Integrating RTM-GWAS and meta‑QTLs uncovers the genomic regions and candidate genes for the first fruit branch node and its height in upland cotton

doi:10.21203/rs.3.rs-4306351/v1

Download PDF

Research Article

Integrating RTM-GWAS and meta‑QTLs uncovers the genomic regions and candidate genes for the first fruit branch node and its height in upland cotton

https://doi.org/10.21203/rs.3.rs-4306351/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 22 Aug, 2024

Read the published version in Theoretical and Applied Genetics →

You are reading this latest preprint version

The first fruit branch node (FFBN) and height of the first fruit branch node (HFFBN) are two important traits related to plant architecture and early maturation in upland cotton. Several studies have been performed to elucidate the genetic basis of FFBN and HFFBN in cotton using biparental and natural populations. In this study, by using 9,244 SNP linkage disequilibrium block (SNPLDB) loci from 315 upland cotton accessions, we carried out restricted two‐stage multilocus and multiallele genome-wide association studies (RTM‐GWASs) and identified promising haplotypes/alleles of the four stable and true major SNPLDB loci that were significantly associated with FFBN and HFFBN. Additionally, a meta-QTL analysis was conducted on 274 original QTLs reported in 27 studies, and 40 MQTLs for FFBN and HFFBN were detected. Through the integration of RTM-GWAS and meta‑QTL analyses, two stable and true major SNPLDBs (LDB_5_15144433 and LDB_16_37952328) that were distributed in the two MQTLs were identified. Ultimately, 142 genes were annotated in the two genomic regions, and three candidate genes for FFBN and HFFBN were identified in the genomic region (A05:14.64–15.64 Mb) by RNA-Seq and qRT‒PCR. The results of virus-induced gene silencing (VIGS) experiments indicated that GhE6was a key gene related to HFFBN and that GhDRM1 and GhGES were important genes associated with early flowering in upland cotton. These findings will aid in the future identification of molecular markers and genetic resources for developing elite early-maturing cultivars with ideal plant characteristics.

upland cotton

FFBN

HFFBN

RTM-GWAS

meta‑QTL

candidate genes

Two genomic regions and a candidate gene for FFBN and HFFBN were identified through the integration of RTM-GWAS and meta‑QTL analyses.

Cotton (Gossypium spp.) is a vitally important industrial crop worldwide because of its natural fibers and the high levels of edible oil and protein in its seeds (Yang et al. 2023). Upland cotton (Gossypium hirsutum L.) comprises 95% of the global cotton yield and is a flagship species for cotton breeding (Chen et al. 2007). The first fruit branch node (FFBN) and height of the first fruit branch node (HFFBN) are two important agronomic traits of upland cotton, and they are closely related to early maturation, plant architecture and mechanical harvest (Su et al. 2016, 2018).

In general, early-maturing cotton cultivars have several obvious characteristics, including relatively lower FFBN and HFFBN, lower plant height (PH), and moderately compact plant type. In recent years, plant architecture has played an increasingly important role in mechanized harvesting processes in Chinese cotton production (Mao et al. 2014, 2015). For cotton bolls to be picked from the first fruit branch, excellent cultivars that are adaptable to mechanical harvesting should have the basic features of relatively large FFBN and HFFBN values. In conclusion, FFBN and HFFBN are two very important traits for early maturity, plant architecture and mechanical harvesting in upland cotton.

Cotton FFBN and HFFBN are complicated traits controlled by numerous quantitative trait loci (QTLs). Over the last 20 years, numerous QTLs for FFBN and HFFBN have been mapped by linkage analysis of segregating populations in cotton (Fan et al. 2006; Guo et al. 2008, 2009; Li et al. 2012, 2017; Jia et al. 2016, 2023; Zhang et al. 2021). Moreover, simple sequence repeats (SSRs) related to FFBN and HFFBN have been mapped through association analysis based on linkage disequilibrium (Li et al. 2016). Additionally, many single nucleotide polymorphisms (SNPs) related to these two traits have been identified via genome-wide association studies (GWASs) (Su et al. 2016, 2018; Li et al. 2018; Fu et al. 2019; Shen et al. 2019; Li et al. 2021). Although a few QTLs and association loci for FFBN and HFFBN have been mapped in previous studies, elucidation of the genome segments and candidate genes controlling these two traits remains limited.

