Skip to main content
NCBI home page
BMC Plant Biology logoLink to BMC Plant Biology
. 2026 May 6;26:1089. doi: 10.1186/s12870-026-08884-1

Identifying candidate genes for grain protein content in wheat (Triticum aestivum L.) by integrated GWAS and transcriptome analysis

Junxiu Shang 1,2, Yuhong Diao 1,2, Yan Wang 1,2, Xiuqing Zhang 1,2, Xiaohua Yao 1,2, Youhua Yao 1,2,✉
PMCID: PMC13312571  PMID: 42092796

Abstract

Background

Grain protein content and composition are key determinants of wheat nutritional and processing quality, directly influencing the processing characteristics and nutritional value of flour-based products. This trait is regulated by multiple genes in a coordinated manner and is susceptible to genotype-by-environment interactions, resulting in a complex genetic architecture. Elucidating the genetic basis of grain protein content is therefore of great significance for the breeding of high-quality specialty wheat varieties.

Results

In this study, a set of 260 widely collected spring wheat germplasm accessions was used. The crude protein (CP) and soluble protein (SP) contents were measured under four environmental conditions. Using genome-wide association study (GWAS), a total of 20 genetic loci significantly associated with CP and 23 loci significantly associated with SP were identified. Based on the linkage disequilibrium (LD) decay distance (590 kb), LD block analysis was performed, and four core loci located within the LD blocks were selected: two CP-related loci, 2B-161,726,146 and 2D-610,941,094, and two SP-related loci, 1B-562,451,687 and 7A-4,591,015. The medium-gluten wheat Humai 14 (HM14) and strong-gluten wheat Xinchun 48 (XC48), which exhibited significant differences in both CP and SP across the four environments, were selected for transcriptome sequencing at five developmental stages of grain (10, 15, 20, 25, and 30 days after anthesis). A total of 27,499 differentially expressed genes were identified, with two expression peaks observed at 10 and 30 days after anthesis. By integrating GWAS and transcriptomic differentially expressed genes, combined with gene functional annotation and expression pattern analysis, five candidate genes associated with grain protein content were ultimately screened: TaPAE2-like, TaHIP1, TaGlu-B1al-like, Ta1-SST, and Td1-SST-like.

Conclusions

This study provides genetic resources for elucidating the genetic mechanisms of grain protein synthesis in wheat and lays a crucial foundation for molecular marker-assisted breeding.

Keywords: Wheat, Crude protein, Soluble protein, Transcriptome, Candidate genes

Background

Wheat (Triticum aestivum L.) is one of the most important staple crops worldwide, serving as a primary source of energy and protein for human consumption [1]. With the continuous growth of the global population, shifts in dietary patterns, and the rapid development of the food processing industry, market demands for wheat quality have increasingly intensified. In particular, grain protein content and composition directly determine the processing quality and nutritional functionality of wheat-based products [2, 3]. Crude protein is a fundamental indicator for evaluating wheat nutritional quality, directly reflecting the capacity for nitrogen uptake, translocation, and accumulation during the wheat growth period [4, 5]. Soluble protein represents the proportion of nitrogen that remains in a soluble state under specific conditions [6, 7]. In plants and microorganisms, soluble proteins function as osmotic regulators, enhancing cellular water retention capacity and protecting biological membranes, and are therefore commonly used as indicators of plant stress resistance [6, 8]. However, in food science, soluble proteins contribute to nutritional value, emulsification, and functional properties [9]. Both CP and SP are typical complex quantitative traits, governed by multiple genes and susceptible to genotype-by-environment interactions [10]. In recent years, GWAS has emerged as a powerful approach for dissecting the associations between traits and genetic variants at the genomic level by leveraging the historical recombination and genetic diversity of natural populations, and has become a key tool for identifying quantitative trait nucleotides (QTNs) and candidate genes controlling important traits [11, 12]. Compared with bi-parental linkage mapping, GWAS offers higher mapping resolution and is particularly suitable for multi-parental or natural populations [13]. To date, numerous studies have employed GWAS to dissect the genetic basis of grain protein-related traits in wheat. For instance, most quantitative trait loci (QTLs) for grain protein content have been reported on chromosomes 6B and 7B [14, 15], among which the Gpc-B1 marker on chromosome 6B has been utilized in wheat quality breeding programs [16]. GWAS studies have also revealed that the QGpc.ipk-6 A locus on chromosome 6 A negatively regulates grain protein content [17]. More recently, a study on durum wheat identified a novel genetic marker significantly associated with protein content on chromosome 4B and determined that the serine acetyltransferase 2 gene (SAT2) is a key candidate gene. In this gene, the missense mutation Gly325Ser in the ninth exon and an intronic SNP individually increased protein content by 1.33% and 0.92%, respectively [18]. However, existing studies have predominantly focused on total protein or storage proteins, while research on soluble protein—which is closely associated with processing quality—remains largely confined to the realms of plant physiology and stress responses. A systematic understanding of the genetic regulation of soluble protein in wheat grains is still lacking, a gap that constrains the breeding progress of high-quality specialty wheat varieties.

To address the issues mentioned above, this study employed a multi‑locus model within the framework of GWAS. Compared with conventional single-locus models, the multi-locus approach incorporates all potential associated markers identified during initial screening into a single model for joint estimation [19]. This model specification more closely reflects the true genetic architecture of quantitative traits governed by multiple genes [19, 20]. This strategy not only inherently avoids the omission of weak but genuine signals caused by stringent multiple testing corrections such as Bonferroni adjustment due to its shrinkage properties, but also demonstrates robust control over background genetic variation when analyzing natural populations with complex population structure [19]. By integrating polygenic background information with variable selection methods, the multi-locus model effectively distinguishes population stratification from true association signals without relying on traditional stratification correction approaches, thereby achieving a better balance between detection power and false-positive control [21]. Furthermore, this study integrated transcriptome sequencing (RNA-seq) with GWAS in a joint analysis strategy, enabling candidate gene validation at the level of expression regulation. This approach has been successfully applied to the discovery and functional validation of key genes in various crops [22, 23], with the aim of screening candidate genes co-expressed with association signals.

