these accessions is provided in Supplementary Table S1. Experimental Design for Plant Height Measurement Phenotypic data for plant height were collected from the 307 accessions at the USDA-ARS research station in Salinas, CA (Chitwood et al., 2016). The experiment utilized pasteurized sandy loam soil in a greenhouse setting. Each accession was grown in plastic pots (10 × 10 × 10 cm) filled with a 2:1 mixture of sand and soil (by volume). A randomized complete block design (RCBD) with three replications was implemented, with 10 plants per accession. Plant height was measured 55 days after planting as the distance from the soil surface to the highest leaf tip. Descriptive statistics, including mean, range, standard deviation (SD), and standard error (SE), were calculated using JMP Genomics 17 (SAS Institute, Cary, NC). The trait distribution was visualized using GAPIT 3, and the mean plant height per accession was used for genome-wide association study (GWAS) analysis. DNA extraction and whole-genome sequencing DNA was extracted from freshly harvested leaves pooled from 5 to 10 plants per genotype. High-quality DNA was fragmented into 350-bp segments using a Covaris Ultrasonic Processor, and sequencing libraries were prepared following a standardized protocol (Van Dijk et al., 2014). Whole-genome resequencing (WGR) was conducted using paired-end sequencing on the Illumina NovaSeq platform at ~10× genome coverage per sample, generating approximately 10 gigabases of sequence data per genotype. Sequencing was performed by BGI (https://www.bgi.com/). For read alignment, the Monoe-Viroflay spinach genome (Collins et al., 2019) was used as the reference genome, obtained from SpinachBase. Alignment was performed using the Burrows-Wheeler Aligner (BWA v0.7.8-r455) (Li & Durbin, 2009). BAM files were sorted, and duplicate reads were removed using SAMtools (v0.1.19-44428cd). BAM files from the same sample were merged using the Picard toolkit (v1.111). SNP and InDel calling was performed using GATK (v3.5) (McKenna et al., 2010), initially identifying ~6 million raw SNPs. Filtering criteria included a minor allele frequency (MAF) >5%, a missing data rate <7%, and a heterozygosity rate <15%. After filtering, 15,058 high-quality SNPs were retained across the six spinach chromosomes (Fig. 1) and the SNP dataset was published at FigShare database with the link: https://figshare.com/account/articles/28603517. Principal component analysis and genetic diversity Genetic diversity and principal component analysis (PCA) were conducted using the Genomic Association and Prediction Integrated Tool (GAPIT) version 3 (Wang & Zhang, 2021; https://zzlab.net/GAPIT/index.html). PCA was performed using eigenvalue decomposition with component numbers ranging from 2 to 10. A neighbor-joining (NJ) phylogenetic tree was constructed to assess genetic relationships among the accessions. Genome-wide association study GWAS was conducted using five statistical models implemented in GAPIT 3: the generalized linear model (GLM), mixed linear model (MLM), multiple loci mixed model (MLMM), Fixed and Random Model Circulating Probability Unification (FarmCPU), and the Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK) model (Wang & Zhang, 2021; https://zzlab.net/GAPIT/index.html). Association significance was determined using a Bonferroni-corrected threshold (0.05/total SNPs), corresponding to a logarithm of odds (LOD) score of 5.48. Candidate Gene Identification Candidate genes near significant SNPs were identified within a 50-kb flanking region using Monoe-Viroflay genome annotations. Genome annotation data were accessed through SpinachBase (http://www.spinachbase.org/) or via FTP (http://spinachbase.org/ftp/genome/Monoe-Viroflay/). Genomic prediction for plant height Genomic prediction model Genomic prediction (GP) was performed using several models implemented in R packages. Ridge regression best linear unbiased prediction (RR-BLUP) was conducted using the 'rrBLUP' package. Four Bayesian models—Bayes A (BA), Bayes B (BB), Bayesian LASSO (BL), and Bayesian Ridge Regression (BRR)—were implemented using the 'BGLR' package . Additionally, Random Forest (RF) analysis was performed using the 'randomForest' package, and Support Vector Machines (SVM) were applied using the 'kernlab' package. These approaches have been previously utilized in genomic prediction studies (Bao et al., 2014; Ravelombola et al., 2019, 2020, 2021; Shi et al. 2012, 2022). Genomic prediction using different SNP sets We examined ten randomly selected subsets of SNPs, ranging from 6 to 15,058 markers, designated as r6, r50, r100, r200, r500, r1000, r2000, r5000, r10000, and all.15,058SNPs. Additionally, a marker set (m6) was derived from a GWAS conducted on a panel of 307 accessions using four models—GLM, MLM, FarmCPU, and BLINK—implemented in GAPIT3. Genomic estimated breeding values (GEBVs) were calculated for each of the 11 SNP sets across all seven models. Each combination underwent 100 iterations, and the mean correlation coefficients (r-values) along with standard errors (SE) were computed to assess model performance. Boxplots illustrating the performance of GP models across different SNP sets were generated using the 'ggplot2' package in R. GP by GWAS-derived SNP markers GWAS-derived SNP markers from the whole panel and self-prediction First, GWAS was conducted using four models (GLM, MLM, FarmCPU, and BLINK), and the associated SNP markers were identified from these models in the entire GWAS panel (307 spinach accessions). Secondly, GP was performed using the GWAS-derived SNP markers, with the whole panel serving as both the training population (TP) and validation population (VP). GP was performed as described in the previous section on GP using different SNP sets. GWAS-derived SNP markers from 80% of the whole panel Both cross- and across-population predictions were performed for tallness using GWAS-derived associated SNP markers. The entire panel (307 accessions) was divided into two subsets: 80% as the training population (TP) (246 accessions) and 20% as the validation population (VP) (61 accessions). GWAS was performed on the 246 accessions using the GLM, MLM, FarmCPU, and BLINK models in GAPIT3. Associated SNPs with a LOD score (-log(P)) > 4.0 were selected from the four models and used to run the GP model 100 times, calculating GEBVs and estimating the average r-value each time. This process was repeated five times, and the mean r-value across the five replications was obtained as the prediction accuracy (average r-value). Three GP types were tested: ‘Across-prediction’, ‘Cross-prediction’, and ‘Cross_self.prediction’. Across_prediction uses GWAS-derived SNP markers from the training set (80% of the population, 246 accessions) to predict the validation set (20% – 61 accessions). Across_self.prediction uses GWAS-derived SNP markers from the training set (80% of the population) to predict itself. Cross_prediction uses all associated SNP markers from the five repeats to predict the entire population (307 accessions). Additionally, GP was performed with five GP models (RF, BA, BB, BL, and BRR), and GEBVs were calculated for all models. Each replication in each model was run 100 times, and mean r-values along with standard errors (SE) were computed. Boxplots illustrating GP model performance across SNP sets were generated using ggplot2 in R. GWAS-derived SNP markers using GAGBLUP in GAPIT3 GP was conducted using the GAGBLUP (BLINK) model in GAPIT3 on the entire population of 307 accessions, referred to as the reference prediction (cross_self.prediction), where the 307 accessions were used as both the training population (TP) and validation population (VP). Additionally, following the same approach as described above, the entire panel (307 accessions) was divided into two subsets: 80% as the TP (246 accessions) and 20% as the VP (61 accessions). GWAS was performed using the BLINK model only in GAPIT3, and the associated SNPs with a LOD score (-log(P)) > 5.48 were selected to run the GAGBLUP model in GAPIT3. Both across- and cross-population predictions were performed. The across-population prediction (Across-prediction) was performed using the associated SNP markers from the TP (246 accessions) to predict the GEBVs in the VP (61 accessions). Cross-population prediction was performed using the associated SNP markers from the TP (246 accessions) to predict the GEBVs in the TP itself (246 accessions). Results Phenotyping of tallness Phenotypic data for plant height (tallness) across the 307 spinach accessions (Supplementary Table S1) exhibited a near-normal distribution (Fig. 2), with heights ranging from 4.5 to 16.2 cm. The shortest accession, PI 303138, measured 4.5 cm, while the tallest, PI 177557, reached 16.2 cm, approximately 11.7 times taller. The mean plant height was 8.8 cm (SD = 1.9), with a coefficient of variation of 21.3%. The observed variation in plant height demonstrates the suitability of this panel for GWAS. Seven accessions—PI 445784, PI 192945, PI 664497, PI 478393, PI 177558, and PI 433209—were identified as exceptionally tall, each exceeding 13 cm in height. These accessions represent valuable genetic resources for breeding programs aimed at enhancing plant height in spinach. PCA and phylogenetic analysis PCA performed using GAPIT 3 on 307 spinach accessions, each with 15,058 SNPs across six chromosomes, identified three distinct sub-populations when PCA = 3. These sub-populations were further explored through phylogenetic analysis, with trees constructed using the neighbor-joining (NJ) method. The phylogenetic trees distinctly positioned each accession within one of the three sub-populations. The results are presented in a 3D PCA plot (Fig. 3A), a PCA eigenvalue plot (Fig. 3B), and fan-shaped and unrooted phylogenetic trees (Fig. 3C and 3D). These analyses highlight the three sub-populations (Q1, Q2, and Q3) for each accession, as detailed in Supplementary Table S1. Association study In this study, association analyses for plant height (tallness) were conducted using four models—GLM, MLM, FarmCPU, and BLINK—within GAPIT3, with PCA set to 3. Significant deviations from the expected LOD (-log 10 (P-value)) distributions were observed in the QQ plots, comparing the observed versus expected values. These deviations were consistent across multiple QQ plots generated from the four models in the 307 spinach accessions (Fig. 4 right, Supplementary Fig. S1 right), indicating the presence of SNP associations linked to plant height within the analyzed population. The association analysis results for the tallness trait were visualized in Manhattan plots (Fig. 4 left, Supplement Fig. S1 left) using the FarmCPU, GLM, and BLINK models in GAPIT3. Each SNP is represented as a point on the plot, with chromosomal positions indicated on the x-axis and the –log 10 ( P -value) on the y-axis. SNPs with LOD values exceeding the significance threshold of 5.48 were considered significantly associated with the tallness trait. According to the results from the four models, six SNPs were found to be significantly associated, each displaying a LOD value greater than 5.48 in at least one model (Table 1). Table 1 Six SNP markers associated with tallness trait in spinach based on four models in GAPIT3 (GLM, MLM, FarmCPU, and BLINK) and t -test SNP Chr Pos LOD = [-log10(P-value)] Allele (Short) Allele (Tall) Model GAPIT3 t- test BLINK FarmCPU GLM MLM SOVchr2_68062488 2 68062488 3.35 5.67 3.51 2.97 1.09 A G FarmCPU SOVchr4_38323167 4 38323167 5.57 6.47 4.63 4.3 2.09 A G FarmCPU,BLINK SOVchr4_188084338 4 188084338 5.74 5.28 4.94 4.85 2.55 T C BLINK SOVchr5_70192260 5 70192260 2.04 5.53 1.08 1.5 1.31 C A FarmCPU SOVchr6_8139833 6 8139833 7.64 3.43 6.1 5.47 5.19 G T BLINK,GLM SOVchr6_91175684 6 91175684 5.6 6.04 3.9 3.76 2.12 G C FarmCPU,BLINK Notably, SOVchr6_8139833 consistently exhibited a LOD value greater than 5.48 across three models (BLINK, GLM, and MLM), indicating a strong association, although a lower LOD value of 3.43 was recorded in the FarmCPU model. Similarly, SOVchr4_188084338 was exclusively associated with the BLINK model, showing a LOD of 5.74, while the other three models reported lower LOD values (>1.08). Additionally, significant associations were observed for SOVchr6_91175684 and SOVchr4_38323167 on chromosomes 4 and 6, highlighting their potential role in the genetic regulation of plant height in spinach. Furthermore, SOVchr2_68062488 and SOVchr5_70192260 were associated with the BLINK model, each displaying a LOD value of 5.53, while other models reported LOD values greater than 1.08. The identification of these SNPs, particularly those with LOD values exceeding 5.48 on chromosomes 2, 4, 5, and 6, underscores their significance as genetic markers linked to plant height. These findings provide valuable insights into the genetic architecture of tallness in spinach and offer promising targets for future breeding programs. The distribution of these SNPs among the 307 spinach accessions revealed distinct differences in plant height among six genotypes with different allele combinations for each SNP, as illustrated in Supplementary Fig. S2, further reinforcing their association with this trait. Candidate gene identification/detection A total of four SNP markers associated with the tallness