Common GWAS procedures, including the general linear model (GLM) and mixed linear model (MLM), are single-locus GWAS methods, and they often are primarily used to identify major QTLs, each with only two alleles, in plants (He et al. 2017; Liu et al. 2021a). Compared with single-locus GWAS models, the restricted two-stage multilocus and multiallele GWAS (RTM-GWAS) method has the obvious advantage of enabling a larger number of QTLs to be identified, and this method can be deemed effective for detecting QTL allele constitutions (He et al. 2017). To date, the RTM-GWAS method has been widely used to explore QTLs for target traits in soybean and cotton. For instance, a greater number of QTL alleles for the flowering period, number of main stem nodes, seed isoflavone content, and shade and drought tolerance were identified via the RTM-GWAS method in soybean (Meng et al. 2016; Khan et al. 2018; Fu et al. 2020; Liu et al. 2021a, b; Su et al. 2023). Additionally, we revealed QTL alleles associated with flowering time (FT), the flower and boll period (FBP), the whole growth period (WGP), PH, fruit branch length, fruit branch angle, fiber length, fiber strength, fiber micronaire, fiber uniformity, and fiber elongation in upland cotton (Su et al. 2020a, b; Wang et al. 2022, 2023). However, to date, studies identifying QTL alleles for cotton FFBN and HFFBN via the RTM-GWAS procedure have not been reported.

Meta-analysis is a highly efficient approach for integrating QTL mapping results and narrowing down genomic regions through the compiling of rich QTL data from various studies (Welcker et al. 2011). This analytical method is beneficial for identifying consistent and reliable meta-quantitative trait loci (MQTLs), which are not affected by genetic background or growth environment (Arcade et al. 2004; Sosnowski et al. 2012). Hence, MQTL analysis has been widely used to obtain good integration results for the QTLs of many crop traits, such as traits related to yield and drought tolerance in rice (Khowaja et al. 2009; Khahani et al. 2020), yield and insect resistance in maize (Wang et al. 2013, 2020b; Badji et al. 2018), general crop yield components (Yang et al. 2021), and resistance to drought, stem rust, leaf rust, Fusarium head blight and tan spot in wheat (Kumar et al. 2020; Liu et al. 2020a, b; Yu et al. 2014; Soriano and Royo 2015; Tai et al. 2021; Venske et al. 2019; Cai et al. 2019; Zheng et al. 2020). Similarly, MQTL analysis of some traits in cotton has also been performed. For example, QTL clusters and hotspots related to morphology, yield, fiber quality and drought tolerance were identified in cotton (Said et al. 2013, 2015). Moreover, MQTLs for resistance to cotton disease, including Verticillium wilt, Fusarium wilt and root-knot nematodes, have been reported in previous studies (Said et al. 2013, 2015; Zhang et al. 2015; Huo et al. 2023). Although six and one QTLs for FFBN and HFFBN, respectively, have been identified through QTL studies in cotton (Said et al. 2013), no MQTLs for the two target traits have been identified.

Here, we performed RTM-GWAS analyses of FFBN and HFFBN in 315 upland cotton lines and conducted a meta-analysis of cotton FFBN- and HFFBN-QTLs published in the last 20 years. The reliable major QTLs for the two traits were mapped by integrating the results of the RTM-GWAS and MQTL analyses. Then, key candidate genes that affect the two traits were identified in key genomic regions by transcriptome data analysis and virus-induced gene silencing (VIGS) technology. Our aim was to decipher the genomic regions and candidate genes for cotton FFBN and HFFBN. This work will aid the development of early-maturing cotton cultivars with ideal plant characteristics in future cotton breeding programs.

Cotton accessions and field experimental design

A natural population consisting of 315 upland cotton lines was used, and the relevant field trials were carried out as described previously (Su et al. 2020a, b). The natural population included 290 accessions from China and 25 lines from foreign regions (FRs, 21 accessions from the USA and four lines from the former Soviet Union) (Table S1). According to the geographic and cotton-growing areas of China, the Chinese accessions were grouped into four geographic regions: the Northwest Inland Region (NIR, 96 accessions), the Yellow River Region (YRR, 123 accessions), the Yangtze River Region (YZRR, 51 accessions), and the Northern Specific Early-maturity Region (NSER, 20 accessions). Field experiments were implemented with a total of 315 upland cotton lines in three regional environments, which were designated E₁ (Anyang, Henan, China, in 2014), E₂ (Anyang, Henan, China, in 2015) and E₃ (Shihezi, Xinjiang, China, in 2014). The 315 lines were tested in single-row test plots of 4×0.8 m (length×width) with 0.2 m spacing between plants in the E₁ and E₂ experiments and were planted in double-row test plots of 2×0.76 m (length×width) with 0.1 m space between plants in the E₃experiment. The accessions were subjected to a randomized block design with three replicates in each environment. The sowing dates were, respectively, 19 April, 19 April, and 25 April in the tested environments E₁, E₂ and E₃.