Therefore, this study employed 260 wheat germplasm accessions as research materials. Crude protein content and soluble protein content were measured across four environments, and a mrMLM model was applied to perform GWAS for these two traits based on high-density SNP array data. By integrating LD block partitioning, gene functional annotation, and transcriptome data from different grain developmental stages, the genomic intervals harboring association signals were delineated, and key candidate genes regulating grain protein content in wheat were identified. This study provides theoretical foundations and genetic resources for molecular design breeding of high-quality wheat varieties.

Results

Phenotypic evaluation

To dissect the genotypic variation underlying grain protein content, phenotypic evaluation was conducted on 260 wheat germplasm resources across four distinct ecological environments (Table S1). All measured traits exhibited normal or skewed distributions (Fig. 1A, B), with descriptive statistics presented in Table 1. CP ranged from 8.92% to 20.98% across the four environments. Among them, E3 exhibited the lowest mean CP value (11.93%) and the highest coefficient of variation (CV =13.65%), indicating the greatest phenotypic dispersion under this environment, whereas E4 showed the lowest CV (8.35%). The mean CP value based on best linear unbiased prediction (BLUP) was 13.85%, with a CV of 7.02%, which was lower than that of individual environments, suggesting that multi-environment joint analysis effectively reduces environmental noise. SP ranged from 3.02% to 12.81%, with CV values ranging from 17.83% to 23.75%; the highest CV was observed in E1 (23.75%) and the lowest in E4 (17.83%). The mean SP value based on BLUP was 6.49%, with a CV of 13.98%. The broad-sense heritability (H²) was 0.75 for CP and 0.71 for SP, both classified as high heritability traits, indicating strong genetic control and suitability for GWAS.

Fig. 1.

Fig. 1

Open in a new tab

Frequency distribution and correlation heatmap of grain protein traits. A Normal distribution plot for CP. B Normal distribution plot for SP. C Heatmap of CP across four environments. D Heatmap of SP across four environments. E Correlation heatmap of CP, SP, and their BLUP values. Significance levels are indicated as P < 0.05 (*) and P < 0.01 (**)

Table 1.

Descriptive statistics for CP and SP traits in 260 wheat accessions

Trait Enviroment Mean±SD CV% Min Max Skewness Kurtosis H2 Environment
CP(%) E1 14.87±1.38a 9.25 11.99 20.47 0.97 1.48 0.75 P < 0.01
E2 14.11±1.30c 9.23 11.38 20.98 1.13 3.03
E3 11.93±1.63e 13.65 8.92 16.7 0.54 -0.28
E4 14.50±1.21b 8.35 11.75 18.24 0.36 -0.11
BLUP 13.85±0.97d 7.02 11.77 17.72 1.01 1.83
SP(%) E1 6.32±1.50b 23.75 3.68 12.81 1.25 1.95 0.71 P < 0.01
E2 6.67±1.51a 22.71 3.02 11.18 0.17 -0.22
E3 6.70±1.29a 19.22 3.55 10.05 0.25 -0.25
E4 6.29±1.12b 17.83 3.49 10.25 0.6 0.87
BLUP 6.49±0.91ab 13.98 4.22 9.11 0.29 0.2
Open in a new tab

CP Crude protein, SP Soluble protein

E1: Planted in 2023 at the Germplasm Resource Innovation Experimental Base of the College of Agriculture and Forestry, Qinghai University, Xining City;E2: Planted in 2024 at the Germplasm Resource Innovation Experimental Base of the College of Agriculture and Forestry, Qinghai University, Xining City;E3: Planted in 2024 at the Cereal Crops Experimental Base in Guide County, Hainan Tibetan Autonomous Prefecture, Qinghai Province;E4: Planted in 2024 at the Xiangride Town Experimental Base in Haixi Mongol and Tibetan Autonomous Prefecture, Qinghai Province. H2: broad-sense heritability, Different lowercase letters indicate significant differences (P < 0.05) among different ecological environments and BLUP values

Correlation analysis of grain protein content across four environments (Fig. 1C, D) showed that the correlation coefficients of CP ranged from 0.18 to 0.63, with extremely significant positive correlations observed in all environments (P < 0.01); the correlation coefficients of SP ranged from 0.29 to 0.47, and SP also exhibited extremely significant positive correlations across all environments (P < 0.01). Correlation analysis of CP, SP, and their BLUP values across the four environments (Fig. 1E) showed that the correlation coefficients ranged from − 0.11 to 0.79. Except for the significant positive correlations observed between E1-CP and E1-SP, E2-CP and E2-SP, E2-CP and E3-SP, and E2-CP and the BLUP value of SP (P < 0.05), all other correlations between CP and SP across different environments were not significant. Analysis of variance further indicated that the effects of environment on both CP and SP were highly significant (P < 0.01).

GWAS analysis for CP and SP traits

In this study, based on 182,549 high-quality SNP markers covering the 21 chromosomes of wheat, the LD decay distance was calculated, with the distance at which LD decayed to half being 590 kb (Fig. S1), which was subsequently used as the candidate interval for LD block analysis. Based on the BLUP values, GWAS for grain protein content was conducted using five models within the mrMLM package (FASTmrEMMA, ISIS EM-BLASSO, mrMLM, FASTmrMLM, and pLARmEB). The results identified 20 loci significantly associated with CP (Fig. 2A, B, Table S2). These significant loci were mainly distributed on chromosomes 2B, 2D, 3B, 3D, 6D, and 7B. Additionally, 23 loci significantly associated with SP were identified (Fig. 3A, B, Table S3), distributed across 18 chromosomes, including 1A, 1B, 1D, 2B, 2D, 3A, 3B, 3D, 4A, 4B, 5A, 5B, 5D, 6A, 6B, 6D, 7A, and 7D. To further delineate key genetic regions, LD block analysis was conducted using 590 kb upstream and downstream of each significant locus as candidate intervals, from which loci located within LD blocks with strong linkage intensity were selected. These included two CP-related loci, 2B-161,726,146 and 2D-610,941,094 (Fig. 2C, E), and two SP-related loci, 1B-562,451,687 and 7A-4,591,015 (Fig. 3C, E). Genotype–phenotype association analysis for these four loci (Figs. 2D and F and 3D and F) revealed that for the CP-related locus 2B-161,726,146, samples with the GG genotype had significantly higher CP than those with the CC genotype; for locus 2D-610,941,094, samples with the CG genotype had significantly higher CP than those with the CC genotype. For the SP-related locus 1B-562,451,687, samples with the AA genotype had significantly higher SP content than those with the CC genotype; for locus 7A-4,591,015, samples with the TT and TC genotypes had significantly higher SP than those with the CC genotype. The identification of these favorable allelic variants provides important molecular marker resources for wheat protein quality improvement.