trait in spinach were found to be in close proximity to five genes within a 50 kb range. One of these genes, SOV4g016060 (U6 snRNA-associated Sm-like protein LSm5), located between 38,326,318 bp and 38,334,619 bp on chromosome 4, was identified as a potential candidate gene related to the SNP marker SOVchr4_38323167. Additionally, two genes on chromosome 4, SOV4g059190 (outer envelope membrane protein 7-like) positioned between 188,080,640 bp and 188,082,495 bp, and SOV4g059200 (epimerase domain-containing protein) located between 188,085,284 bp and 188,087,037 bp, were identified as potential candidates near the SNP marker SOVchr4_188084338. Moreover, the genes SOV5g028680 (cleavage and polyadenylation specificity factor subunit 2) on chromosome 5, spanning from 70,188,248 bp to 70,192,130 bp, and SOV6g020520 (LETM1 and EF-hand domain-containing protein 1 mitochondrial) on chromosome 6, positioned from 91,176,079 bp to 91,179,277 bp, were identified as potential candidates for tallness near the SNP markers SOVchr5_70192260 and SOVchr6_91175684, respectively (Table 2). Table 2 Five candidate genes within50 kb distance from one associated SNP marker for tallness trait in spinach Gene Start_pos (bp) End_pos (bp) Gene size (bp) Annotation_gene_name SNP Chr Comment SOV4g016060 38326318 38334619 8301 U6 snRNA-associated Sm-like protein LSm5 SOVchr4_38323167 4 <4kb SOV4g059190 188080640 188082495 1855 outer envelope membrane protein 7-like SOVchr4_188084338 <2kb SOV4g059200 188085284 188087037 1753 Epimerase domain-containing protein SOVchr4_188084338 <1kb SOV5g028680 70188248 70192130 3882 Cleavage and polyadenylation specificity factor subunit 2 SOVchr5_70192260 5 130bp SOV6g020520 91176079 91179277 3198 LETM1 and EF-hand domain-containing protein 1 mitochondrial SOVchr6_91175684 6 395bb Genomic prediction for genomic selection of tallness trait Genomic prediction using different SNP sets All seven GP models—BA, BB, BL, BRR, rrBLUP, RF, and SVM—showed similar r-values across SNP sets, ranging from r6 to all.15058SNPs, with r-values averaging from 0.08 (r6) to 0.15 (all.15058SNPs). These results demonstrated that r-values increased as more SNPs were used (Table S2, Fig. 5 and S3). However, the overall prediction accuracy remained low, as indicated by these r-values. GP by GWAS-derived SNP markers GWAS-derived SNP markers from whole panel and self-prediction GWAS-derived SNP markers were identified from the entire GWAS panel (307 spinach accessions), with the whole panel used as both the training population (TP) and validation population (VP). The m6 set (6 GWAS-derived SNP markers) showed higher r-values (Fig. 6, Supplementary Table S3), validating their association with the tallness trait within the panel. However, the r-values are likely to decrease when these markers are applied in across-population predictions. GWAS-derived SNP markers from 80% of the whole panel Across all scenarios, GWAS-derived SNP markers generally produced r-values of 0.51m ranged from 0.47 in RF model to 0.54 in BRR in cross-population predictions, and 0.55, ranged from 0.46 in RF to 0.58 in BA, BL, and BRR in across-self.population prediction; but dropped significantly to 0.12, ranged from 0.10 in RF to 0.12 in other four Bayesian models in across-population predictions. (Table S3, Fig. 7). These results confirm the association of GWAS-derived SNP markers with tallness trait but it doesn’t support the GS implementation in selecting tallness trait in spinach breeding programs (Table S3, Fig. 7). GWAS-derived SNP markers using GAGBLUP in GAPIT3 GP was conducted using the GAGBLUP (BLINK) model in GAPIT3 (Fig. 8). The reference prediction (self-prediction = All.population.set) and cross-population prediction yielded r-values of 0.41 and 0.39, respectively (Fig. 8). However, the r-value dropped significantly to 0.13 in across-population predictions. These findings suggest that GP using only the significant SNP markers identified by GAGBLUP may not be highly effective for selecting the tallness trait in spinach through GS across populations. Genetic Prediction Using Difference Genomic Models Seven GP models (BA, BB, BRR, BL, rrBLUP, cBLUP, and gBLUP) were employed to estimate GP (r-values) for both cross- and across-population prediction and all seven models exhibited similar r-values (Supplementary Tables S2 and S3, Fig. 5, 6, 7, and 8; Supplementary Fig. S3) with several seniors. Among the ten randomly selected SNP sets, all seven models had the r-value averaged 0.11 or 0.12 (Table S2, Fig. 5 & S3). When using the 6 associated SNP markers (m6) as the SNP set, the average r-value was 0.51, 0.51, 0.51, 0.53, 0.50, 0.44, and 0.41 from BA, BB, BL, BRR, rrBLUP, RF, and SVM, respectively (Table S2, Fig. 6), showing the BRR had the highest r-value of 0.53. The BRR also showed the highest PA (r-value) when using GWAS-derived SNP markers (Table S3; Fig. 7). These results suggest that the BRR model is particularly well-suited for predicting tallness in spinach and it is recommended for GS of tallness trait in spinach molecular breeding programs. Discussion Phenotyping of tallness The 307 spinach accessions exhibited significant phenotypic variation in tallness, highlighting the complexity and quantitative nature of this trait. In addition, the observed range, spanning 4.5 cm to 16.2 cm, indicates a broad genetic base, which is essential for successful GWAS and breeding programs aimed at improving plant height. This diversity is consistent with findings in other crops, such as rice and maize, where height is influenced by multiple genes with small effects (Huang & Han, 2016; Yu & Buckler, 2006). Polygenic traits often result in continuous phenotypic variation, which is exactly what we observed in this spinach population. The identification of particularly tall accessions, such as PI445784 and PI192945, suggests the presence of favorable alleles in these accessions that could be valuable in breeding programs. This finding aligns with studies on wheat and barley, where specific alleles have been identified and exploited to successfully enhance plant height (Chitwood et al., 2016). The coefficient of variation of 21.3% further indicates substantial phenotypic variability, which is beneficial for selection and increases the likelihood of detecting significant genetic associations. Similar levels of variability have proven advantageous in other crops, supporting the use of diverse panels in GWAS to identify key loci associated with target traits (Magar et al., 2021). Thus, the observation of substantial variability in this study confirms the suitability of this spinach panel for uncovering the genetic underpinnings of tallness and paves the way for more effective breeding strategies. PCA and phylogenetic analysis The population structure and genetic diversity of spinach have been extensively explored using various methodologies, including SNP markers and phylogenetic analyses (Shi et al., 2016). Spinach exhibits significant variability in key traits essential for breeding and crop improvement, such as plant height and leaf morphology (Rashid et al., 2020b). In this study, we utilized high-density SNP data and principal component analysis (PCA) to assess the genetic diversity of 307 spinach accessions. The results revealed three distinct sub-populations, consistent with earlier studies that highlighted the complex genetic structure of spinach germplasm (Patterson et al., 2006; Saitou & Nei, 1987). Association study In the association study on spinach tallness, which utilized multiple models (MLM, GLM, FarmCPU, and BLINK) within the GAPIT3 framework, we detected consistent deviations in the QQ plots, suggesting that the identified SNPs likely contribute to the observed phenotypic variation in height. This finding echoes previous research on other crops, such as rice, maize, and wheat, where plant height has been demonstrated as a complex polygenic trait influenced by multiple loci (Huang & Han, 2016; Yu & Buckler, 2006). The identification of significant SNPs in these crops has been crucial not only for understanding the genetic basis of height but also for guiding breeding programs aimed at improving this trait. Specifically, certain alleles have been exploited to enhance barley stature, further illustrating the value of SNP identification for crop improvement (Chitwood et al., 2016). The present GWAS identified six SNPs that exceeded the significance threshold, making them suitable candidates for marker-assisted selection. Targeted breeding based on genetic markers has been successfully applied in crops such as maize, rice, and wheat to develop superior cultivars with improved yield and adaptability (Collard & Mackill, 2008; Huang & Han, 2016; Yu & Buckler, 2006). Candidate gene identification/detection In this study, four SNP markers were identified within 50 kb of five candidate genes, suggesting these genes may play important roles in controlling plant height. Both SOV4g016060 (U6 snRNA-associated Sm-like protein LSm5), located near SOVchr4_38323167 on chromosome 4, and SOV5g028680 (cleavage and polyadenylation specificity factor subunit 2), near SOVchr5_70192260 on chromosome 5, are involved in RNA processing—a critical function previously linked to growth regulation in multiple crops. For instance, in maize, genes involved in RNA processing have been shown to influence plant height by regulating the expression of growth-related genes (Huang & Han, 2016). Similarly, in rice, RNA processing genes can affect both height and yield (Gong et al., 2021). In barley, genes associated with RNA processing have been found to regulate flowering time and overall plant stature (Nitcher et al., 2013). Two additional candidate genes on chromosome 4, SOV4g059190 (outer envelope membrane protein 7-like) and SOV4g059200 (epimerase domain-containing protein), both located near SOVchr4_188084338, are associated with metabolic and transport processes. These processes are crucial for cell elongation and biomass accumulation, as previously evidenced in rice and maize. In rice, genes related to metabolic pathways have been linked to the regulation of internode elongation, a key factor in determining plant height (Yu & Buckler, 2006). In maize, the transport of nutrients and growth regulators is critical for the development of tall plants (Hütsch & Schubert, 2018). The last candidate gene identified in this study, SOV6g020520 (LETM1 and EF-hand domain-containing protein 1 mitochondrial), located near SOVchr6_91175684 on chromosome 6, suggests a role for mitochondrial function in regulating spinach height. Mitochondria are essential for energy production, which is necessary to sustain the metabolic demands of growing plants. Studies in wheat and barley have demonstrated that mitochondrial function is closely linked to plant vigor and height, with efficient energy production supporting taller growth (Lozano et al., 2009). Genomic prediction for genomic selection of tallness trait Integration of GP models into breeding programs has become an essential tool for enhancing crop traits, such as plant height, through genomic selection. In this study, we evaluated the performance of seven GP models in predicting tallness in 307 spinach accessions using both randomly selected SNP sets and GWAS-derived SNP marker sets. The seven GP models—BA, BB, BL, BRR, rrBLUP, RF, and SVM—showed similar r-values across SNP sets from r6 to all.15058SNPs, averaging from 0.08 (r6) to 0.16 (r1000) (Table S2, Fig. 5 & S3). The r-value generally increased as the number of SNPs in the set increased, but the improvement plateaued after 1,000 SNPs. The results demonstrated that increasing the number of SNPs from six to 15,058 led to a progressive rise in prediction accuracy (r-value), stabilizing around 1,000 SNPs across all models. These findings underscore the necessity of utilizing a sufficient number of markers to achieve reliable predictions, consistent with previous research emphasizing the importance of genome-wide coverage for accurate GP (Heslot et al., 2012). However, all r-values were low, indicating that GP may not be efficient for predicting the tallness trait. Despite the general trend of higher SNP numbers correlating with improved accuracy, the GWAS-derived SNP set (m6), comprising only six markers, exhibited a relatively high average r-value of 0.49, with the BRR model achieving a peak r-value of 0.53. These outcomes suggest that even a small number of strategically selected SNPs can provide substantial predictive power, particularly when those markers are closely linked to the trait of interest (Zhao et al., 2021). This parallels findings in other crops, where GWAS-derived markers have proven instrumental in improving prediction models for complex traits (Minamikawa et al., 2021). Among the seven GP models assessed, the BRR model consistently outperformed the others, particularly when using the m6 SNP set and other GWAS-derived SNP marker sets. This result suggests that the BRR model is suitable for predicting tallness in spinach. The model’s superior performance may be attributed to its robust ability to capture additive genetic variance, a key factor in polygenic traits like plant height, where the contributing genes have individually small effects (Crossa et al., 2017; Goddard & Hayes, 2007). Conclusion