Investigation of the FFBN and HFFBN phenotypic characteristics

The FFBN and HFFBN of the 315 upland cotton accessions were investigated using methods described previously (Su et al. 2016, 2018). The FFBN was the number of nodes on the first fruit branch, and the HFFBN was the distance between the cotyledon node and the FFBN. The broad-sense heritability (h²) for FFBN and HFFBN was calculated using the formula h²= σg²/(σg² + σge²/n + σɛ²/nr), as described previously (Hanson et al. 1956). Analysis of variance (ANOVA) for the FFBN and HFFBN were calculated using SPSS 26.0 software.

RTM-GWAS procedures

The RTM-GWAS analyses (He et al. 2017) were performed using 9,244 organized SNP linkage disequilibrium blocks (SNPLDBs) (Su et al. 2020a, b) and the phenotypic values of the FFBN and HFFBN. The procedure was used to detect the QTLs for FFBN and HFFBN by using the total phenotypic values and individual environment phenotypic values, respectively. The top ten eigenvectors of the genetic similarity coefficient matrix were constructed by utilizing the SNPLDBs, and they were integrated as covariates for group structure correction. In the first phase, a preselection of SNPLDBs was conducted using a single-locus linear model at the 5% significance level. In the second stage, SNPLDB loci associated with FFBN and HFFBN were screened out by a stepwise regression method under a multilocus model at the 5% significance level. Finally, the values of the allele effects were computed for any significant associations with FFBN and HFFBN, and they were established as a QTL-allele matrix.

Haplotype analysis and frequency distribution of the major SNPLDB loci

The significant SNPLDBs that were simultaneously identified by the total phenotypic values and by one or more individual–environment phenotypic values with large values (− lgP ≥ 10.00) were regarded as stable major loci. For the stable major SNPLDB loci, the phenotypic effect of each haplotype/allele was determined, and box plots were drawn using OriginPro 2018C software. Then, the stable major SNPLDB loci that had significant differences in phenotypic value were selected among several haplotypes/alleles, and they were regarded as the stable and true major SNPLDB loci for the FFBN and HFFBN. Moreover, we selected the 50 lines with the maximum FFBN and HFFBN values and the 50 lines with the minimum FFBN and HFFBN values among the 315 upland cotton accessions, and the promising haplotype/allele frequencies of the stable and true major SNPLDBs were determined for both groups. Additionally, the phenotypic values and the promising haplotype/allele frequencies were calculated and compared among the upland cotton lines from the five different geographic regions (NSER, YRR, YZRR, NIR and FR).

Meta-QTL analysis

On the basis of two widely used multimarker genetic maps employed extensively in multiple QTL mapping studies (Chen et al. 2015; Fang & Yu 2012), a reference genetic map was generated using the R package LRmerge and an optimized "synthesis" approach (Venske et al. 2019). All genetic maps and original QTLs for cotton FFBN and HFFBN from previous studies were collected. Then, the genetic maps were integrated into the reference genetic map, and a consensus map was established. Finally, the original QTLs that could be projected onto the consensus map were deemed the mapped QTLs with BioMercator v4.2.3 software (Sosnowski et al. 2012). MQTL analysis was also performed with BioMercator v4.2.3 software using the mapped QTLs. When the number of mapped QTLs was ≤10, the approach described by Goffinet and Gerber was applied to the meta-analysis (Goffinet and Gerber 2000). Alternatively, when the number of mapped QTLs on each chromosome was >10, the method proposed by Veyrieras and Charcosset was utilized for meta-analysis (Veyrieras and Charcosset 2007). Through the two methods, MQTLs for FFBN and HFFBN were identified, and the average values of the mapped QTLs projected onto the consensus map were computed as the R² and logarithm of odds (LOD) values of the MQTLs. Then, we utilized CottonGen (https://www.cottongen.org/) to search for the physical locations of MQTL flanking markers, and the physical intervals of the MQTLs were determined. Finally, the significant SNPLDB loci were mapped on the MQTL intervals for FFBN and HFFBN with MapChart v2.32 software.