Fig. 2.

Fig. 2

Open in a new tab

GWAS and LD Block Analysis for CP. A Manhattan plot for the mrMLM model; B Q-Q plot; C The locus 2B-161,726,146 associated with CP and its corresponding LD pattern; D Box plot displaying the distribution of the CP for SNP locus 2B-161,726,146 across different genotypes; E Significant association of the locus 2D-610,941,094 with CP and its corresponding LD pattern; F Box plot showing the distribution of CP at SNP locus 2D-610,941,094 across different genotypes

Fig. 3.

Fig. 3

Open in a new tab

GWAS and LD Block Analysis for SP. A Manhattan plot for the mrMLM model; B Q-Q plot; C The locus 1B-562,451,687 associated with SP and its corresponding LD pattern; D Box plot displaying the distribution of the SP for SNP locus 1B-562,451,687 across different genotypes; E Significant association of the locus 7A-4,591,015 with SP and its corresponding LD pattern; F Box plot showing the distribution of SP at SNP locus 7A-4,591,015 across different genotypes

Transcriptome analysis of CP and SP

To explore the transcriptional regulatory basis of grain protein formation in wheat, a pair of wheat varieties with contrasting grain protein traits across all four environments was selected from the 260 accessions evaluated. These were the strong-gluten wheat variety XC48 and the medium-gluten wheat variety HM14 (Fig. 4A). When |log2FC| ≥ 1 and padjust < 0.01, genes were considered significantly differentially expressed. A total of 28,884 up-regulated genes and 18,886 down-regulated genes were identified, resulting in 27,499 differentially expressed genes(DEGs) after removing duplicates (Fig. 4B, C; Table S4). At 10 days after anthesis (DAA), the number of DEGs between the two varieties reached a peak (16,573), indicating that this stage represents a critical phase for variety trait differentiation. Subsequently, the number of DEGs exhibited a decreasing trend, declining from 10,140 at 15 DAA to 4,640 at 25 DAA, reflecting a gradual reduction in transcriptional differences between the two varieties as grain development progressed. By 30 DAA, the number of DEGs increased again to 10,409, suggesting that distinct gene regulatory pathways may be activated in the two varieties during the late stage of grain maturation. Collectively, these results indicate that the expression differences of grain protein-related genes between HM14 and XC48 exhibit pronounced developmental stage specificity, being particularly significant at both the early and late stages of development. Heatmap analysis further classified these DEGs into two distinct expression patterns, suggesting that they play different regulatory roles during grain protein formation (Fig. 4D).

Fig. 4.

Fig. 4

Open in a new tab

Comparative RNA-seq analysis of grain development in HM14 and XC48 wheat cultivars. A Phenotypic comparison of the two cultivars across four environments (E1–E4); B Venn diagram of DEGs at different developmental stages; C Statistical count of DEGs between the two cultivars at various developmental time points; D Clustering heatmap of expression patterns for all DEGs

To elucidate the functions of these 27,499 DEGs, Gene Ontology (GO) enrichment analysis was performed. A total of 2,130 GO terms were identified, including 1,245 associated with biological process (BP), 166 with cellular component (CC), and 719 with molecular function (MF) (Fig. 5A, Table S5). Within the BP category, “carbohydrate metabolic process” was the most prominent, encompassing 1,070 genes. For CC, “cellular anatomical entity” exhibited the highest enrichment score, involving 11,442 genes. For MF, “nutrient reservoir activity” received the highest score, comprising 119 genes. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used to further elucidate the biological functions and interactions of the identified genes (Fig. 5B, Table S6). Among the KEGG annotation terms, the three most significantly enriched pathways were “Photosynthesis – antenna proteins,” “Starch and sucrose metabolism,” and “Plant hormone signal transduction,” all of which were significantly associated with the target traits under investigation.

Fig. 5.

Fig. 5

Open in a new tab

GO and KEGG enrichment analysis. A GO analysis: the x axis represents GO terms, and the y axis shows the enrichment factor. B KEGG analysis: the y axis indicates pathway names, and the x axis displays the ratio of the enrichment factor to the number of annotated genes (background count). A larger enrichment factor reflects a higher degree of enrichment. The dot size corresponds to the number of genes in the respective pathway, while the dot color represents different ranges of the adjusted P value (padjust)

Candidate gene analysis for CP and SP

To screen candidate genes associated with grain protein formation, we integrated GWAS and DEGs analyses. Based on LD analysis, 18 and 24 genes were identified within the 590-kb upstream and downstream intervals flanking the significant loci associated with CP and SP , respectively (Table S7). To focus on key genes, we performed an intersection analysis between the genes located within LD blocks identified by GWAS and the DEGs, yielding 21 overlapping genes. Their annotation information, enrichment terms, and expression levels are detailed in Table S8. Based on gene functional annotation, GO and KEGG enrichment pathways, and DEGs between the two varieties, candidate genes associated with CP and SP were further screened. Among them, two genes related to CP were identified: TaPAE2-like (TraesCS2B03G0433900), annotated as pectin acetylesterase 2-like; and TaHIP1 (TraesCS2D03G1151900), annotated as E3 ubiquitin-protein ligase HIP1. The genes associated with SP included TaGlu-B1al-like (TraesCS1B03G0904700), annotated as high molecular weight glutenin subunit DX5-like; Ta1-SST (TraesCS7A03G0018100), annotated as sucrose: sucrose 1-fructosyltransferase; and Td1-SST-like (TraesCS7A03G0019700), annotated as sucrose: sucrose 1-fructosyltransferase-like. Td1-SST-like is a homologous isoform with functions similar to Ta1-SST and is also involved in carbohydrate metabolism and fructan biosynthesis-related processes.

To validate the reliability of the RNA-seq data, quantitative real-time PCR (qRT-PCR) was performed for these five genes (Figs. 6B and C and 7B, C and D), and their expression patterns across different grain developmental stages in the transcriptome were analyzed (Figs. 6A and 7A). The results showed that the expression trends from qRT-PCR were highly consistent with those from transcriptome analysis, confirming the reliability of the data and suggesting that these genes may play important roles in grain protein formation.

Fig. 6.