Phenotypic data revealed significant variability in plant height (tallness), with seven accessions exhibiting exceptional tallness identified as promising candidates for breeding programs. Six single nucleotide polymorphisms (SNPs) on chromosomes 2, 4, 5, and 6 were strongly associated with tallness, with notable contributions from markers on chromosome 6. Five candidate genes, involved in RNA processing, metabolic pathways, and mitochondrial function, were located within 50 kb of these SNPs, warranting further investigation as potential regulators of the tallness trait. The genomic prediction (GP) models, particularly the BRR model, demonstrated high predictive accuracy, even when applied to a small GWAS-derived SNP set. This finding supports the utility of these markers for forecasting genetic potential for plant height. These insights offer breeders valuable tools for targeted selection and genotyping, facilitating the development of spinach varieties optimized for mechanical harvesting and market preferences. By integrating genomic insights with traditional breeding methods, this study establishes a strong foundation for future strategies aimed at improving spinach height, contributing to more sustainable and economically viable agricultural practices. Declarations Author contributions AS and BM designed the research and experimental strategy. BM, IA, HA, QL, KC, YQ, RD, AW, and DJH conducted the experiments. AS, IA, and HX analyzed the data. IA drafted and HX and AS edited and revised the manuscript. All authors contributed to the study and approved the final manuscript for publication. Funding This research was funded by the USDA-NIFA SCRI project # 2023-51181-41321; USDA NIFA Hatch (Project numbers ARK0VG2018, ARK02440, and ARK02609), and a scholarship from the Saudi Arabia government, the Saudi Arabian Cultural Mission (SACM), and the University of Tabuk, Saudi Arabia. Acknowledgments This research was supported by the USDA-NIFA SCRI and USDA NIFA Hatch grants. The authors are grateful to the scientists who contributed to this project, as well as to the reviewers and editors for their constructive feedback. Data availability The data presented in this study are available within the article and its Supplementary Material. Whole genome resequencing (WGR) data aligned with the reference genome are available at NCBI under BioProject ID: PRJNA860974. The SNP datasets can be accessed through Tables, Figures, Supplementary Tables, and Supplementary Figures. The SNP data are also available on FigShare at https://figshare.com/account/articles/28603517. The accession numbers used in this study are provided in the article and Supplementary Materials for reference. Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. References Barili, L D, Vale, N M do, Silva, F F e, Carneiro, J E de S, Oliveira, H R de, Vianello, R P, Valdisser, P A M R, & Nascimento, M (2018) Genome prediction accuracy of common bean via Bayesian models Ciência Rural , 48 (8) https://doiorg/101590/0103-8478cr20170497 Bhattarai, G, & Shi, A (2021) Research advances and prospects of spinach breeding, genetics, and genomics Vegetable Research , 1 (1), 1–18 Cai, X, Xu, C, Wang, X, Wang, S, Zhang, Z, Fei, Z, & Wang, Q (2018) Construction of genetic linkage map using genotyping-by-sequencing and identification of QTLs associated with leaf color in spinach Euphytica , 214 , 1–11 Chitwood, J, Shi, A, Mou, B, Evans, M, Clark, J, Motes, D, Chen, P, & Hensley, D (2016) Population structure and association analysis of bolting, plant height, and leaf erectness in spinach HortScience , 51 (5) https://doiorg/1021273/hortsci515481 Collard, B C Y, & Mackill, D J (2008) Marker-assisted selection: An approach for precision plant breeding in the twenty-first century In Philosophical Transactions of the Royal Society B: Biological Sciences (Vol 363, Issue 1491) https://doiorg/101098/rstb20072170 Collins, K, Zhao, K, Jiao, C, Xu, C, Cai, X, Wang, X, Ge, C, Dai, S, Wang, Q, Wang, Q, Fei, Z, & Zheng, Y (2019) SpinachBase: A central portal for spinach genomics Database , 2019 (1) https://doiorg/101093/database/baz072 Crossa, J, Pérez, P, Hickey, J, Burgueño, J, Ornella, L, Cerón-Rojas, J, & de los Campos, G (2017) Genomic prediction in CIMMYT maize and wheat breeding programs Heredity , 112 (1), 48–60 Frary, A, Gul, D, Keles, D, others, & Doganlar, S (2010) Salt Tolerance in Solanum Pennellii: Antioxidant Response and Related QTL BMC Plant Biology , 10 (1), 1–16 Gaynor, R C, Gorjanc, G, Bentley, A R, Ober, E S, Howell, P, Jackson, R, Mackay, I J, & Hickey, J M (2017) A two-part strategy for using genomic selection to develop inbred lines Crop Science , 57 (5) https://doiorg/102135/cropsci2016090742 Goddard, M E, & Hayes, B J (2007) Genomic selection based on dense genotypes inferred from sparse genotypes Journal of Animal Breeding and Genetics , 124 (6), 423–430 Gong, L sha, Qu, S jie, Huang, G min, Guo, Y ling, Zhang, M cai, Li, Z hu, Zhou, Y yi, & Duan, L sheng (2021) Improving maize grain yield by formulating plant growth regulator strategies in North China Journal of Integrative Agriculture , 20 (2) https://doiorg/101016/S2095-3119(20)63453-0 Heffner, E L, Sorrells, M E, & Jannink, J L (2009) Genomic selection for crop improvement Crop Science , 49 (1), 1–12 Heslot, N, Yang, H P, Sorrells, M E, & Jannink, J L (2012) Genomic selection in plant breeding: A comparison of models Crop Science , 52 (1) https://doiorg/102135/cropsci2011060297 Huang, X, & Han, B (2016) Natural variations and genome-wide association studies in crop plants Annual Review of Plant Biology , 67 , 101–122 https://doiorg/101146/annurev-arplant-043015-111846 Hütsch, B W, & Schubert, S (2018) Maize harvest index and water use efficiency can be improved by inhibition of gibberellin biosynthesis Journal of Agronomy and Crop Science , 204 (2) https://doiorg/101111/jac12250 Jones, R B, Phillips, S, & Patterson, R (2019) Mechanization in leafy vegetable production: Impacts on crop yield and quality Agronomy Journal , 111 (4), 1534–1542 Legarra, A, Robert-Granié, C, Croiseau, P, Guillaume, F, & Fritz, S (2011) Improved Lasso for genomic selection Genetics Research , 93 (1) https://doiorg/101017/S0016672310000534 Li, H, & Durbin, R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform Bioinformatics , 25 (14) https://doiorg/101093/bioinformatics/btp324 Lozano, R, Giménez, E, Cara, B, Capel, J, & Angosto, T (2009) Genetic analysis of reproductive development in tomato In International Journal of Developmental Biology (Vol 53, Issues 8–10) https://doiorg/101387/ijdb072440rl Magar, B T, Acharya, S, Gyawali, B, Timilsena, K, Upadhayaya, J, & Shrestha, J (2021) Genetic variability and trait association in maize (Zea mays L) varieties for growth and yield traits Heliyon , 7 (9) https://doiorg/101016/jheliyon2021e07939 McKenna, A, Hanna, M, Banks, E, Sivachenko, A, Cibulskis, K, Kernytsky, A, Garimella, K, Altshuler, D, Gabriel, S, Daly, M, & DePristo, M A (2010) The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data Genome Research , 20 (9) https://doiorg/101101/gr107524110 Minamikawa, M F, Kunihisa, M, Noshita, K, Moriya, S, Abe, K, Hayashi, T, Katayose, Y, Matsumoto, T, Nishitani, C, Terakami, S, Yamamoto, T, & Iwata, H (2021) Tracing founder haplotypes of Japanese apple varieties: application in genomic prediction and genome-wide association study Horticulture Research , 8 (1) https://doiorg/101038/s41438-021-00485-3 Nitcher, R, Distelfeld, A, Tan, C, Yan, L, & Dubcovsky, J (2013) Increased copy number at the HvFT1 locus is associated with accelerated flowering time in barley Molecular Genetics and Genomics , 288 (5–6) https://doiorg/101007/s00438-013-0746-8 Patterson, N, Price, A L, & Reich, D (2006) Population structure and eigenanalysis PLoS Genetics , 2 (12) https://doiorg/101371/journalpgen0020190 Rashid, M, Yousaf, Z, Ullah, M N, Munawar, M, Riaz, N, Younas, A, Aftab, A, & Shamsheer, B (2020a) Genetic variability assessment of worldwide spinach accessions by agro-morphological traits Journal of Taibah University for Science , 14 (1), 1637–1650 Rashid, M, Yousaf, Z, Ullah, M N, Munawar, M, Riaz, N, Younas, A, Aftab, A, & Shamsheer, B (2020b) Genetic variability assessment of worldwide spinach accessions by agro-morphological traits Journal of Taibah University for Science , 14 (1), 1637–1650 Ravelombola, W, Shi, A, and Huynh, B (2021) Loci discovery, network-guided approach, and genomic prediction for drought tolerance index in a multi-parent advanced generation intercross (MAGIC) cowpea population Hortic Res 8:24 https://doiorg/101038/s41438-021-00462-w Ravelombola, W, Qin, J, Shi, A, Nice, L, Bao, Y, Lorenz, A, Orf, JH, Young, ND, and Chen, S (2019) Genome-wide association study and genomic selection for soybean chlorophyll content associated with soybean cyst nematode BMC Genomics, 20, 904 https://doiorg/101186/s12864-019-6275-z Ravelombola, W, Qin, J, Shi, A, Nice, L, Bao, Y, Lorenz, A, Orf, JH, Young, ND, and Chen, S (2020) Genome-wide association study and genomic selection for tolerance of soybean biomass reduction under to soybean cyst nematode infestation Plos One doiorg/101371/journalpone0235089 Saitou, N, & Nei, M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees Molecular Biology and Evolution , 4 (4) https://doiorg/101093/oxfordjournalsmolbeva040454 Shi, A, Mou, B, & Cheng, Z M (2016) Genetic diversity and association analysis of leafminer (Liriomyza langei) resistance in spinach (Spinacia oleracea) Genome , 59 (8) https://doiorg/101139/gen-2016-0075 Shi, A, P Gepts, Q Song, H Xiong, TE Michaels, and S Chen 2021 Genome-wide association study and genomic selection for soybean cyst nematode resistance in USDA common bean (Phaseolus vulgaris) core collection Frontiers in Plant Science, 07, https://doiorg/103389/fpls2021624156 Shi, A, G Bhattarai, H Xiong, CA Avila, C Feng, B Liu, V Joshi, L Stein, B Mou, LJ du Toit, and JC Correll 2022 Genome-wide association study and genomic prediction of white rust resistance in USDA GRIN Spinach Germplasm Horticulture Research, Volume 9, 2022, uhac069, https://doiorg/101093/hr/uhac069 Tang, X, Mu, X, Shao, H, Wang, H, & Brestic, M (2015) Global plant-responding mechanisms to salt stress: Physiological and molecular levels and implications in biotechnology In Critical Reviews in Biotechnology (Vol 35, Issue 4) https://doiorg/103109/073885512014889080 Van Dijk, E L, Jaszczyszyn, Y, & Thermes, C (2014) Library preparation methods for next-generation sequencing: Tone down the bias In Experimental Cell Research (Vol 322, Issue 1) https://doiorg/101016/jyexcr201401008 Wang, J, & Zhang, Z (2021) GAPIT Version 3: Boosting Power and Accuracy for Genomic Association and Prediction Genomics, Proteomics and Bioinformatics , 19 (4) https://doiorg/101016/jgpb202108005 Yu, J, & Buckler, E S (2006) Genetic association mapping and genome organization of maize In Current Opinion in Biotechnology (Vol 17, Issue 2) https://doiorg/101016/jcopbio200602003 Zhao, S, Zhang, Q, Liu, M, Zhou, H, Ma, C, & Wang, P (2021) Regulation of plant responses to salt stress In International Journal of Molecular Sciences (Vol 22, Issue 9) MDPI AG https://doiorg/103390/ijms22094609 Additional Declarations No competing interests reported. Supplementary Files sFigureSPTallnes.xlsx sTableSPTallness.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6305818","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":477055100,"identity":"8c6a021b-a9bc-4290-a8e0-8ea763b35c55","order_by":0,"name":"Ibtisam Alatawi","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Ibtisam","middleName":"","lastName":"Alatawi","suffix":""},{"id":477055101,"identity":"0bf5f7e3-b53f-4073-bd11-84f3571aca20","order_by":1,"name":"Haizheng Xiong","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Haizheng","middleName":"","lastName":"Xiong","suffix":""},{"id":477055102,"identity":"bcc9966a-b4bb-4d4b-95b7-dbe9d928fb44","order_by":2,"name":"Hanan Alkabkabi","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Hanan","middleName":"","lastName":"Alkabkabi","suffix":""},{"id":477055103,"identity":"824592ab-0481-4757-a63d-96183c7739ca","order_by":3,"name":"Kenani Chiwina","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Kenani","middleName":"","lastName":"Chiwina","suffix":""},{"id":477055104,"identity":"e86a2bfa-5c08-439a-ab3d-7dad26bf406b","order_by":4,"name":"Beiquan Mou","email":"","orcid":"","institution":"Agricultural Research Service (USDA-ARS)","correspondingAuthor":false,"prefix":"","firstName":"Beiquan","middleName":"","lastName":"Mou","suffix":""},{"id":477055105,"identity":"a66640e1-5675-4b7c-80af-62e4f0411b2c","order_by":5,"name":"Qun Luo","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Qun","middleName":"","lastName":"Luo","suffix":""},{"id":477055106,"identity":"e38ecee2-6f33-480b-8f18-930723bd624a","order_by":6,"name":"Yuejun Qu","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Yuejun","middleName":"","lastName":"Qu","suffix":""},{"id":477055107,"identity":"b7db531a-2ac4-40cd-b6d1-cf98c4b3d956","order_by":7,"name":"Renjie Du","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Renjie","middleName":"","lastName":"Du","suffix":""},{"id":477055108,"identity":"9311fa1d-e294-46b6-af40-b2627ea51da5","order_by":8,"name":"Awais