Transcriptome data analysis

Integration of the major SNPLDB loci and the MQTLs for the target traits and the stable and true major SNPLDBs that were distributed in the MQTLs were used to mine possible candidate genes for FFBN and HFFBN. The genomic regions for the mining of candidate genes were ± 500 kb around the major SNPLDBs, similar to the genomic regions of SNP-linked candidate genes identified in our previous studies (Su et al., 2020a, b). In terms of the upland cotton reference genome v2.1 (Hu et al., 2019), annotated genes were downloaded for the target genomic regions. To detect potential candidate genes for FFBN and HFFBN, we utilized RNA-Seq data from five developmental stages (the cotyledon and one- to four-true-leaf stages) between ZMS50 (small HFFBN cultivar) and GXM11 (large HFFBN cultivar) (Cheng et al., 2021). The mean fragments per kilobase per million (FPKM) values were calculated as the expression levels of the annotated genes. The genes whose FPKM values significantly differed between ZMS50 and GXM11 were considered potential candidate genes.

Expression analysis of the potential candidate genes via qRT‒PCR

To identify the candidate genes controlling FFBN and HFFBN, two cotton cultivars (ZMS50 and ZMS74) with low FFBN and HFFBN and two other cultivars (STS458 and GXM11) with high FFBN and HFFBN were selected, and their shoot apical meristems and the youngest leaves were sampled at the third-true-leaf stage (Wang et al. 2023). Total RNA was subsequently extracted and reverse-transcribed into cDNA using a UnionScript First-strand cDNA Synthesis Mix (with dsDNase, No. SR511; Genesand, China) for qPCR. The qRT‒PCR-specific primers for the four potential candidate genes (the sequences of which are shown in Table S2) were designed on the Primer-BLAST website (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) and synthesized by Sangon Biotech (Shanghai) Co., Ltd. qRT‒PCR was performed on three replicates using FastReal qPCR PreMix (SYBR Green, No. FP217; Tiangen, China). GhActin (GenBank accession no. AY305733) was used as the internal control gene in the experiment. The expression values of GhActin and the potential candidate genes were calculated by using the 2^−ΔΔCT approach (Livak and Schmittgen 2001).

Silencing of the candidate genes by VIGS in upland cotton

To elucidate the functions of the candidate genes for FFBN and HFFBN, VIGS experiments were performed for three candidate genes (GhE6, GhDRM1 and GhGES). The silenced fragments (400–600 bp) of the three candidate genes were inserted into the TRV-based (pYL156) vector as the corresponding constructs, which were then introduced into Agrobacterium strain GV3101 by employing a freeze‒thaw technique. The primers used for the pYL156 vector construction for the three candidate genes are listed in Table S3. Subsequently, the helper vector pYL192 was mixed with strains TRV:GhCLA1, TRV:00, TRV:GhE6, TRV:GhDRM1 and TRV:GhGES at a 1:1 ratio and injected into 8-day-old cotton cotyledons after incubation at 28°C for 3 h. The seedlings infected with the virus were cultured in a dark environment at 25°C for 24 h and then transferred to a normal environment of 16 h light and 8 h darkness. After the positive control (TRV:GhCLA1) albinism emerged, gene-silenced plants whose expression levels were less than 50% of those of the TRV:00 plants were selected via qRT‒PCR for the TRV:GhE6, TRV:GhDRM1 and TRV:GhGES plants. Finally, we investigated four traits, namely, the FFBN, HFFBN, budding time (BT) and FT, of the TRV:00, TRV:GhE6, TRV:GhDRM1 and TRV:GhGES plants.

1. Phenotypic variation in the natural population

Two traits related to plant architecture and early maturity, the FFBN and HFFBN, were investigated in the natural population. We observed that the FFBN varied from 4.57 to 8.90 cm in E₁, 3.56 to 9.89 cm in E₂, and 3.49 to 6.33 cm in E₃, and the HFFBN ranged from 11.20 to 29.30 cm in E₁, 11.11 to 26.00 cm in E₂, and 10.22 to 26.00 cm in E₃, respectively. The data showed that the two target traits exhibited high phenotypic variation among the 315 lines with a normal distribution (Fig. 1). Correlation analysis revealed significant positive correlations between FFBN and HFFBN in all three environments (Fig. 1). The h² values of FFBN and HFFBN were 78.73% and 81.64%, respectively, suggesting high heritability of the target traits and precision of the experimental results. The ANOVA results showed that FFBN and HFFBN were significantly influenced by the genotype, environment, and interaction between them (Table S4). These results indicate that FFBN and HFFBN are controlled by heredity, the environment, and the interaction of the two, and that the first plays a dominant role.