Fig. 6

Open in a new tab

Expression patterns of two candidate genes for CP. A Heatmap showing the transcriptome expression levels of two candidate genes in HM14 and XC48 at five time points from 10 to 30 DAA. B Relative expression levels of the TaPAE2-like gene validated by qRT-PCR. C Relative expression levels of the TaHIP1 gene validated by qRT-PCR. Error bars represent the standard deviation (SD) of three technical replicates.

Fig. 7.

Fig. 7

Open in a new tab

Expression patterns of three candidate genes associated with SP. A Heatmap showing the transcriptome expression levels of three candidate genes in HM14 and XC48 at five time points from 10 to 30 DAA. B Relative expression levels of the TaGlu-B1al-like gene validated by qRT-PCR. C Relative expression levels of the Ta1-SST gene validated by qRT-PCR. D Relative expression levels of the Ta1-SST-like gene validated by qRT-PCR. Error bars represent the SD of three technical replicates.

Discussion

Improving grain protein content in wheat is one of the key objectives for enhancing its nutritional and processing quality [24]. Both CP and SP exhibited continuous phenotypic variation across the four environments, and all traits displayed normal or skewed distributions, consistent with the characteristics of quantitative traits, thereby providing a solid phenotypic foundation for subsequent GWAS analysis [25]. Analysis of variance indicated that environment had significant effects on both CP and SP, and correlation analysis among traits further revealed their complex genetic networks. Grain protein accumulation is a complex quantitative trait regulated by multiple genes in a coordinated manner and is subject to strong genotype–environment interactions [26, 27]. In addition, based on multi-environment phenotypic data, we calculated the BLUP values for each trait and used them as integrated phenotypic values for subsequent GWAS analysis, aiming to consolidate genotypic effects and reduce environmental noise, thereby improving the accuracy of genetic locus detection [28].

Through GWAS analysis, a total of 43 SNP loci significantly associated with CP and SP were identified across 21 chromosomes. Comparison with previously published studies revealed that several of the significant loci identified in this study overlapped with or were in close proximity to QTL regions reported in earlier research. Among them, 3A_15059920 was located within the q3A-1 interval (10.30–15.69 Mb), a region previously reported by Yang et al. [29] to be significantly associated with five quality traits, including grain protein content, wet gluten content, SDS sedimentation volume, dough development time, and grain test weight. In addition, 5B_435082730 was only 0.23 Mb away from QTL_GPC-5B.2, a GPC-related significant locus identified by Tian et al. [14] through GWAS. Furthermore, 3D_308390893 was located 2.50 Mb from AX-94,412,995, a GPC-related significant locus identified by Bashir et al. [30] through GWAS, and 6D_141184719 was located 1.08 Mb from QGPC-6D.1 reported by Liu et al. [31]. These results indicate that grain protein content is regulated by multiple loci [32]. The above results provide important evidence for further dissecting the genetic basis of grain protein content. To further delineate key genetic regions and screen candidate genes, significant loci located within LD blocks with high linkage intensity were selected for subsequent analysis. These included the crude protein-related loci 2B_161726146 and 2D_610941094, as well as the soluble protein-related loci 1B_562451687 and 7A_4591015.

Transcriptome analysis of developing grains from contrasting phenotypic materials (HM14 and XC48) revealed two peaks in the number of differentially expressed genes at 10 and 30 days after anthesis. This finding is consistent with the study by Shewry et al. [33], which tracked protein deposition in wheat grains using SDS-PAGE and reported that gluten proteins were first detected at 10 days after anthesis, accumulated most rapidly between 12 and 35 days after anthesis, and showed little change after 42 days. The observed expression peaks of DEGs in this study are highly consistent with the critical stages of grain protein synthesis and accumulation. By integrating GWAS and transcriptome data, this study identified five core candidate genes significantly associated with grain protein content: TaPAE2-like, TaHIP1, TaGlu-B1al-like, Ta1-SST, and Td1-SST-like. Among them, TaPAE2-like encodes pectin acetylesterase 2. GO annotation indicates that this gene is involved in cell wall organization, is localized to the extracellular region and the plant-type cell wall, and possesses hydrolase activity and pectin acetylesterase activity. Pectin acetylesterase regulates the degree of pectin acetylation by hydrolyzing acetyl ester bonds on galacturonic acid residues in pectin, thereby influencing cell wall structure and function [34]. The expression level of this gene in the strong-gluten wheat variety XC48 was significantly different at both 10 DAA and 30 DAA, suggesting its potential involvement in grain development or protein accumulation processes. TaHIP1 possesses ubiquitin protein ligase activity (MF: ubiquitin protein ligase activity) and metal ion binding function (MF: metal ion binding). E3 ubiquitin ligases mediate the ubiquitination of target proteins, thereby regulating protein degradation and turnover. Studies have shown that RING-type E3 ligases play important roles in regulating starch synthesis and grain development in wheat [35]. In addition, E3 ligase genes have also been reported to be involved in wheat disease resistance responses [36]. Elhadi et al. [37] identified a candidate gene encoding a RING-type E3 ubiquitin-protein ligase on chromosome 5D, which was shown to be involved in regulating grain size in wheat. TaGlu-B1al-like encodes a high-molecular-weight glutenin subunit DX5-like and possesses nutrient reservoir activity (MF: nutrient reservoir activity). The high-molecular-weight glutenin subunits (HMW-GS) encoded by this gene are key determinants of wheat gluten strength and bread-making quality [33]. López-Fernández et al. [38] identified an MTA-QTL_1B.7 region on chromosome 1B containing this gene through GWAS, and this region was significantly associated with gluten aggregation properties and dough strength. Ta1-SST and Td1-SST-like are both annotated as sucrose: sucrose 1-fructosyltransferase (1-SST) and are involved in carbohydrate metabolic processes. 1-SST is a key enzyme in the fructan biosynthesis pathway, catalyzing the irreversible conversion of two sucrose molecules into 1-kestose and glucose [39]. Studies have shown that the expression of 1-SST genes increases during overwintering in winter wheat and is closely associated with fructan accumulation [40]. In wheat varieties with stronger cold tolerance, 1-SST expression levels are higher, accompanied by greater fructan accumulation [40, 41]. The five candidate genes identified in this study provide key entry points for understanding the molecular basis of grain protein accumulation in wheat and serve as valuable genetic resources for improving protein quality through molecular marker-assisted breeding.