Riaz","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Awais","middleName":"","lastName":"Riaz","suffix":""},{"id":477055109,"identity":"9167d228-a98e-4311-918a-51efa4d0bc12","order_by":9,"name":"Derrick J. Harrison","email":"","orcid":"","institution":"University of Arkansas","correspondingAuthor":false,"prefix":"","firstName":"Derrick","middleName":"J.","lastName":"Harrison","suffix":""},{"id":477055110,"identity":"ca5b6681-dd93-47c5-ae42-35271333b229","order_by":10,"name":"Ainong Shi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAp0lEQVRIiWNgGAWjYFACHoYDH4CUBBQTp+XgDJK1MPOQpIW//+zBw7ZtdvmSDcwHb/MQo0XiRl7C4dy2ZMvZDGzJ1kRpYbjBYwDUwmwgx8BjJk2UFvnzZwwOW7bVA7XwfyNOi8GBHIPDjG2HDaQZeNiI02J4I8fgYM+54waSzWzGlnOI0SJ3/ozxhx9l1QYSx5sf3nhDjBYEYCZN+SgYBaNgFIwCfAAAxCcsMVx1qBMAAAAASUVORK5CYII=","orcid":"","institution":"University of Arkansas","correspondingAuthor":true,"prefix":"","firstName":"Ainong","middleName":"","lastName":"Shi","suffix":""}],"badges":[],"createdAt":"2025-03-25 16:53:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6305818/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6305818/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85634854,"identity":"539c3905-a45f-481b-b0de-5f21d40fb73d","added_by":"auto","created_at":"2025-06-30 05:34:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":126889,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of the 15,058 high-quality SNPs within 1-Mb window sizes across six spinach chromosomes\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/105b3a06f4d189987bc78e7f.png"},{"id":85634861,"identity":"358be3bb-0b19-4468-a479-188cfbaf5269","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":36524,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of plant tallness in the 307 spinach accessions\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/c82d58472f1625f97404e3f8.png"},{"id":85634863,"identity":"176e68b0-e4a3-45ff-aeac-0b81ad68f723","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":276531,"visible":true,"origin":"","legend":"\u003cp\u003ePopulation genetic diversity analysis in the association panel consisted of 307 USDA spinach germplasm accessions. (A) 3D graphical plot of the principal component analysis (PCA), (B) PCA.eigenValue plot drawn by GAPIT 3, and Phylogenetic trees (C- fan and D - unrooted) drawn by neighbor-joining (NJ) method in three sub-populations\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/bec188e02b63f0f6eb97bff4.png"},{"id":85635694,"identity":"f3904bdd-c070-4be3-ab49-d6b138698c90","added_by":"auto","created_at":"2025-06-30 05:50:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":432257,"visible":true,"origin":"","legend":"\u003cp\u003eThe multiple Manhattan plot (left) and QQ plot (right) symphonically GLM, MLM, FarmCPU, and BLINK models in GAPIT3 for tallness trait in 307 spinach accessions\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/3fcf8d5fbdfa769f9afbd557.png"},{"id":85635524,"identity":"ac7f0078-0e28-4887-b483-8b93fd6c1864","added_by":"auto","created_at":"2025-06-30 05:42:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":78071,"visible":true,"origin":"","legend":"\u003cp\u003eGenomic prediction (r-value) for the tallness trait in 307 spinach accessions using ten different SNP sets, ranging from 6 to 15,058 randomly selected SNPs, in cross-prediction\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/d0187971e1921d2eab73ae55.png"},{"id":85634872,"identity":"6bf67567-f7ba-4552-ae23-fdf637090d06","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":35739,"visible":true,"origin":"","legend":"\u003cp\u003eGenomic prediction (r-value) of the six GWAS-derived SNP markers (listed in Table 1) in cross-prediction for tallness trait in 307 spinach accessions estimated by seven models: BA, BB, BL, BRR, RF, rrBLUP, and SVM\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/a264cf32e6d90ce4b4869409.png"},{"id":85635692,"identity":"31c145d1-7329-4079-a7b1-dde753dead5f","added_by":"auto","created_at":"2025-06-30 05:50:05","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":43979,"visible":true,"origin":"","legend":"\u003cp\u003eGenomic Prediction (GP) (r-value) for Tallness Using GWAS-Derived SNP Markers: First, perform GWAS using the GLM, MLM, FarmCPU, and BLINK models in GAPIT3 with five-fold cross-validation. Then, select associated SNPs with a LOD score (-log(P)) \u0026gt; 4.0 from the four models and use them to run the GP model 100 times, calculating GEBVs and estimating the average r-value each time. Repeat this process five times, and the figure presents the mean r-value across the five replications. Across_prediction uses GWAS-derived SNP markers from the training set (80% of the population, 246 accessions) to predict the validation set (20% - 61 accessions). Across_self_prediction uses GWAS-derived SNP markers from the training set (80% of the population) to predict itself. Cross_prediction uses all associated SNP markers to predict the entire population (307 accessions)\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/da2cd60fec618f28db2975a8.png"},{"id":85634877,"identity":"0ee26523-08d8-479a-bf93-b59c38618696","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":24394,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eGenomic prediction (GP) (r-value) for tallness using the GAGBLUP (BLINK) model in GAPIT3\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/bf11124625901002b9d6afae.png"},{"id":85635695,"identity":"36754400-7368-46bb-964e-d77af8eb7232","added_by":"auto","created_at":"2025-06-30 05:50:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2102127,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/14739749-cdc8-4398-95b6-a1765582dd71.pdf"},{"id":85634882,"identity":"ded6fff5-9d42-43d9-8b46-cbedbb11b054","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":457897,"visible":true,"origin":"","legend":"","description":"","filename":"sFigureSPTallnes.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/da2b7719f9c8fd1f084edfb7.xlsx"},{"id":85634870,"identity":"ae922ada-95e9-4f71-bda0-2e85c4b3de9a","added_by":"auto","created_at":"2025-06-30 05:34:05","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":36495,"visible":true,"origin":"","legend":"","description":"","filename":"sTableSPTallness.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6305818/v1/ece420254693f796c23745f0.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Genome-wide association study and genome prediction of tallness trait in spinach tallness phenotyping","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSpinach (\u003cem\u003eSpinacia oleracea\u003c/em\u003e L.) is a highly nutritious leafy vegetable, widely cultivated in the United States and globally (Shi et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Tang et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Its increasing demand is driven by consumer awareness of its rich nutritional profile, including essential vitamins, minerals, antioxidants, and bioactive compounds such as carotenoids and flavonoids (Frary et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Rashid et al., 2020). Among its key agronomic traits, plant height plays a crucial role in spinach production and management (Jones et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Taller spinach varieties are preferred for their compatibility with mechanical harvesting, which enhances operational efficiency, reduces labor costs, and improves economic returns (Li et al., 2020). However, optimizing plant height requires a balance, as taller plants must also resist lodging\u0026mdash;a condition where plants collapse under adverse weather, leading to yield loss (Jones et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePlant height in spinach, like in other major crops such as rice and maize, is a polygenic trait governed by multiple genetic factors (Huang \u0026amp; Han, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Traditional quantitative trait loci (QTL) mapping approaches have been useful in identifying large-effect loci but often fail to detect small-effect loci that collectively influence complex traits (Yu \u0026amp; Buckler, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). This limitation underscores the need for genome-wide approaches such as genome-wide association studies (GWAS) and genomic selection (GS), which enable the identification of multiple loci contributing to plant height and improve breeding efficiency through genome-wide marker predictions.\u003c/p\u003e \u003cp\u003eThe substantial phenotypic variation observed in spinach plant height reflects its rich genetic diversity, making it an excellent candidate for advanced genomic studies and breeding efforts (Huang \u0026amp; Han, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Yu \u0026amp; Buckler, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). GWAS has been a powerful tool for dissecting complex traits by identifying associations between single-nucleotide polymorphisms (SNPs) and phenotypic variation. In spinach, GWAS has successfully identified genetic loci controlling plant height, downy mildew resistance, and leaf morphology (Cai et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). By leveraging high-density SNP markers, GWAS facilitates the discovery of key genetic regions associated with important traits, supporting marker-assisted selection (MAS) in breeding programs. For instance, previous studies have identified height-related SNPs on chromosomes 2 and 6, linked to increased plant tallness (Shi et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). The high resolution of GWAS enables the detection of both major and minor loci, enhancing genetic improvement strategies without compromising other critical traits such as leaf texture, flavor, or pest resistance (Jones et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The integration of GWAS with traditional breeding methods can significantly improve selection efficiency, supporting the development of spinach varieties optimized for both agricultural productivity and consumer preferences.\u003c/p\u003e \u003cp\u003eGenomic selection (GS) is an advanced breeding approach that utilizes genome-wide markers to predict genetic potential before trait expression (Goddard \u0026amp; Hayes, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). While GS has been successfully implemented in maize, wheat, and rice to enhance yield, disease resistance, and stress tolerance (Crossa et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), its application in spinach remains limited (Bhattarai \u0026amp; Shi, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Nevertheless, studies in other crops highlight the potential of GS to accelerate breeding cycles and improve cultivar development (Gaynor et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In spinach, GS could be particularly valuable for optimizing plant height, biomass, and leaf morphology by enabling early selection of superior genotypes, reducing reliance on time-intensive field evaluations (Heffner et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Expanding the application of GS in spinach breeding holds promise for improving agricultural efficiency and developing high-performing cultivars suited to market demands.\u003c/p\u003e \u003cp\u003eThis study had two primary objectives: (1) to perform a GWAS to identify SNP markers associated with plant height in spinach, and (2) to implement genomic prediction (GP) models to assess the accuracy of these markers in predicting plant height. We utilized a dataset of 15,058 high-quality SNPs obtained from whole-genome resequencing of 307 USDA-GRIN spinach accessions, forming the basis for GWAS and GP analyses. Our findings contribute to a deeper understanding of the genetic architecture of plant height in spinach and provide valuable resources for breeding programs aimed at improving mechanical harvesting efficiency and overall crop performance.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003ch2\u003ePlant material\u003c/h2\u003e\n\u003cp\u003eA total of 307 spinach accessions were obtained from the USDA-GRIN spinach germplasm repository. These accessions represented 30 countries, with the majority originating from Turkey (n = 96), the United States (n = 52), Afghanistan (n = 21), North Macedonia (n = 18), China (n = 16), Iran (n = 13), and Belgium (n = 11), collectively accounting for 74.9% of the total collection. Phenotypic assessments for plant height were conducted, and whole-genome resequencing was performed to generate genotypic data. Detailed information on these accessions is provided in Supplementary Table S1.\u003c/p\u003e\n\u003ch2\u003eExperimental Design for Plant Height Measurement\u003c/h2\u003e\n\u003cp\u003ePhenotypic data for plant height were collected from the 307 accessions at the USDA-ARS research station in Salinas, CA (Chitwood et al., 2016). The experiment utilized pasteurized sandy loam soil in a greenhouse setting. Each accession was grown in plastic pots (10 \u0026times; 10 \u0026times; 10 cm) filled with a 2:1 mixture of sand and soil (by volume). A randomized complete block design (RCBD) with three replications was implemented, with 10 plants per accession. Plant height was measured 55 days after planting as the distance from the soil surface to the highest leaf tip. Descriptive statistics, including mean, range, standard deviation (SD), and standard error (SE), were calculated using JMP Genomics 17 (SAS Institute, Cary, NC). The trait distribution was visualized using GAPIT 3, and the mean plant height per accession was used for genome-wide association study (GWAS) analysis.