2. Marker‑trait associations using the RTM-GWAS

In our previous research, a total of 13,391 high-quality SNP loci were detected in the association panel, and 9,244 SNPLDB loci with multiple haplotypes/alleles were organized as molecular markers (Su et al., 2020a). In the present study, to identify the significant SNPLDB loci related to FFBN and HFFBN, RTM-GWAS analyses were performed using SNPLDB markers constructed previously (Su et al., 2020a). According to the single-environment phenotypic values of FFBN and HFFBN, 28, 37 and 23 significant associations with FFBN were identified in E₁, E₂ and E₃, respectively (Fig. S1), and 51, 37 and 21 significant associations with HFFBN were identified in E₁, E₂ and E₃, respectively (Fig. S2). Moreover, 52 and 54 significant associations with FFBN and HFFBN, respectively, were detected using the total phenotypic values of the two target traits (Fig. 2, Table S5). Among these associations with the total phenotypic values of FFBN and HFFBN, eight associations involving seven SNPLDB loci with high -log₁₀(P) values were found in one or two single-environment associations (Fig. 2, Table 1). Among the seven SNPLDB loci, one locus (LDB_16_37952328) was concurrently associated with FFBN and HFFBN in two different single environments (E₁ and E₂) and was distinctly labeled with the numbers 1 and 7, respectively (Fig. 2, Table 1). Three SNPLDB loci (LDB_13_75038211_75038228, LDB_12_103278546 and LDB_5_15144433) associated with FFBN were detected in one environment (E₁) and were marked with numbers 2, 3 and 4, respectively (Fig. 2a, Table S2). In addition, three SNPLDB loci (LDB_11_112414092_112414298, LDB_18_37434283 and LDB_26_48647858) for the HFFBN were also identified in one environment (E₁) and were labeled with numbers 5, 6 and 8, respectively (Fig. 2b, Table 1). These results imply that the seven SNPLDBs involved in the eight associations labeled 1–8 should have stable major associations with FFBN and HFFBN.

3. Effects and matrices of the SNPLDB alleles with significant associations with FFBN and HFFBN

To investigate the allele effects of the significant associations with FFBN and HFFBN, the effect values of each SNPLDB allele for all the significant associations were determined though the RTM-GWAS procedure. Of the 52 significant associations with the FFBN, 38 and 14 single- and multi-locus associations were detected, respectively. Furthermore, 171 SNPLDB alleles, including 87 negative and 84 positive alleles, were identified for the 52 significant associations. The negative allele effect values varied from -0.84 to -0.0045, and the positive allele effect values ranged from 0.0054 to 1.37 (Fig. 3a). Among the 54 significant associations with HFFBN, 40 and 14 contained, respectively, one and multiple loci. Moreover, 170 SNPLDB alleles, including 84 negative and 86 positive alleles, were detected for the 54 significant associations with HFFBN. The allelic effect values of the 84 negative alleles ranged from −1.39 to −0.031 cm, and the effect values of the 86 positive alleles varied from 0.0094 to 4.30 cm (Fig. 3b). Additionally, we found that not all eight labeled associations with larger –log₁₀(P) values were loci with larger heredity effects.

Furthermore, to clearly display the allele effects of all the significant associations, we constructed diminutive allele effect matrices in the natural population. The allele effects of the 52 significant SNPLDB loci related to the FFBN were subsequently built into a 52 × 315 (locus × line) matrix for the 315 upland cotton lines (Fig. 3c). Similarly, the allele effects of the 54 significant SNPLDB loci associated with the HFFBN were organized into a 54 × 315 (locus × line) matrix in the natural population (Fig. 3d). The two SNPLDB allele matrices for FFBN and HFFBN from RTM-GWAS, which are abbreviated forms of hereditary constitution, provide comprehensive information, including positive and negative allele effects of the significant SNPLDB loci for each germplasm.