Conclusions

In this study, a total of 43 genetic loci significantly associated with CP and SP were identified through GWAS. Based on the LD decay distance (590 kb), LD block analysis was performed using regions 590 kb upstream and downstream of each significant locus as candidate intervals. From this, four significant loci with high linkage intensity located within LD blocks were selected: two CP-related loci, 2B-161,726,146 and 2D-610,941,094, and two SP-related loci, 1B-562,451,687 and 7A-4,591,015. On this basis, transcriptome data from key grain developmental stages were integrated with gene functional annotations and DEGs, ultimately leading to the identification of five candidate genes associated with grain protein content. Among them, TaPAE2-like and TaHIP1 were identified as candidate genes related to CP, while TaGlu-B1al-like, Ta1-SST, and Td1-SST-like were identified as candidate genes related to SP. These candidate genes provide an important foundation for the genetic improvement of grain protein content in wheat.

Methods

Experimental material

A total of 260 wheat germplasm accessions were used in this study, including 213 from China, 41 from CIMMYT, and 6 from Sweden (Table S9). The experimental materials were planted under different ecological conditions in Qinghai Province over two consecutive years (2023–2024). In 2023, the materials were grown at the Germplasm Resources Innovation Experimental Base of the Academy of Agriculture and Forestry Sciences, Qinghai University, in Xining City (36°43′ N, 101°45′ E, 2261 m). In 2024, they were planted at three locations: the same base in Xining, the Triticeae Crop Experimental Base in Guide County, Hainan Tibetan Autonomous Prefecture (35°55′ N, 101°47′ E, 2200 m), and the Xiangride Town Experimental Base in Haixi Mongolian and Tibetan Autonomous Prefecture (36°03′ N, 97°47′ E, 2999 m). Sowing dates were determined according to the specific conditions of each location. The planting followed a sequential arrangement design, with each accession manually sown in three rows of 20 seeds per row. The row length was 1.5 m, and the row spacing was 20 cm. The same cultivation and management practices were applied across all four environments. Guard rows were established around the experimental plots, and field management was consistent with standard practices.

The tested materials grew normally and reached maturity at all four experimental sites. At maturity, two rows of each accession were harvested for threshing. The harvested seeds were stored at room temperature for two months, after which the wheat kernels were cleaned and used for further analysis. CP was determined using a Perten DA7200 multifunctional near-infrared analyzer, with three replicate measurements performed for each accession in each environment. SP was measured using the BCA Protein Assay Kit (Suzhou Comin Biotechnology Co., Ltd.). For each wheat variety, 200 g of kernels were ground, and the powder was passed through an 80-mesh sieve. Then, 0.1 g of the dry sample powder was weighed, and subsequent steps were carried out strictly according to the kit instructions. Three replicate measurements were performed for each accession from each environment.

BLUP values and H² were calculated using the “lme4” package in R software. Data analysis was performed using Microsoft Excel 2016. Variance analysis was conducted using SPSS, normal distribution plots were generated using Origin, and correlation analysis among traits was also performed.

GWAS analysis

The genotypic data used in this study were derived from an unpublished high-density wheat SNP array (100 K) obtained from our research group in a previous study. Raw data were quality-controlled using PLINK software to obtain a high-quality core SNP dataset. The quality control criteria were as follows: minor allele frequency (MAF) ≥ 0.05 and SNP call rate ≥ 95% [23]. The SNP array consisted of 182,549 SNPs distributed across the 21 wheat chromosomes. LD decay was calculated using PopLDdecay [42] to assess the level of LD in the 260 wheat accessions.

GWAS was conducted using the mrMLM package in R software (https://github.com/YuanmingZhang65/mrMLM), employing five multi-locus models included in the package: FASTmrEMMA, ISIS EM-BLASSO, mrMLM, FASTmrMLM, and pLARmEB. The significance threshold for association results was set at a logarithm of odds (LOD) score ≥ 5, in accordance with standard screening criteria [19]. Subsequently, LDBlockShow [43] was used to identify LD regions containing significant SNPs, which were considered potential candidate gene regions for further analysis.

RNA-seq analysis

From the 260 accessions, two approved cultivars—the strong‑gluten variety XC48 and the medium‑gluten variety HM14—were selected based on their significant differences in crude protein and soluble protein content across four environments. The two cultivars were planted in 2025 at the Germplasm Innovation Experimental Base of Qinghai University, Xining, Qinghai Province. Grain samples were collected at 10, 15, 20, 25, and 30 DAA for gene expression analysis. Six biological replicates were collected per sample: three were used for grain protein content determination, and the other three were used for grain transcriptome sequencing. After sampling, the grains were immediately frozen in liquid nitrogen and stored at − 80°C.

Total RNA was extracted from the grain samples using TRIzol® Reagent (Shanghai Majorbio Bio‑pharm Technology Co., Ltd., Shanghai, China) according to the manufacturer’s instructions. RNA quality was assessed using a 5300 Bioanalyzer (Agilent), and RNA concentration was quantified using an ND-2000 spectrophotometer (NanoDrop Technologies). Only high‑quality RNA samples (OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, RQN ≥ 6.5, 28S:18S ≥ 1.0, > 1 µg) were used for library construction. Sequencing libraries were prepared using 1 µg of total RNA with the Illumina® Stranded mRNA Prep Ligation kit following the manufacturer’s instructions (Illumina, San Diego, CA). The libraries were sequenced on the NovaSeq X Plus/DNBSEQ T7 platform with 2 × 150 bp paired-end reads (Shanghai Majorbio Bio‑pharm Technology Co., Ltd.).

Raw paired-end reads were trimmed and quality‑controlled using fastp [44] with default parameters. A total of 186.43 Gb of clean data was generated, with each of the 30 samples yielding over 5.62 Gb of clean data and Q30 base percentages exceeding 95.45%. The raw data have been deposited in the National Center for Biotechnology Information (NCBI) under BioProject accession number PRJNA1402758. Clean reads were aligned to the reference genome (Chinese Spring v2.1) using HISAT2 [45] with orientation mode, achieving alignment rates ranging from 74.81% to 93.55%. The mapped reads for each sample were assembled using StringTie [46] with a reference‑based approach. Gene expression levels were normalized to fragments per kilobase per million reads (FPKM). A total of 156,052 expressed genes and 181,418 expressed transcripts were detected across all samples.