\u003c/p\u003e\n\u003ch2\u003eDNA extraction and whole-genome sequencing\u003c/h2\u003e\n\u003cp\u003eDNA was extracted from freshly harvested leaves pooled from 5 to 10 plants per genotype. High-quality DNA was fragmented into 350-bp segments using a Covaris Ultrasonic Processor, and sequencing libraries were prepared following a standardized protocol (Van Dijk et al., 2014). Whole-genome resequencing (WGR) was conducted using paired-end sequencing on the Illumina NovaSeq platform at ~10\u0026times; genome coverage per sample, generating approximately 10 gigabases of sequence data per genotype. Sequencing was performed by BGI (https://www.bgi.com/).\u003c/p\u003e\n\u003cp\u003eFor read alignment, the Monoe-Viroflay spinach genome (Collins et al., 2019) was used as the reference genome, obtained from SpinachBase. Alignment was performed using the Burrows-Wheeler Aligner (BWA v0.7.8-r455) (Li \u0026amp; Durbin, 2009). BAM files were sorted, and duplicate reads were removed using SAMtools (v0.1.19-44428cd). BAM files from the same sample were merged using the Picard toolkit (v1.111). SNP and InDel calling was performed using GATK (v3.5) (McKenna et al., 2010), initially identifying ~6 million raw SNPs.\u003c/p\u003e\n\u003cp\u003eFiltering criteria included a minor allele frequency (MAF) \u0026gt;5%, a missing data rate \u0026lt;7%, and a heterozygosity rate \u0026lt;15%. After filtering, 15,058 high-quality SNPs were retained across the six spinach chromosomes (Fig. 1) and the SNP dataset was published at FigShare database with the link: https://figshare.com/account/articles/28603517.\u003c/p\u003e\n\u003ch2\u003ePrincipal component analysis and genetic diversity\u003c/h2\u003e\n\u003cp\u003eGenetic diversity and principal component analysis (PCA) were conducted using the Genomic Association and Prediction Integrated Tool (GAPIT) version 3 (Wang \u0026amp; Zhang, 2021; https://zzlab.net/GAPIT/index.html). PCA was performed using eigenvalue decomposition with component numbers ranging from 2 to 10. A neighbor-joining (NJ) phylogenetic tree was constructed to assess genetic relationships among the accessions.\u003c/p\u003e\n\u003ch2\u003eGenome-wide association study\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eGWAS was conducted using five statistical models implemented in GAPIT 3: the generalized linear model (GLM), mixed linear model (MLM), multiple loci mixed model (MLMM), Fixed and Random Model Circulating Probability Unification (FarmCPU), and the Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK) model (Wang \u0026amp; Zhang, 2021; https://zzlab.net/GAPIT/index.html). Association significance was determined using a Bonferroni-corrected threshold (0.05/total SNPs), corresponding to a logarithm of odds (LOD) score of 5.48.\u003c/p\u003e\n\u003ch2\u003eCandidate Gene Identification\u003c/h2\u003e\n\u003cp\u003eCandidate genes near significant SNPs were identified within a 50-kb flanking region using Monoe-Viroflay genome annotations. Genome annotation data were accessed through SpinachBase (http://www.spinachbase.org/) or via FTP (http://spinachbase.org/ftp/genome/Monoe-Viroflay/).\u003c/p\u003e\n\u003ch2\u003eGenomic prediction for plant height\u003c/h2\u003e\n\u003ch2\u003eGenomic prediction model\u003c/h2\u003e\n\u003cp\u003eGenomic prediction (GP) was performed using several models implemented in R packages. Ridge regression best linear unbiased prediction (RR-BLUP) was conducted using the \u0026apos;rrBLUP\u0026apos; package. Four Bayesian models\u0026mdash;Bayes A (BA), Bayes B (BB), Bayesian LASSO (BL), and Bayesian Ridge Regression (BRR)\u0026mdash;were implemented using the \u0026apos;BGLR\u0026apos; package . Additionally, Random Forest (RF) analysis was performed using the \u0026apos;randomForest\u0026apos; package, and Support Vector Machines (SVM) were applied using the \u0026apos;kernlab\u0026apos; package. These approaches have been previously utilized in genomic prediction studies (Bao et al., 2014; Ravelombola et al., 2019, 2020, 2021; Shi et al. 2012, 2022).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eGenomic prediction using different SNP sets\u003c/h2\u003e\n\u003cp\u003eWe examined ten randomly selected subsets of SNPs, ranging from 6 to 15,058 markers, designated as r6, r50, r100, r200, r500, r1000, r2000, r5000, r10000, and all.15,058SNPs. Additionally, a marker set (m6) was derived from a GWAS conducted on a panel of 307 accessions using four models\u0026mdash;GLM, MLM, FarmCPU, and BLINK\u0026mdash;implemented in GAPIT3. Genomic estimated breeding values (GEBVs) were calculated for each of the 11 SNP sets across all seven models. Each combination underwent 100 iterations, and the mean correlation coefficients (r-values) along with standard errors (SE) were computed to assess model performance. Boxplots illustrating the performance of GP models across different SNP sets were generated using the \u0026apos;ggplot2\u0026apos; package in R.\u003c/p\u003e\n\u003ch2\u003eGP by GWAS-derived SNP markers\u003c/h2\u003e\n\u003ch2\u003eGWAS-derived SNP markers from the whole panel and self-prediction\u003c/h2\u003e\n\u003cp\u003eFirst, GWAS was conducted using four models (GLM, MLM, FarmCPU, and BLINK), and the associated SNP markers were identified from these models in the entire GWAS panel (307 spinach accessions). Secondly, GP was performed using the GWAS-derived SNP markers, with the whole panel serving as both the training population (TP) and validation population (VP). GP was performed as described in the previous section on GP using different SNP sets.\u003c/p\u003e\n\u003ch2\u003eGWAS-derived SNP markers from 80% of the whole panel\u003c/h2\u003e\n\u003cp\u003eBoth cross- and across-population predictions were performed for tallness using GWAS-derived associated SNP markers. The entire panel (307 accessions) was divided into two subsets: 80% as the training population (TP) (246 accessions) and 20% as the validation population (VP) (61 accessions). GWAS was performed on the 246 accessions using the GLM, MLM, FarmCPU, and BLINK models in GAPIT3. Associated SNPs with a LOD score (-log(P)) \u0026gt; 4.0 were selected from the four models and used to run the GP model 100 times, calculating GEBVs and estimating the average r-value each time. This process was repeated five times, and the mean r-value across the five replications was obtained as the prediction accuracy (average r-value). Three GP types were tested: \u0026lsquo;Across-prediction\u0026rsquo;, \u0026lsquo;Cross-prediction\u0026rsquo;, and \u0026lsquo;Cross_self.prediction\u0026rsquo;.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eAcross_prediction uses GWAS-derived SNP markers from the training set (80% of the population, 246 accessions) to predict the validation set (20% \u0026ndash; 61 accessions).\u003c/li\u003e\n \u003cli\u003eAcross_self.prediction uses GWAS-derived SNP markers from the training set (80% of the population) to predict itself.\u003c/li\u003e\n \u003cli\u003eCross_prediction uses all associated SNP markers from the five repeats to predict the entire population (307 accessions).\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eAdditionally, GP was performed with five GP models (RF, BA, BB, BL, and BRR), and GEBVs were calculated for all models. Each replication in each model was run 100 times, and mean r-values along with standard errors (SE) were computed. Boxplots illustrating GP model performance across SNP sets were generated using ggplot2 in R.\u003c/p\u003e\n\u003ch2\u003eGWAS-derived SNP markers using GAGBLUP in GAPIT3\u003c/h2\u003e\n\u003cp\u003eGP was conducted using the GAGBLUP (BLINK) model in GAPIT3 on the entire population of 307 accessions, referred to as the reference prediction (cross_self.prediction), where the 307 accessions were used as both the training population (TP) and validation population (VP). Additionally, following the same approach as described above, the entire panel (307 accessions) was divided into two subsets: 80% as the TP (246 accessions) and 20% as the VP (61 accessions). GWAS was performed using the BLINK model only in GAPIT3, and the associated SNPs with a LOD score (-log(P)) \u0026gt; 5.48 were selected to run the GAGBLUP model in GAPIT3. Both across- and cross-population predictions were performed. The across-population prediction (Across-prediction) was performed using the associated SNP markers from the TP (246 accessions) to predict the GEBVs in the VP (61 accessions). Cross-population prediction was performed using the associated SNP markers from the TP (246 accessions) to predict the GEBVs in the TP itself (246 accessions).\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003ePhenotyping of tallness\u003c/h2\u003e\n\u003cp\u003ePhenotypic data for plant height (tallness) across the 307 spinach accessions (Supplementary Table S1) exhibited a near-normal distribution (Fig. 2), with heights ranging from 4.5 to 16.2 cm. The shortest accession, PI 303138, measured 4.5 cm, while the tallest, PI 177557, reached 16.2 cm, approximately 11.7 times taller. The mean plant height was 8.8 cm (SD = 1.9), with a coefficient of variation of 21.3%. The observed variation in plant height demonstrates the suitability of this panel for GWAS.\u003c/p\u003e\n\u003cp\u003eSeven accessions\u0026mdash;PI 445784, PI 192945, PI 664497, PI 478393, PI 177558, and PI 433209\u0026mdash;were identified as exceptionally tall, each exceeding 13 cm in height. These accessions represent valuable genetic resources for breeding programs aimed at enhancing plant height in spinach.\u003c/p\u003e\n\u003ch2\u003ePCA and phylogenetic analysis\u003c/h2\u003e\n\u003cp\u003ePCA performed using GAPIT 3 on 307 spinach accessions, each with 15,058 SNPs across six chromosomes, identified three distinct sub-populations when PCA = 3. These sub-populations were further explored through phylogenetic analysis, with trees constructed using the neighbor-joining (NJ) method. The phylogenetic trees distinctly positioned each accession within one of the three sub-populations. The results are presented in a 3D PCA plot (Fig. 3A), a PCA eigenvalue plot (Fig. 3B), and fan-shaped and unrooted phylogenetic trees (Fig. 3C and 3D). These analyses highlight the three sub-populations (Q1, Q2, and Q3) for each accession, as detailed in Supplementary Table S1.\u003c/p\u003e\n\u003ch2\u003eAssociation study\u003c/h2\u003e\n\u003cp\u003eIn this study, association analyses for plant height (tallness) were conducted using four models\u0026mdash;GLM, MLM, FarmCPU, and BLINK\u0026mdash;within GAPIT3, with PCA set to 3. Significant deviations from the expected LOD (-log\u003csub\u003e10\u003c/sub\u003e(P-value)) distributions were observed in the QQ plots, comparing the observed versus expected values. These deviations were consistent across multiple QQ plots generated from the four models in the 307 spinach accessions (Fig. 4 right, Supplementary Fig. S1 right), indicating the presence of SNP associations linked to plant height\u0026nbsp;\u003cbr\u003e\u0026nbsp;within the analyzed population.\u003c/p\u003e\n\u003cp\u003eThe association analysis results for the tallness trait were visualized in Manhattan plots (Fig. 4 left, Supplement Fig. S1 left) using the FarmCPU, GLM, and BLINK models in GAPIT3. Each SNP is represented as a point on the plot, with chromosomal positions indicated on the x-axis and the \u0026ndash;log\u003csub\u003e10\u003c/sub\u003e(\u003cem\u003eP\u003c/em\u003e-value) on the y-axis. SNPs with LOD values exceeding the significance threshold of 5.48 were considered significantly associated with the tallness trait. According to the results from the four models, six SNPs were found to be significantly associated, each displaying a LOD value greater than 5.48 in at least one model (Table 1).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e Six SNP markers associated with tallness trait in spinach based on four models in GAPIT3 (GLM, MLM, FarmCPU, and BLINK) and \u003cem\u003et\u003c/em\u003e-test\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"750\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSNP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 34px;\"\u003e\n \u003cp\u003eChr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 74px;\"\u003e\n \u003cp\u003ePos\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"5\" style=\"width: 252px;\"\u003e\n \u003cp\u003eLOD = [-log10(P-value)]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 69px;\"\u003e\n \u003cp\u003eAllele\u003cbr\u003e\u0026nbsp;(Short)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 69px;\"\u003e\n \u003cp\u003eAllele\u003cbr\u003e\u0026nbsp;(Tall)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 120px;\"\u003e\n \u003cp\u003eModel\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" style=\"width: 214px;\"\u003e\n \u003cp\u003eGAPIT3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 38px;\"\u003e\n \u003cp\u003e\u003cem\u003et-\u003c/em\u003etest\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003eBLINK\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003eFarmCPU\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 44px;\"\u003e\n \u003cp\u003eGLM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003eMLM\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr2_68062488\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e68062488\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e3.