4. Mining of promising haplotypes/alleles with stable associations with FFBN and HFFBN

To identify promising haplotypes/alleles of FFBN and HFFBN, we selected seven SNPLDB loci involved in eight stable associations with the two target traits (Table 1) for further study. Among the eight stable associations with FFBN and HFFBN, one SNPLDB locus, LDB_16_37952328, which was concurrently related to FFBN and HFFBN, had three allelic variations (CC, CT and TT), and the FFBN and HFFBN of the lines with a positive allele (CC) were significantly greater than those of the lines with a negative allele (TT). In the natural population, for the SNPLDB locus LDB_5_15144433 associated with FFBN, the AA allele accessions had significantly greater FFBN than did the CC allele accessions (Fig. 4). LDB_11_112414092_112414298 associated with HFFBN produced five haplotypes (AACC, AACT, AATT, ACCC and CCCC), and the AACC haplotype lines had significantly larger HFFBN than did the AATT haplotype accessions. There were three alleles (CC, CT and TT) related to HFFBN in LDB_26_48647858, and the CC allele accessions had significantly higher HFFBN than did the AA allele accessions. For the other three associated loci (LDB_12_103278546, LDB_13_75038211_75038228, and LDB_18_37434283), however, there were no significant differences in the FFBN or HFFBN among the lines with different alleles (Fig. 4), implying that the three SNPLDBs might not be truly associated with the FFBN or HFFBN. Keeping mechanized harvesting of cotton cultivars with relatively larger FFBN and HFFBN in mind, the alleles with larger HFFBN were identified as promising haplotypes/alleles. In summary, we identified four promising haplotypes/alleles, namely, LDB_5_15144433 (AA), LDB_11_112414092_112414298 (AACC), LDB_16_37952328 (CC) and LDB_26_48647858 (CC), for five stable associations with FFBN and HFFBN. The results suggest that the genomic regions containing the four promising haplotypes/alleles of the five stable and true associations might be major QTLs regulating the FFBN or HFFBN in cotton.

5. Distribution frequency of promising haplotypes/alleles for the five stable and true major associations with FFBN and HFFBN in cotton accessions

To check the promising haplotype/allele effects of the five stable and true major associations with FFBN and HFFBN, we selected the 50 lines with the maximum and minimum phenotypic values, respectively, to determine the cumulative effect of promising alleles on the target traits. The promising haplotype/allele frequencies in the 50 lines with maximum phenotypic values were markedly greater than those in the 50 lines with minimum phenotypic values for the five stable and true major associations with FFBN and HFFBN (Fig. 5a, b). Moreover, we performed ANOVA for the FFBN and HFFBN values among lines from five diverse cultivation regions (YRR, YZRR, NIR, NSER and FR), and there was a striking difference in the FFBN and HFFBN among these regions (Fig. 5c, d). The mean FFBN values from highest to lowest were YZRR, FR, YRR, NIR and NSER. The ANOVA results showed that the FFBN values of the NSER lines were significantly lower than those of the FR, YZRR and YRR lines, and there was no significant difference in the FFBN between the NIR and NSER lines. We also found that the HFFBN significantly differed among the five diverse cultivated areas (Fig. 5d). The HFFBN of the NIR lines was significantly higher than that of the lines from the YZRR and YRR, and there were no significant differences in the HFFBN among the NIR, YZRR and FR lines. The results demonstrated that the NSER lines had relatively lower FFBN and HFFBN than did the YZRR, YRR and FR lines. Interestingly, we observed that the NIR lines had the highest HFFBN values, while they had lower FFBN values, indicating that the NIR lines had long internodes on the cotton plants. To investigate the regional distribution of promising haplotypes/alleles, we compared the frequency distributions of the four stable and true major SNPLDB loci among cotton lines from diverse cultivated areas. For the three major SNPLDB loci, except for LDB_5_15144433, the YZRR and NIR accessions had higher frequencies of promising haplotypes/alleles, and the YRR and NSER accessions had lower frequencies of promising haplotypes/alleles (Fig. 5e). These results imply that promising haplotypes/alleles might experience artificial selection during cotton breeding procedures.

6. Integration of MQTLs and the SNPLDB loci associated with FFBN and HFFBN

The 27 previous studies, published mainly from 2010 to 2020, that included QTL mapping of FFBN and HFFBN in cotton were systematically analyzed, and 167 and 107 original FFBN and HFFBN QTLs, accounting for 61% and 39%, respectively, were identified among the original QTLs (Table S6). The original QTL distribution was uneven across the26 chromosomes pairs of upland cotton, with approximately 39.05% (107/274) and 60.95% (167/274) on the A_t and D_t subgenomes, respectively. On chromosomes A11, D01, D03 and D08, 17, 19, 55 and 19 original QTLs were distributed, respectively, while on chromosomes D04 and D05, only two and three original QTLs were found, respectively. To merge the original QTLs from the 27 previous studies, an upgraded consensus map was established via 16 individual genetic maps from the 27 previous papers and the reference genetic map, which contained 5,044 SSR markers with an overall length of 3,830.90 cM and a mean length of 147.35 cM for each chromosome. Then, of the 274 original QTLs, 203 were projected onto the consensus map. The meta-analysis results showed that the 203 mapped QTLs constituted 40 MQTL regions (19.70%), and each MQTL contained at least two mapped QTLs. The MQTLs were unevenly dispersed across different upland cotton chromosomes. The four most common MQTLs were identified on chromosome D01, while no MQTLs were identified on chromosomes A06, A12, D04, D05, D06 or D13 (Fig. 6). Of the 40 MQTLs, 22 were composed of three or more mapped QTLs, and three MQTLs, namely, MQTL-D03.1, MQTL-D03.2 and MQTL-D08.2, contained seven, 17 and six mapped QTLs, respectively.