Data analysis was performed using the Majorbio Cloud platform (www.majorbio.com). To identify DEGs, transcript expression levels were calculated using the FPKM method, and gene abundances were quantified using RSEM [47]. Differential expression analysis was conducted using DESeq2 [48] with thresholds of |log2FC| ≥ 1 and padjust < 0.01. Functional enrichment analyses, including GO and KEGG pathway enrichment, were performed to identify DEGs significantly enriched in GO terms and metabolic pathways using Goatools and KOBAS [49], respectively, with a Bonferroni‑corrected P‑value ≤ 0.05 as the significance threshold.

qRT-PCR analysis

cDNA was synthesized using the TaKaRa PrimeScript RT Master Mix Kit, and all reaction steps were carried out strictly in accordance with the manufacturer’s instructions. Primers were designed using NCBI (Table S10). qRT-PCR was performed using the TaKaRa TB Green Premix Ex Taq II Kit on a LightCycler 480 system (Roche), with Tubulin used as the reference gene. Three technical replicates were performed for each sample, and data were analyzed using the 2⁻ΔΔCt method [50]. Gene expression data were organized using Microsoft Excel 2010 and visualized using GraphPad Prism 10. Results are presented as mean ± SD.

Abbreviations

BLUP

Best Linear Unbiased Prediction

CP

Crude Protein

DAA

Days After Anthesis

DEGs

Differentially Expressed Genes

E1

Planted in 2023 at the Germplasm Resource Innovation Experimental Base, College of Agriculture and Forestry, Qinghai University, Xining City, Qinghai Province, China

E2

Planted in 2024 at the Same Base in Xining City, Qinghai Province, China

E3

Planted in 2024 at the Cereal Crops Experimental Base in Guide County, Hainan Tibetan Autonomous Prefecture, Qinghai Province, China

E4

Planted in 2024 at the Xiangride Town Experimental Base in Haixi Mongol and Tibetan Autonomous Prefecture, Qinghai Province, China

FPKM

Fragments Per Kilobase Per Million Mapped Reads

GO

Gene Ontology

GWAS

Genome-wide association study

H²

Broad-Sense Heritability

HM14

The Wheat Cultivar Humai 14

KEGG

Kyoto Encyclopedia of Genes and Genomes

LD

Linkage Disequilibrium

LOD

Logarithm of Odds

MAF

Minor Allele Frequency

NCBI

National Center for Biotechnology Information

qRT-PCR

Quantitative real-time reverse transcription PCR

QTLs

Quantitative Trait Loci

QTNs

Quantitative Trait Nucleotides

RNA-seq

RNA sequencing

SD

Standard Deviation

SNP

Single Nucleotide Polymorphism

SP

Soluble Protein

XC48

The Wheat Cultivar Xinchun 48

Authors’ contributions

J.S. contributed to software processing, data analysis, writing, and editing. H.D. contributed to software processing, data analysis, and editing. Y.W. contributed to investigation, manuscript review, and editing. X.Z. contributed to software processing, project supervision, data analysis, and writing. X.Y. contributed to software processing, investigation, and writing. Y.Y. contributed to project supervision, funding acquisition, writing, review, and editing. All authors read and approved the final manuscript.

Funding

This study was supported by the Qinghai Province “Kunlun Talents · High-Level Innovative and Entrepreneurial Talent” Leading Talent Training Program(QHKLYC-GDCXCY-2025-079).

Data availability

All experimental materials were provided by the research group of the Crop Breeding and Cultivation Institute at the Qinghai Academy of Agricultural and Forestry Sciences. All data generated in this study are included in the supplementary tables and figures. Raw sequencing data have been deposited in the NCBI under BioProject accession number PRJNA1402758.