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e5.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e3.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e2.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eFarmCPU\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr4_38323167\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e38323167\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e5.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e6.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e4.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e4.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e2.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 120px;\"\u003e\n \u003cp\u003eFarmCPU,BLINK\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr4_188084338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e188084338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e5.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e5.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e4.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e4.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 38px;\"\u003e\n \u003cp\u003e2.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003eT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eBLINK\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr5_70192260\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e70192260\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e2.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e5.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e1.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e1.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e1.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eFarmCPU\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr6_8139833\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e8139833\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e7.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e3.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e6.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e5.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e5.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003eT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eBLINK,GLM\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr6_91175684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 34px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 74px;\"\u003e\n \u003cp\u003e91175684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 55px;\"\u003e\n \u003cp\u003e5.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003e6.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 44px;\"\u003e\n \u003cp\u003e3.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 46px;\"\u003e\n \u003cp\u003e3.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e2.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 69px;\"\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 120px;\"\u003e\n \u003cp\u003eFarmCPU,BLINK\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eNotably, SOVchr6_8139833 consistently exhibited a LOD value greater than 5.48 across three models (BLINK, GLM, and MLM), indicating a strong association, although a lower LOD value of 3.43 was recorded in the FarmCPU model. Similarly, SOVchr4_188084338 was exclusively associated with the BLINK model, showing a LOD of 5.74, while the other three models reported lower LOD values (\u0026gt;1.08). Additionally, significant associations were observed for SOVchr6_91175684 and SOVchr4_38323167 on chromosomes 4 and 6, highlighting their potential role in the genetic regulation of plant height in spinach. Furthermore, SOVchr2_68062488 and SOVchr5_70192260 were associated with the BLINK model, each displaying a LOD value of 5.53, while other models reported LOD values greater than 1.08.\u003c/p\u003e\n\u003cp\u003eThe identification of these SNPs, particularly those with LOD values exceeding 5.48 on chromosomes 2, 4, 5, and 6, underscores their significance as genetic markers linked to plant height. These findings provide valuable insights into the genetic architecture of tallness in spinach and offer promising targets for future breeding programs. The distribution of these SNPs among the 307 spinach accessions revealed distinct differences in plant height among six genotypes with different allele combinations for each SNP, as illustrated in Supplementary Fig. S2, further reinforcing their association with this trait.\u003c/p\u003e\n\u003ch2\u003eCandidate gene identification/detection\u003c/h2\u003e\n\u003cp\u003eA total of four SNP markers associated with the tallness trait in spinach were found to be in close proximity to five genes within a 50 kb range. One of these genes, \u003cem\u003eSOV4g016060\u003c/em\u003e (U6 snRNA-associated Sm-like protein LSm5), located between 38,326,318 bp and 38,334,619 bp on chromosome 4, was identified as a potential candidate gene related to the SNP marker SOVchr4_38323167.\u003c/p\u003e\n\u003cp\u003eAdditionally, two genes on chromosome 4, \u003cem\u003eSOV4g059190\u003c/em\u003e (outer envelope membrane protein 7-like) positioned between 188,080,640 bp and 188,082,495 bp, and \u003cem\u003eSOV4g059200\u003c/em\u003e (epimerase domain-containing protein) located between 188,085,284 bp and 188,087,037 bp, were identified as potential candidates near the SNP marker SOVchr4_188084338.\u003c/p\u003e\n\u003cp\u003eMoreover, the genes \u003cem\u003eSOV5g028680\u0026nbsp;\u003c/em\u003e(cleavage and polyadenylation specificity factor subunit 2) on chromosome 5, spanning from 70,188,248 bp to 70,192,130 bp, and \u003cem\u003eSOV6g020520\u003c/em\u003e (LETM1 and EF-hand domain-containing protein 1 mitochondrial) on chromosome 6, positioned from 91,176,079 bp to 91,179,277 bp, were identified as potential candidates for tallness near the SNP markers SOVchr5_70192260 and SOVchr6_91175684, respectively (Table 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u0026nbsp;\u003c/strong\u003eFive candidate genes within50 kb distance from one associated SNP marker for tallness trait in spinach\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"888\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGene\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStart_pos\u003cbr\u003e\u0026nbsp;(bp)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEnd_pos\u003cbr\u003e\u0026nbsp;(bp)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGene size\u0026nbsp;\u003cbr\u003e\u0026nbsp;(bp)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnnotation_gene_name\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSNP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 37px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eChr\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eComment\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cem\u003eSOV4g016060\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e38326318\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e38334619\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e8301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003eU6 snRNA-associated Sm-like protein LSm5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr4_38323167\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" style=\"width: 37px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e\u0026lt;4kb\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cem\u003eSOV4g059190\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e188080640\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e188082495\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e1855\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003eouter envelope membrane protein 7-like\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr4_188084338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e\u0026lt;2kb\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cem\u003eSOV4g059200\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e188085284\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e188087037\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e1753\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003eEpimerase domain-containing protein\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr4_188084338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e\u0026lt;1kb\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cem\u003eSOV5g028680\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e70188248\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e70192130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e3882\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003eCleavage and polyadenylation specificity factor subunit 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr5_70192260\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 37px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e130bp\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cem\u003eSOV6g020520\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 106px;\"\u003e\n \u003cp\u003e91176079\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e91179277\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e3198\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 288px;\"\u003e\n \u003cp\u003eLETM1 and EF-hand domain-containing protein 1 mitochondrial\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eSOVchr6_91175684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 37px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e395bb\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eGenomic prediction for genomic selection of tallness trait\u003c/h2\u003e\n\u003cp\u003eGenomic prediction using different SNP sets\u003c/p\u003e\n\u003cp\u003eAll seven GP models\u0026mdash;BA, BB, BL, BRR, rrBLUP, RF, and SVM\u0026mdash;showed similar r-values across SNP sets, ranging from r6 to all.15058SNPs, with r-values averaging from 0.08 (r6) to 0.15 (all.15058SNPs). These results demonstrated that r-values increased as more SNPs were used (Table S2, Fig. 5 and S3). However, the overall prediction accuracy remained low, as indicated by these r-values.\u003c/p\u003e\n\u003ch4\u003eGP by GWAS-derived SNP markers\u003c/h4\u003e\n\u003ch5\u003eGWAS-derived SNP markers from whole panel and self-prediction\u003c/h5\u003e\n\u003cp\u003eGWAS-derived SNP markers were identified from the entire GWAS panel (307 spinach accessions), with the whole panel used as both the training population (TP) and validation population (VP). The m6 set (6 GWAS-derived SNP markers) showed higher r-values (Fig. 6, Supplementary Table S3), validating their association with the tallness trait within the panel. However, the r-values are likely to decrease when these markers are applied in across-population predictions.\u003c/p\u003e\n\u003ch2\u003eGWAS-derived SNP markers from 80% of the whole panel\u003c/h2\u003e\n\u003cp\u003eAcross all scenarios, GWAS-derived SNP markers generally produced r-values of 0.51m ranged from 0.47 in RF model to 0.54 in BRR in cross-population predictions, and 0.55, ranged from 0.46 in RF to 0.58 in BA, BL, and BRR in across-self.population prediction; but dropped significantly to 0.12, ranged from 0.10 in RF to 0.12 in other four Bayesian models in across-population predictions. (Table S3, Fig. 7). These results confirm the association of GWAS-derived SNP markers with tallness trait but it doesn\u0026rsquo;t support the GS implementation in selecting tallness trait in spinach breeding programs (Table S3, Fig. 7).\u003c/p\u003e\n\u003ch2\u003eGWAS-derived SNP markers using GAGBLUP in GAPIT3\u003c/h2\u003e\n\u003cp\u003eGP was conducted using the GAGBLUP (BLINK) model in GAPIT3 (Fig. 8). The reference prediction (self-prediction = All.population.set) and cross-population prediction yielded r-values of 0.41 and 0.39, respectively (Fig. 8). However, the r-value dropped significantly to 0.13 in across-population predictions. These findings suggest that GP using only the significant SNP markers identified by GAGBLUP may not be highly effective for selecting the tallness trait in spinach through GS across populations.\u003c/p\u003e\n\u003ch2\u003eGenetic Prediction Using Difference Genomic Models\u003c/h2\u003e\n\u003cp\u003eSeven GP models (BA, BB, BRR, BL, rrBLUP, cBLUP, and gBLUP) were employed to estimate GP (r-values) for both cross- and across-population prediction and all seven models exhibited similar r-values (Supplementary Tables S2 and S3, Fig. 5, 6, 7, and 8; Supplementary Fig. S3) with several seniors.\u0026nbsp;\u003c/p\u003e\n\u003col class=\"decimal_type\" style=\"list-style-type: lower-roman;\"\u003e\n \u003cli\u003eAmong the ten randomly selected SNP sets, all seven models had the r-value averaged 0.11 or 0.12 (Table S2, Fig. 5 \u0026amp; S3).