To integrate the MQTL information with the significant SNPLDBs related to FFBN and HFFBN, we compared the 40 MQTLs with the 52 and 54 SNPLDB loci for the two target traits. The results showed that 11 and 15 SNPLDB loci associated with FFBN and HFFBN, respectively, were distributed among the MQTLs (Fig. 6). Among the five stable and true major SNPLDB associations, two loci (LDB_16_37952328 and LDB_5_15144433), which were involved in three (FFBN-1, FFBN-4 and HFFBN-3), were distributed among the two MQTLs.

7. Identification of potential candidate genes forFFBN and HFFBN

To identify candidate genes related to FFBN and HFFBN, we focused on two stable and true major loci (LDB_5_15144433 and LDB_16_37952328) that were distributed among the MQTLs. Referring to the linkage disequilibrium decay distance (500 kb) of the SNPs (Su et al., 2020a, b), we selected two genome segments, D03:37.45–38.45 Mb and A05:14.64–15.64 Mb, to mine potential candidate genes for FFBN and HFFBN. Based on the reference genome v2.1 (Hu et al., 2019), 142 genes (Table S7) were annotated in the two genomic regions. Then, we analyzed the expression levels of the 142 genes using previously published transcriptome data (Cheng et al., 2021). The data were obtained from two upland cotton cultivars, including one lower-FFBN (5.31) and -HFFBN (13.20 cm) cultivar, ZMS50, and one higher-FFBN (6.69) and -HFFBN (19.69 cm) cultivar, GXM11 (Fig. 7a, b). Using standardized gene FPKM values from the RNA-Seq data, we detected 89 of the 142 annotated genes that were highly expressed in the shoot apical meristems and the youngest leaves during five growth periods. The heatmaps showed that four genes, namely, GH_A05G1602 (GhCIT), GH_A05G1606 (GhE6), GH_A05G1609 (GhDRM1) and GH_A05G1635 (GhGES), were highly differentially expressed in the cotyledon and 1–4-true-leaf stages between ZMS50 and GXM11 (Fig. 7c, d). To validate the four differentially expressed genes (DEGs), we conducted qRT‐PCR experiments on shoot apical meristems and the youngest leaves of the third-true-leaf stage between two low-FFBN and -HFFBN cultivars (ZMS50 and ZMS74) and two high-FFBN and HFFBN accessions (GXM11 and STS458). We found that four genes, GhCIT, GhE6, GhDRM1 and GhGES, were significantly differentially expressed between the shoot apical meristems and the youngest leaves of the third-true-leaf stage between the small and large FFBN and HFFBN accessions (Fig. 7e). The results indicated four candidate genes, GhCIT, GhE6, GhDRM1 and GhGES, for cotton FFBN and HFFBN.

8. Functional analysis of the candidate genes for FFBN and HFFBN in upland cotton

To confirm the biological function of the candidate genes for FFBN and HFFBN, we selected three highly significant DEGs (GhE6, GhDRM1 and GhGES) for the VIGS experiments. Eight days after virus injection, the TRV:GhCLA1 plants became albino (Fig. S3a), confirming that the VIGS assay was successful. Moreover, the expression levels of the three candidate genes were effectively lower in most of the gene-silenced plants than in the TRV:00 plants (Fig. S3b–d). To investigate the relationship between the gene expression levels and the target characteristics, we selected the gene-silenced plants in which the expression levels of the three candidate genes were half or less than those in the TRV:00 plants. Compared with the TRV:00 plants, the TRV:GhDRM1 and TRV:GhGES plants exhibited late-flowering phenotypes (Figure 8a), and the TRV:GhE6 plants exhibited a high HFFBN (Figure 8b, c). The BT and FT of the TRV:GhDRM1 and TRV:GhGES plants were significantly earlier than those of the TRV:00 plants, and the HFFBN and PH of the TRV:GhE6 plants were significantly higher than those of the TRV:00 plants. (Figure 8d). The results demonstrated that GhE6 is a key gene related to HFFBN and PH and that GhDRM1 and GhGES are important genes associated with early flowering in cotton.