Declarations

Ethics approval and consent to participate

All procedures were conducted following the guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.KaplanEvlice A, Cetiner B, Pehlivan A, Kara R. Wheat Quality. In: Advances in Wheat Breeding: Towards Climate Resilience and Nutrient Security. Edited by Zencirci N, Altay F, Baloch FS, Nadeem MA, Ludidi N. Singapore: Springer Nature Singapore;2024. p. 453–477. 10.1007/978-981-99-9478-6_9.
  • 2.MetheP, Kawade S, Oak M. Exploring the texture and quality delight: cookies enhanced with coarse cereals–wheat composite flours. Cereal Res Commun. 2025;53(4):2545–56. 10.1007/s42976-025-00684-x. [Google Scholar]
  • 3.SainiP, Sinha ASK, Prasad K. Wheat bran modification treatments for enhancing the quality and nutritional profile of traditional cookies. Discover Food. 2025;5(1):424. 10.1007/s44187-025-00681-3. [Google Scholar]
  • 4.JungJG, Kim J-H, Kim C, Shim D. Assessment of crude protein in wheat kernels using SWIR hyperspectral imaging combined with deep learning-based segmentation. J Agric Food Res. 2026;26:102679. 10.1016/j.jafr.2026.102679. [Google Scholar]
  • 5.ShewryP. Wheat grain proteins: Past, present, and future. Cereal Chem. 2023;100(1):9–22. 10.1002/cche.10585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.ZayasJF. Solubility of Proteins. In: Functionality of Proteins in Food. Edited by Zayas JF. Berlin, Heidelberg: Springer Berlin Heidelberg; 1997: 6–75. 10.1007/978-3-642-59116-7.
  • 7.LiuC, Zhao J, Gupta S, Carrillo M. Quantification of Crude and Soluble Protein Content. In: Plant-Based Proteins: Production, Physicochemical, Functional, and Sensory Properties. Edited by Li Y. New York, NY: Springer US; 2025. p. 107–122. 10.1007/978-1-0716-4272-6_9.
  • 8.DangHH, Bartel L, England KA, Cavagnero S. Exploring kinetically controlled protein solubility under physiologically relevant conditions. Biophysical Journal. 2022;121(3, Supplement 1):184a. 10.1016/j.bpj.2021.11.1801.
  • 9.GrossmannL, McClements DJ. Current insights into protein solubility: A review of its importance for alternative proteins. Food Hydrocolloids. 2023;137:108416. 10.1016/j.foodhyd.2022.108416. [Google Scholar]
  • 10.BoyeC, Nirmalan S, Ranjbaran A, Luca F. Genotype × environment interactions in gene regulation and complex traits. Nat Genet. 2024;56(6):1057–68. 10.1038/s41588-024-01776-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.ChirmadeS, Wang Z, Mastromatteo S, Sanders E, Thiruvahindrapuram B, Nalpathamkalam T, Pellecchia G, Lin F, Keenan K, Patel RV et al. GWAS SVatalog: a visualization tool to aid fine-mapping of GWAS loci with structural variations. Heredity. 2026;135(3):199-210. 10.1038/s41437-025-00809-2. [DOI] [PMC free article] [PubMed]
  • 12.JoshiB, Singh S, Asif Iquebal M, Sarika, Kumar D, Sawant SV, Jena SN. Tools and Techniques Used for Quantitative Trait Loci (QTL) and Association Mapping Studies. In: GWAS and QTL Mapping in Horticultural Crops: Vegetable Crops. Edited by Dey SS, Asif Iquebal M, Sarika, Bhatia Dey R. Singapore: Springer Nature Singapore; 2026: 61–113. 10.1007/978-981-95-2306-1_2.
  • 13.HeJ, Gai J. Genome-Wide Association Studies (GWAS). In: Plant Genotyping: Methods and Protocols. Edited by Shavrukov Y. New York, NY: Springer US; 2023. p. 123–146. 10.1007/978-1-0716-3024-2_9. [DOI] [PubMed]
  • 14.TianY, Liu P, Kong D, Nie Y, Xu H, Han X, Sang W, Li W. Genome-wide association analysis and KASP markers development for protein quality traits in winter wheat. BMC Plant Biol. 2025;25(1):149. 10.1186/s12870-025-06171-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.GuoY, Wang G, Guo X, Chi S, Yu H, Jin K, Huang H, Wang D, Wu C, Tian J et al. Genetic dissection of protein and starch during wheat grain development using QTL mapping and GWAS. Front Plant Sci.2023;14:1189887. 10.3389/fpls.2023.1189887. [DOI] [PMC free article] [PubMed]
  • 16.UauyC, Brevis JC, Dubcovsky J. The high grain protein content gene Gpc-B1 accelerates senescence and has pleiotropic effects on protein content in wheat. J Exp Bot. 2006;57(11):2785–2794. 10.1093/jxb/erl047. [DOI] [PubMed] [Google Scholar]
  • 17.MuqaddasiQH, Brassac J, Ebmeyer E, Kollers S, Korzun V, Argillier O, Stiewe G, Plieske J, Ganal MW, Röder MS. Prospects of GWAS and predictive breeding for European winter wheat’s grain protein content, grain starch content, and grain hardness. Sci Rep. 2020;10(1):12541. 10.1038/s41598-020-69381-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.ErmolaevA, Bespalova L, Korobkova V, Yanovsky A, Nazarova L, Kroupina A, Chernook A, Mudrova A, Voronezhskaya V, Kroupin P et al. High-quality bonds: serine acetyltransferase 2 gene revealed by GWAS is associated with grain protein content in spring durum wheat. Front Plant Sci. 2025;16:1632673. 10.3389/fpls.2025.1632673. [DOI] [PMC free article] [PubMed]
  • 19.ZhangYW, Tamba CL, Wen YJ, Li P, Ren WL, Ni YL, Gao J, Zhang YM. mrMLM v4.0.2: An R Platform for Multi-locus Genome-wide Association Studies. Genomics Proteom Bioinf. 2020;18(4):481–487. 10.1016/j.gpb.2020.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.StaffTPG. Correction: Iterative Usage of Fixed and Random Effect Models for Powerful and Efficient Genome-Wide Association Studies. PLoS Genet. 2016;12(3):e1005957. 10.1371/journal.pgen.1005957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.WangS-B, Feng J-Y, Ren W-L, Huang B, Zhou L, Wen Y-J, Zhang J, Dunwell JM, Xu S, Zhang Y-M. Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Sci Rep. 2016;6(1):19444. 10.1038/srep19444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.HeF, Xu M, Liu H, Xu Y, Long R, Kang J, Yang Q, Chen L. Unveiling alfalfa root rot resistance genes through an integrative GWAS and transcriptome study. BMC Plant Biol. 2025;25(1):58. 10.1186/s12870-024-05903-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.LuanH, Gao J, Wu Y, Yang J, Shen Y, Sun M, Liu F, Xu M, Xu X, Sun M, et al. Identification of candidate genes for grain size in barley through combined GWAS and transcriptome analysis. Journal of Plant Research2025;138(5):789-806. 10.1007/s10265-025-01657-1. [DOI] [PubMed]
  • 24.KhalidA, Hameed A, Tahir MF. Wheat quality: A review on chemical composition, nutritional attributes, grain anatomy, types, classification, and function of seed storage proteins in bread making quality. Front Nutr. 2023; 10:1053196 . 10.3389/fnut.2023.1053196. [DOI] [PMC free article] [PubMed]