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWhen using the 6 associated SNP markers (m6) as the SNP set, the average r-value was 0.51, 0.51, 0.51, 0.53, 0.50, 0.44, and 0.41 from BA, BB, BL, BRR, rrBLUP, RF, and SVM, respectively (Table S2, Fig. 6), showing the BRR had the highest r-value of 0.53.\u003c/li\u003e\n \u003cli\u003eThe BRR also showed the highest PA (r-value) when using GWAS-derived SNP markers (Table S3; Fig. 7).\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThese results suggest that the BRR model is particularly well-suited for predicting tallness in spinach and it is recommended for GS of tallness trait in spinach molecular breeding programs.\u003c/p\u003e"},{"header":"Discussion","content":"\u003ch2\u003ePhenotyping of tallness\u003c/h2\u003e\n\u003cp\u003eThe 307 spinach accessions exhibited significant phenotypic variation in tallness, highlighting the complexity and quantitative nature of this trait. In addition, the observed range, spanning 4.5 cm to 16.2 cm, indicates a broad genetic base, which is essential for successful GWAS and breeding programs aimed at improving plant height. This diversity is consistent with findings in other crops, such as rice and maize, where height is influenced by multiple genes with small effects (Huang \u0026amp; Han, 2016; Yu \u0026amp; Buckler, 2006). Polygenic traits often result in continuous phenotypic variation, which is exactly what we observed in this spinach population.\u003c/p\u003e\n\u003cp\u003eThe identification of particularly tall accessions, such as PI445784 and PI192945, suggests the presence of favorable alleles in these accessions that could be valuable in breeding programs. This finding aligns with studies on wheat and barley, where specific alleles have been identified and exploited to successfully enhance plant height (Chitwood et al., 2016).\u003c/p\u003e\n\u003cp\u003eThe coefficient of variation of 21.3% further indicates substantial phenotypic variability, which is beneficial for selection and increases the likelihood of detecting significant genetic associations. Similar levels of variability have proven advantageous in other crops, supporting the use of diverse panels in GWAS to identify key loci associated with target traits (Magar et al., 2021). Thus, the observation of substantial variability in this study confirms the suitability of this spinach panel for uncovering the genetic underpinnings of tallness and paves the way for more effective breeding strategies.\u003c/p\u003e\n\u003ch2\u003ePCA and phylogenetic analysis\u003c/h2\u003e\n\u003cp\u003eThe population structure and genetic diversity of spinach have been extensively explored using various methodologies, including SNP markers and phylogenetic analyses (Shi et al., 2016). Spinach exhibits significant variability in key traits essential for breeding and crop improvement, such as plant height and leaf morphology (Rashid et al., 2020b). In this study, we utilized high-density SNP data and principal component analysis (PCA) to assess the genetic diversity of 307 spinach accessions. The results revealed three distinct sub-populations, consistent with earlier studies that highlighted the complex genetic structure of spinach germplasm (Patterson et al., 2006; Saitou \u0026amp; Nei, 1987).\u003c/p\u003e\n\u003ch2\u003eAssociation study\u003c/h2\u003e\n\u003cp\u003eIn the association study on spinach tallness, which utilized multiple models (MLM, GLM, FarmCPU, and BLINK) within the GAPIT3 framework, we detected consistent deviations in the QQ plots, suggesting that the identified SNPs likely contribute to the observed phenotypic variation in height. This finding echoes previous research on other crops, such as rice, maize, and wheat, where plant height has been demonstrated as a complex polygenic trait influenced by multiple loci (Huang \u0026amp; Han, 2016; Yu \u0026amp; Buckler, 2006). The identification of significant SNPs in these crops has been crucial not only for understanding the genetic basis of height but also for guiding breeding programs aimed at improving this trait. Specifically, certain alleles have been exploited to enhance barley stature, further illustrating the value of SNP identification for crop improvement (Chitwood et al., 2016).\u003c/p\u003e\n\u003cp\u003eThe present GWAS identified six SNPs that exceeded the significance threshold, making them suitable candidates for marker-assisted selection. Targeted breeding based on genetic markers has been successfully applied in crops such as maize, rice, and wheat to develop superior cultivars with improved yield and adaptability (Collard \u0026amp; Mackill, 2008; Huang \u0026amp; Han, 2016; Yu \u0026amp; Buckler, 2006).\u003c/p\u003e\n\u003ch2\u003eCandidate gene identification/detection\u003c/h2\u003e\n\u003cp\u003eIn this study, four SNP markers were identified within 50 kb of five candidate genes, suggesting these genes may play important roles in controlling plant height. Both \u003cem\u003eSOV4g016060\u003c/em\u003e (U6 snRNA-associated Sm-like protein LSm5), located near SOVchr4_38323167\u003cem\u003e\u0026nbsp;\u003c/em\u003eon chromosome 4, and \u003cem\u003eSOV5g028680\u003c/em\u003e (cleavage and polyadenylation specificity factor subunit 2), near SOVchr5_70192260\u003cem\u003e\u0026nbsp;\u003c/em\u003eon chromosome 5, are involved in RNA processing\u0026mdash;a critical function previously linked to growth regulation in multiple crops. For instance, in maize, genes involved in RNA processing have been shown to influence plant height by regulating the expression of growth-related genes (Huang \u0026amp; Han, 2016). Similarly, in rice, RNA processing genes can affect both height and yield (Gong et al., 2021). In barley, genes associated with RNA processing have been found to regulate flowering time and overall plant stature (Nitcher et al., 2013).\u003c/p\u003e\n\u003cp\u003eTwo additional candidate genes on chromosome 4, \u003cem\u003eSOV4g059190\u003c/em\u003e (outer envelope membrane protein 7-like) and \u003cem\u003eSOV4g059200\u003c/em\u003e (epimerase domain-containing protein), both located near SOVchr4_188084338, are associated with metabolic and transport processes. These processes are crucial for cell elongation and biomass accumulation, as previously evidenced in rice and maize. In rice, genes related to metabolic pathways have been linked to the regulation of internode elongation, a key factor in determining plant height (Yu \u0026amp; Buckler, 2006). In maize, the transport of nutrients and growth regulators is critical for the development of tall plants (H\u0026uuml;tsch \u0026amp; Schubert, 2018).\u003c/p\u003e\n\u003cp\u003eThe last candidate gene identified in this study, \u003cem\u003eSOV6g020520\u003c/em\u003e (LETM1 and EF-hand domain-containing protein 1 mitochondrial), located near SOVchr6_91175684 on chromosome 6, suggests a role for mitochondrial function in regulating spinach height. Mitochondria are essential for energy production, which is necessary to sustain the metabolic demands of growing plants. Studies in wheat and barley have demonstrated that mitochondrial function is closely linked to plant vigor and height, with efficient energy production supporting taller growth (Lozano et al., 2009).\u003c/p\u003e\n\u003ch2\u003eGenomic prediction for genomic selection of tallness trait\u003c/h2\u003e\n\u003cp\u003eIntegration of GP models into breeding programs has become an essential tool for enhancing crop traits, such as plant height, through genomic selection. In this study, we evaluated the performance of seven GP models in predicting tallness in 307 spinach accessions using both randomly selected SNP sets and GWAS-derived SNP marker sets.\u003c/p\u003e\n\u003cp\u003eThe seven GP models\u0026mdash;BA, BB, BL, BRR, rrBLUP, RF, and SVM\u0026mdash;showed similar r-values across SNP sets from r6 to all.15058SNPs, averaging from 0.08 (r6) to 0.16 (r1000) (Table S2, Fig. 5 \u0026amp; S3). The r-value generally increased as the number of SNPs in the set increased, but the improvement plateaued after 1,000 SNPs. The results demonstrated that increasing the number of SNPs from six to 15,058 led to a progressive rise in prediction accuracy (r-value), stabilizing around 1,000 SNPs across all models. These findings underscore the necessity of utilizing a sufficient number of markers to achieve reliable predictions, consistent with previous research emphasizing the importance of genome-wide coverage for accurate GP (Heslot et al., 2012). However, all r-values were low, indicating that GP may not be efficient for predicting the tallness trait.\u003c/p\u003e\n\u003cp\u003eDespite the general trend of higher SNP numbers correlating with improved accuracy, the GWAS-derived SNP set (m6), comprising only six markers, exhibited a relatively high average r-value of 0.49, with the BRR model achieving a peak r-value of 0.53. These outcomes suggest that even a small number of strategically selected SNPs can provide substantial predictive power, particularly when those markers are closely linked to the trait of interest (Zhao et al., 2021). This parallels findings in other crops, where GWAS-derived markers have proven instrumental in improving prediction models for complex traits (Minamikawa et al., 2021).\u003c/p\u003e\n\u003cp\u003eAmong the seven GP models assessed, the BRR model consistently outperformed the others, particularly when using the m6 SNP set and other GWAS-derived SNP marker sets. This result suggests that the BRR model is suitable for predicting tallness in spinach. The model\u0026rsquo;s superior performance may be attributed to its robust ability to capture additive genetic variance, a key factor in polygenic traits like plant height, where the contributing genes have individually small effects (Crossa et al., 2017; Goddard \u0026amp; Hayes, 2007).\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003ePhenotypic data revealed significant variability in plant height (tallness), with seven accessions exhibiting exceptional tallness identified as promising candidates for breeding programs. Six single nucleotide polymorphisms (SNPs) on chromosomes 2, 4, 5, and 6 were strongly associated with tallness, with notable contributions from markers on chromosome 6. Five candidate genes, involved in RNA processing, metabolic pathways, and mitochondrial function, were located within 50 kb of these SNPs, warranting further investigation as potential regulators of the tallness trait. The genomic prediction (GP) models, particularly the BRR model, demonstrated high predictive accuracy, even when applied to a small GWAS-derived SNP set. This finding supports the utility of these markers for forecasting genetic potential for plant height. These insights offer breeders valuable tools for targeted selection and genotyping, facilitating the development of spinach varieties optimized for mechanical harvesting and market preferences. By integrating genomic insights with traditional breeding methods, this study establishes a strong foundation for future strategies aimed at improving spinach height, contributing to more sustainable and economically viable agricultural practices.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAS and BM designed the research and experimental strategy. BM, IA, HA, QL, KC, YQ, RD, AW, and DJH conducted the experiments. AS, IA, and HX analyzed the data. IA drafted and HX and AS edited and revised the manuscript. All authors contributed to the study and approved the final manuscript for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded by the USDA-NIFA SCRI project # 2023-51181-41321; USDA NIFA Hatch (Project numbers ARK0VG2018, ARK02440, and ARK02609), and a scholarship from the Saudi Arabia government, the Saudi Arabian Cultural Mission (SACM), and the University of Tabuk, Saudi Arabia.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by the USDA-NIFA SCRI and USDA NIFA Hatch grants. The authors are grateful to the scientists who contributed to this project, as well as to the reviewers and editors for their constructive feedback.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data presented in this study are available within the article and its Supplementary Material. Whole genome resequencing (WGR) data aligned with the reference genome are available at NCBI under BioProject ID: PRJNA860974. The SNP datasets can be accessed through Tables, Figures, Supplementary Tables, and Supplementary Figures. The SNP data are also available on FigShare at https://figshare.com/account/articles/28603517. The accession numbers used in this study are provided in the article and Supplementary Materials for reference.