  • 25.RayD, Chatterjee N. Effect of non-normality and low count variants on cross-phenotype association tests in GWAS. Eur J Hum Genet. 2020;28(3):300– 312. 10.1038/s41431-019-0514-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.XuX, Zhou L, Taylor J, Casa R, Fan C, Song X, Yang G, Huang W, Li Z. The 500-meter long-term winter wheat grain protein content dataset for China from multi-source data. Sci Data. 2024;11(1):1025. 10.1038/s41597-024-03866-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.PaullAE, Anderson JA, THE EFFECTS OF AMOUNT AND DISTRIBUTION OF RAINFALL ON THE PROTEIN CONTENT OF WESTERN CANADIAN WHEAT. Can J Res. 1942;20c(4):212–227. 10.1139/cjr42c-021. [Google Scholar]
  • 28.Machadoe Silva C, Mezzomo HC, Alves RS, de Resende MDV, Nardino M. Optimizing selection of wheat genotypes through simulated individual BLUP and modified simulated individual BLUP. Agron J. 2023;115(3):1237-1247. 10.1002/agj2.21289.
  • 29.YangY, Chai Y, Zhang X, Lu S, Zhao Z, Wei D, Chen L, Hu Y-G. Multi-Locus GWAS of Quality Traits in Bread Wheat: Mining More Candidate Genes and Possible Regulatory Network. Front Plant Sci. 2020; 11:1091. 10.3389/fpls.2020.01091. [DOI] [PMC free article] [PubMed]
  • 30.BashirL, Budhlakoti N, Pradhan AK, Sharma D, Jain A, Rehman SS, Kondal V, Jacob SR, Bhardwaj R, Gaikwad K, et al. Identification of quantitative trait nucleotides for grain quality in bread wheat under heat stress. Sci Rep. 2025;15(1):6641. 10.1038/s41598-025-91199-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.LiuZ, Lai X, Chen Y, Zhao P, Wang X, Ji W, Xu S. Selection and application of four QTLs for grain protein content in modern wheat cultivars. J Integr Agric. 2024;23(8):2557–2570. 10.1016/j.jia.2023.09.006. [Google Scholar]
  • 32.HongY, Zhang M, Zhu J, Zhang Y, Lv C, Guo B, Wang F, Xu R. Genome-wide association studies reveal novel loci for grain size in two-rowed barley (Hordeum vulgare L). Theor Appl Genet. 2024;137(3):58. 10.1007/s00122-024-04562-8. [DOI] [PubMed] [Google Scholar]
  • 33.ShewryPR, Underwood C, Wan Y, Lovegrove A, Bhandari D, Toole G, Mills ENC, Denyer K, Mitchell RAC. Storage product synthesis and accumulation in developing grains of wheat. J Cereal Sci. 2009;50(1):106–112. 10.1016/j.jcs.2009.03.009. [Google Scholar]
  • 34.PhilippeF, Pelloux J, Rayon C. Plant pectin acetylesterase structure and function: new insights from bioinformatic analysis. BMC Genomics. 2017;18(1):456. 10.1186/s12864-017-3833-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.ParveenA, Rahim MS, Sharma A, Mishra A, Kumar P, Fandade V, Kumar P, Bhandawat A, Verma SK, Roy J. Genome-wide analysis of RING-type E3 ligase family identifies potential candidates regulating high amylose starch biosynthesis in wheat (Triticum aestivum L). Sci Rep. 2021;11(1):11461. 10.1038/s41598-021-90685-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.YangF, Liu J, Guo Y, He Z, Rasheed A, Wu L, Cao S, Nan H, Xia X. Genome-Wide Association Mapping of Adult-Plant Resistance to Stripe Rust in Common Wheat (Triticum aestivum). Plant Dis. 2020;104(8):2174–2180. 10.1094/PDIS-10-19-2116-RE. [DOI] [PubMed] [Google Scholar]
  • 37.ElhadiGM, Kamal NM, Gorafi YS, Yamasaki Y, Takata K, Tahir ISA, Itam MO, Tanaka H, Tsujimoto H. Exploitation of Tolerance of Wheat Kernel Weight and Shape-Related Traits from Aegilops tauschii under Heat and Combined Heat-Drought Stresses. In: International Journal of Molecular Sciences.2021;.22(4), 1830 . 10.3390/ijms22041830. [DOI] [PMC free article] [PubMed]
  • 38.López-FernándezM, Chozas A, Benavente E, Alonso-Rueda E, Isidro y Sánchez J, Pascual L, Giraldo P. Genome wide association mapping of end-use gluten properties in bread wheat landraces (Triticum aestivum L). J Cereal Sci. 2024;118:103956. 10.1016/j.jcs.2024.103956. [Google Scholar]
  • 39.SchroevenL, Lammens W, Van Laere A, Van den Ende W. Transforming wheat vacuolar invertase into a high affinity sucrose:sucrose 1-fructosyltransferase. New Phytol. 2008;180(4):822–831. 10.1111/j.1469-8137.2008.02603.x. [DOI] [PubMed] [Google Scholar]
  • 40.KawakamiA, Yoshida M. Molecular Characterization of Sucrose:Sucrose 1-Fructosyltransferase and Sucrose:Fructan 6-Fructosyltransferase Associated with Fructan Accumulation in Winter Wheat during Cold Hardening. Biosci Biotechnol Biochem. 2002;66(11):2297–2305. 10.1271/bbb.66.2297. [DOI] [PubMed] [Google Scholar]
  • 41.HouJ, Huang X, Sun W, Du C, Wang C, Xie Y, Ma Y, Ma D. Accumulation of water-soluble carbohydrates and gene expression in wheat stems correlates with drought resistance. J Plant Physiol. 2018;231:182–191. 10.1016/j.jplph.2018.09.017. [DOI] [PubMed] [Google Scholar]
  • 42.ZhangC, Dong S-S, Xu J-Y, He W-M, Yang T-L. PopLDdecay: a fast and effective tool for linkage disequilibrium decay analysis based on variant call format files. Bioinformatics. 2019;35(10):1786–1788. 10.1093/bioinformatics/bty875. [DOI] [PubMed] [Google Scholar]
  • 43.DongS-S, He W-M, Ji J-J, Zhang C, Guo Y, Yang T-L. LDBlockShow: a fast and convenient tool for visualizing linkage disequilibrium and haplotype blocks based on variant call format files. Brief Bioinform. 2021;22(4):bbaa227. 10.1093/bib/bbaa227. [DOI] [PubMed] [Google Scholar]
  • 44.ChenS, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–890. 10.1093/bioinformatics/bty560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.KimD, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods. 2015;12(4):357–360. 10.1038/nmeth.3317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Pertea M, Pertea GM, Antonescu CM, Chang T-C, Mendell JT, Salzberg SL. StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol. 2015;33(3):290–295. 10.1038/nbt.3122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.LiB, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12(1):323. 10.1186/1471-2105-12-323 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.LoveMI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.XieC, Mao X, Huang J, Ding Y, Wu J, Dong S, Kong L, Gao G, Li C-Y, Wei L. KOBAS 2.0: a web server for annotation and identification of enriched pathways and diseases. Nucleic Acids Res. 2011;39(suppl2):W316–322. 10.1093/nar/gkr483. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.PfafflMW. A new mathematical model for relative quantification in real-time RT–PCR. Nucleic Acids Res. 2001;29(9):e45–45. 10.1093/nar/29.9.e45. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All experimental materials were provided by the research group of the Crop Breeding and Cultivation Institute at the Qinghai Academy of Agricultural and Forestry Sciences. All data generated in this study are included in the supplementary tables and figures. Raw sequencing data have been deposited in the NCBI under BioProject accession number PRJNA1402758.

Articles from BMC Plant Biology are provided here courtesy of BMC

RESOURCES