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBarili, L D, Vale, N M do, Silva, F F e, Carneiro, J E de S, Oliveira, H R de, Vianello, R P, Valdisser, P A M R, \u0026amp; Nascimento, M (2018) Genome prediction accuracy of common bean via Bayesian models \u003cem\u003eCi\u0026ecirc;ncia Rural\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(8) https://doiorg/101590/0103-8478cr20170497\u003c/li\u003e\n\u003cli\u003eBhattarai, G, \u0026amp; Shi, A (2021) Research advances and prospects of spinach breeding, genetics, and genomics \u003cem\u003eVegetable Research\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(1), 1\u0026ndash;18\u003c/li\u003e\n\u003cli\u003eCai, X, Xu, C, Wang, X, Wang, S, Zhang, Z, Fei, Z, \u0026amp; Wang, Q (2018) Construction of genetic linkage map using genotyping-by-sequencing and identification of QTLs associated with leaf color in spinach \u003cem\u003eEuphytica\u003c/em\u003e, \u003cem\u003e214\u003c/em\u003e, 1\u0026ndash;11\u003c/li\u003e\n\u003cli\u003eChitwood, J, Shi, A, Mou, B, Evans, M, Clark, J, Motes, D, Chen, P, \u0026amp; Hensley, D (2016) Population structure and association analysis of bolting, plant height, and leaf erectness in spinach \u003cem\u003eHortScience\u003c/em\u003e, \u003cem\u003e51\u003c/em\u003e(5) https://doiorg/1021273/hortsci515481\u003c/li\u003e\n\u003cli\u003eCollard, B C Y, \u0026amp; Mackill, D J (2008) Marker-assisted selection: An approach for precision plant breeding in the twenty-first century In \u003cem\u003ePhilosophical Transactions of the Royal Society B: Biological Sciences\u003c/em\u003e (Vol 363, Issue 1491) https://doiorg/101098/rstb20072170\u003c/li\u003e\n\u003cli\u003eCollins, K, Zhao, K, Jiao, C, Xu, C, Cai, X, Wang, X, Ge, C, Dai, S, Wang, Q, Wang, Q, Fei, Z, \u0026amp; Zheng, Y (2019) SpinachBase: A central portal for spinach genomics \u003cem\u003eDatabase\u003c/em\u003e, \u003cem\u003e2019\u003c/em\u003e(1) https://doiorg/101093/database/baz072\u003c/li\u003e\n\u003cli\u003eCrossa, J, P\u0026eacute;rez, P, Hickey, J, Burgue\u0026ntilde;o, J, Ornella, L, Cer\u0026oacute;n-Rojas, J, \u0026amp; de los Campos, G (2017) Genomic prediction in CIMMYT maize and wheat breeding programs \u003cem\u003eHeredity\u003c/em\u003e, \u003cem\u003e112\u003c/em\u003e(1), 48\u0026ndash;60\u003c/li\u003e\n\u003cli\u003eFrary, A, Gul, D, Keles, D, others, \u0026amp; Doganlar, S (2010) Salt Tolerance in Solanum Pennellii: Antioxidant Response and Related QTL \u003cem\u003eBMC Plant Biology\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(1), 1\u0026ndash;16\u003c/li\u003e\n\u003cli\u003eGaynor, R C, Gorjanc, G, Bentley, A R, Ober, E S, Howell, P, Jackson, R, Mackay, I J, \u0026amp; Hickey, J M (2017) A two-part strategy for using genomic selection to develop inbred lines \u003cem\u003eCrop Science\u003c/em\u003e, \u003cem\u003e57\u003c/em\u003e(5) https://doiorg/102135/cropsci2016090742\u003c/li\u003e\n\u003cli\u003eGoddard, M E, \u0026amp; Hayes, B J (2007) Genomic selection based on dense genotypes inferred from sparse genotypes \u003cem\u003eJournal of Animal Breeding and Genetics\u003c/em\u003e, \u003cem\u003e124\u003c/em\u003e(6), 423\u0026ndash;430\u003c/li\u003e\n\u003cli\u003eGong, L sha, Qu, S jie, Huang, G min, Guo, Y ling, Zhang, M cai, Li, Z hu, Zhou, Y yi, \u0026amp; Duan, L sheng (2021) Improving maize grain yield by formulating plant growth regulator strategies in North China \u003cem\u003eJournal of Integrative Agriculture\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(2) https://doiorg/101016/S2095-3119(20)63453-0\u003c/li\u003e\n\u003cli\u003eHeffner, E L, Sorrells, M E, \u0026amp; Jannink, J L (2009) Genomic selection for crop improvement \u003cem\u003eCrop Science\u003c/em\u003e, \u003cem\u003e49\u003c/em\u003e(1), 1\u0026ndash;12\u003c/li\u003e\n\u003cli\u003eHeslot, N, Yang, H P, Sorrells, M E, \u0026amp; Jannink, J L (2012) Genomic selection in plant breeding: A comparison of models \u003cem\u003eCrop Science\u003c/em\u003e, \u003cem\u003e52\u003c/em\u003e(1) https://doiorg/102135/cropsci2011060297\u003c/li\u003e\n\u003cli\u003eHuang, X, \u0026amp; Han, B (2016) Natural variations and genome-wide association studies in crop plants \u003cem\u003eAnnual Review of Plant Biology\u003c/em\u003e, \u003cem\u003e67\u003c/em\u003e, 101\u0026ndash;122 https://doiorg/101146/annurev-arplant-043015-111846\u003c/li\u003e\n\u003cli\u003eH\u0026uuml;tsch, B W, \u0026amp; Schubert, S (2018) Maize harvest index and water use efficiency can be improved by inhibition of gibberellin biosynthesis \u003cem\u003eJournal of Agronomy and Crop Science\u003c/em\u003e, \u003cem\u003e204\u003c/em\u003e(2) https://doiorg/101111/jac12250\u003c/li\u003e\n\u003cli\u003eJones, R B, Phillips, S, \u0026amp; Patterson, R (2019) Mechanization in leafy vegetable production: Impacts on crop yield and quality \u003cem\u003eAgronomy Journal\u003c/em\u003e, \u003cem\u003e111\u003c/em\u003e(4), 1534\u0026ndash;1542\u003c/li\u003e\n\u003cli\u003eLegarra, A, Robert-Grani\u0026eacute;, C, Croiseau, P, Guillaume, F, \u0026amp; Fritz, S (2011) Improved Lasso for genomic selection \u003cem\u003eGenetics Research\u003c/em\u003e, \u003cem\u003e93\u003c/em\u003e(1) https://doiorg/101017/S0016672310000534\u003c/li\u003e\n\u003cli\u003eLi, H, \u0026amp; Durbin, R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform \u003cem\u003eBioinformatics\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(14) https://doiorg/101093/bioinformatics/btp324\u003c/li\u003e\n\u003cli\u003eLozano, R, Gim\u0026eacute;nez, E, Cara, B, Capel, J, \u0026amp; Angosto, T (2009) Genetic analysis of reproductive development in tomato In \u003cem\u003eInternational Journal of Developmental Biology\u003c/em\u003e (Vol 53, Issues 8\u0026ndash;10) https://doiorg/101387/ijdb072440rl\u003c/li\u003e\n\u003cli\u003eMagar, B T, Acharya, S, Gyawali, B, Timilsena, K, Upadhayaya, J, \u0026amp; Shrestha, J (2021) Genetic variability and trait association in maize (Zea mays L) varieties for growth and yield traits \u003cem\u003eHeliyon\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(9) https://doiorg/101016/jheliyon2021e07939\u003c/li\u003e\n\u003cli\u003eMcKenna, A, Hanna, M, Banks, E, Sivachenko, A, Cibulskis, K, Kernytsky, A, Garimella, K, Altshuler, D, Gabriel, S, Daly, M, \u0026amp; DePristo, M A (2010) The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data \u003cem\u003eGenome Research\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(9) https://doiorg/101101/gr107524110\u003c/li\u003e\n\u003cli\u003eMinamikawa, M F, Kunihisa, M, Noshita, K, Moriya, S, Abe, K, Hayashi, T, Katayose, Y, Matsumoto, T, Nishitani, C, Terakami, S, Yamamoto, T, \u0026amp; Iwata, H (2021) Tracing founder haplotypes of Japanese apple varieties: application in genomic prediction and genome-wide association study \u003cem\u003eHorticulture Research\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1) https://doiorg/101038/s41438-021-00485-3\u003c/li\u003e\n\u003cli\u003eNitcher, R, Distelfeld, A, Tan, C, Yan, L, \u0026amp; Dubcovsky, J (2013) Increased copy number at the HvFT1 locus is associated with accelerated flowering time in barley \u003cem\u003eMolecular Genetics and Genomics\u003c/em\u003e, \u003cem\u003e288\u003c/em\u003e(5\u0026ndash;6) https://doiorg/101007/s00438-013-0746-8\u003c/li\u003e\n\u003cli\u003ePatterson, N, Price, A L, \u0026amp; Reich, D (2006) Population structure and eigenanalysis \u003cem\u003ePLoS Genetics\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(12) https://doiorg/101371/journalpgen0020190\u003c/li\u003e\n\u003cli\u003eRashid, M, Yousaf, Z, Ullah, M N, Munawar, M, Riaz, N, Younas, A, Aftab, A, \u0026amp; Shamsheer, B (2020a) Genetic variability assessment of worldwide spinach accessions by agro-morphological traits \u003cem\u003eJournal of Taibah University for Science\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 1637\u0026ndash;1650\u003c/li\u003e\n\u003cli\u003eRashid, M, Yousaf, Z, Ullah, M N, Munawar, M, Riaz, N, Younas, A, Aftab, A, \u0026amp; Shamsheer, B (2020b) Genetic variability assessment of worldwide spinach accessions by agro-morphological traits \u003cem\u003eJournal of Taibah University for Science\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 1637\u0026ndash;1650\u003c/li\u003e\n\u003cli\u003eRavelombola, W, Shi, A, and Huynh, B (2021) Loci discovery, network-guided approach, and genomic prediction for drought tolerance index in a multi-parent advanced generation intercross (MAGIC) cowpea population Hortic Res 8:24 https://doiorg/101038/s41438-021-00462-w\u003c/li\u003e\n\u003cli\u003eRavelombola, W, Qin, J, Shi, A, Nice, L, Bao, Y, Lorenz, A, Orf, JH, Young, ND, and Chen, S (2019) Genome-wide association study and genomic selection for soybean chlorophyll content associated with soybean cyst nematode BMC Genomics, 20, 904 https://doiorg/101186/s12864-019-6275-z\u003c/li\u003e\n\u003cli\u003eRavelombola, W, Qin, J, Shi, A, Nice, L, Bao, Y, Lorenz, A, Orf, JH, Young, ND, and Chen, S (2020) Genome-wide association study and genomic selection for tolerance of soybean biomass reduction under to soybean cyst nematode infestation Plos One doiorg/101371/journalpone0235089\u003c/li\u003e\n\u003cli\u003eSaitou, N, \u0026amp; Nei, M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees \u003cem\u003eMolecular Biology and Evolution\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(4) https://doiorg/101093/oxfordjournalsmolbeva040454\u003c/li\u003e\n\u003cli\u003eShi, A, Mou, B, \u0026amp; Cheng, Z M (2016) Genetic diversity and association analysis of leafminer (Liriomyza langei) resistance in spinach (Spinacia oleracea) \u003cem\u003eGenome\u003c/em\u003e, \u003cem\u003e59\u003c/em\u003e(8) https://doiorg/101139/gen-2016-0075\u003c/li\u003e\n\u003cli\u003eShi, A, P Gepts, Q Song, H Xiong, TE Michaels, and S Chen 2021 Genome-wide association study and genomic selection for soybean cyst nematode resistance in USDA common bean (Phaseolus vulgaris) core collection Frontiers in Plant Science, 07, https://doiorg/103389/fpls2021624156\u003c/li\u003e\n\u003cli\u003eShi, A, G Bhattarai, H Xiong, CA Avila, C Feng, B Liu, V Joshi, L Stein, B Mou, LJ du Toit, and JC Correll 2022 Genome-wide association study and genomic prediction of white rust resistance in USDA GRIN Spinach Germplasm Horticulture Research, Volume 9, 2022, uhac069, https://doiorg/101093/hr/uhac069\u003c/li\u003e\n\u003cli\u003eTang, X, Mu, X, Shao, H, Wang, H, \u0026amp; Brestic, M (2015) Global plant-responding mechanisms to salt stress: Physiological and molecular levels and implications in biotechnology In \u003cem\u003eCritical Reviews in Biotechnology\u003c/em\u003e (Vol 35, Issue 4) https://doiorg/103109/073885512014889080\u003c/li\u003e\n\u003cli\u003eVan Dijk, E L, Jaszczyszyn, Y, \u0026amp; Thermes, C (2014) Library preparation methods for next-generation sequencing: Tone down the bias In \u003cem\u003eExperimental Cell Research\u003c/em\u003e (Vol 322, Issue 1) https://doiorg/101016/jyexcr201401008\u003c/li\u003e\n\u003cli\u003eWang, J, \u0026amp; Zhang, Z (2021) GAPIT Version 3: Boosting Power and Accuracy for Genomic Association and Prediction \u003cem\u003eGenomics, Proteomics and Bioinformatics\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(4) https://doiorg/101016/jgpb202108005\u003c/li\u003e\n\u003cli\u003eYu, J, \u0026amp; Buckler, E S (2006) Genetic association mapping and genome organization of maize In \u003cem\u003eCurrent Opinion in Biotechnology\u003c/em\u003e (Vol 17, Issue 2) https://doiorg/101016/jcopbio200602003\u003c/li\u003e\n\u003cli\u003eZhao, S, Zhang, Q, Liu, M, Zhou, H, Ma, C, \u0026amp; Wang, P (2021) Regulation of plant responses to salt stress In \u003cem\u003eInternational Journal of Molecular Sciences\u003c/em\u003e (Vol 22, Issue 9) MDPI AG https://doiorg/103390/ijms22094609\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Genome-wide association study (GWAS), genomic prediction (GP), plant height, single-nucleotide polymorphism (SNP), Spinacia oleracea L., spinach, tallness","lastPublishedDoi":"10.21203/rs.3.rs-6305818/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6305818/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePlant height (tallness) is a crucial agronomic trait in spinach (Spinacia oleracea L.), affecting mechanical harvesting efficiency and overall yield. This study evaluated plant height variation in a panel of 307 USDA germplasm accessions, which were phenotyped for this trait and genotyped using 15,058 single-nucleotide polymorphisms (SNPs) from whole-genome resequencing. A genome-wide association study (GWAS) using GLM, MLM, FarmCPU, and BLINK in GAPIT3 identified six SNPs significantly associated with plant height: SOVchr2_68062488 (68,062,488 bp) on chromosome 2; SOVchr4_38323167 (38,323,167 bp) and SOVchr4_188084338 (188,084,338 bp) on chromosome 4; SOVchr5_70192260 (70,192,260 bp) on chromosome 5; and SOVchr6_8139833 (8,139,833 bp) and SOVchr6_91175684 (91,175,684 bp) on chromosome 6. Additionally, genomic prediction (GP) models estimated genomic estimated breeding values (GEBVs) for plant height, achieving r-values of 0.50 using GWAS-derived SNP markers in cross-population prediction. The integration of GWAS and GP provides valuable insights into the genetic architecture of plant height in spinach, supporting marker-assisted breeding strategies to enhance crop management and economic returns.\u003c/p\u003e","manuscriptTitle":"Genome-wide association study and genome prediction of tallness trait in spinach tallness phenotyping","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-30 05:33:59","doi":"10.21203/rs.3.rs-6305818/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d8f8d959-1c7f-4c4c-b24f-4a4cf324ff48","owner":[],"postedDate":"June 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-06-30T05:33:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-30 05:33:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6305818","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6305818","identity":"rs-6305818","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}