Importance of FFBN and HFFBN in Chinese cotton production

According to the National Bureau of Statistics, in 2023, the cotton planting area in the NIR region of China (Xinjiang and Gansu) was 2.39 million hectares, producing 5.15 million tons of ginned cotton, accounting for 85.71% of the country's cotton planting area and 91.74% of its cotton output. Plant architecture and early maturation are two vital factors in mechanized harvesting processes for cotton production in the NIR region (Mao et al. 2014, 2015). FFBN and HFFBN are important traits related to plant architecture; they are two key indicators used to evaluate early maturation and are significantly and positively correlated with FT and WGP (Godoy and Palomo, 1999; Jia et al., 2016; Guo et al., 2008; Su et al., 2016, 2018; Zhang et al., 2021). In the present study, the broad-sense heritability of FFBN and HFFBN was high, suggesting that these two traits are highly influenced by genetic factors (Fan et al. 2006; Fu et al. 2019). Additionally, previous studies have shown that early-maturing cotton cultivars have lower FFBN and HFFBN (Zhao et al., 1974; Yu et al. 2017). Therefore, FFBN and HFFBN are often considered reliable indices for early-maturing cotton. HFFBN has a great influence on mechanized cotton picking. Past studies have shown that an HFFBN of 18–20 cm is needed to facilitate mechanized cotton picking (Nusirat et al., 2012; Zhang et al., 2022). If the HFFBN of machine-picked cotton is too low, it can easily reduce the net harvesting rate and affect the harvest yield, and can increase the level of impurities and affect the quality of the raw cotton (Chen et al., 2019). Hence, FFBN and HFFBN are two important traits related to mechanized cotton harvesting in Chinese cotton production.

Advantages of mining QTLs by the integration of the RTM-GWAS procedure and MQTL analysis

GWAS based on linkage disequilibrium and high-throughput sequencing technologies is a highly accurate approach for mapping QTLs in natural populations (Fu et al. 2020; Yang et al., 2020). The earliest GWAS method that was widely applied was the MLM on the basis of a single-locus model (Yu et al., 2006), a method that is suitable for detecting a few major loci for target traits and is likely to identify some false-positive loci (Li et al., 2017). Fortunately, a recent RTM-GWAS procedure based on a multiallele model has high power and efficiency in QTL exploration. Hence, compared with the MLM-GWAS method, the RTM-GWAS procedure can enable the identification of a greater number of stable and true QTLs in natural populations (He et al. 2017; Pan et al. 2018; Su et al. 2020; Wang et al. 2022). The MQTL analysis method, which is used to identify coincident QTLs through multiple studies, has been widely used to detect consensus genomic regions of target traits across various genetic backgrounds and different planting environments (Goffinet and Gerber 2000; Veyrieras et al. 2007). Previous studies have shown that GWAS and QTL mapping are two complementary approaches for locating candidate genes and elucidating the hereditary basis of target characteristics in plants (Mahuku et al. 2016; Tao et al. 2013).

RTM-GWAS is a robust and efficient approach for identifying a greater number of significant loci related to target characteristics in natural populations. MQTL analysis is an efficient method for assembling QTLs from distinct populations. Each of the two methods has its own advantages and disadvantages and they therefore complement each other (Sonah et al. 2015). The integration of GWAS results with MQTL analysis has enabled the effective identification of stable and true major QTLs (Daryani et al. 2022). According to our RTM-GWAS results, 52 and 54 significant associations with FFBN and HFFBN, respectively, were identified using the total phenotypic values of the two target traits (Fig. 2). Usually, SNPLDBs that had high -log₁₀(P) values and were simultaneously detected using the total environment and one or more single-environment phenotypic values were considered stable major loci. In accordance with this principle, seven stable major SNPLDB loci for FFBN and HFFBN were discovered in this study (Fig. 2). In the present MQTL analysis, to polymerize the genomic regions controlling the target traits, we collected information on 167 and 107 QTLs for the NFFB and HNFFB, respectively, from 27 articles published between 2010 and 2020. A total of 40 MQTLs were identified, and we found that three important MQTLs (MQTL-D03.1, MQTL-D03.2 and MQTL-D08.2) contained seven, 17 and six mapped QTLs, respectively. Through the integration of MQTLs and significant SNPLDB loci, 11 and 15 SNPLDBs related to the FFBN and HFFBN, respectively, were distributed in the MQTL intervals. More usefully, we found two stable and true major SNPLDB loci (LDB_5_15144433 and LDB_16_37952328) involved in the three associations (FFBN-1, FFBN-4 and HFFBN-3) in the two MQTLs. Hence, genomic regions containing the two major SNPLDB loci were chosen for the identification of candidate genes for FFBN and HFFBN.

Comparison of the SNPLDB loci for FFBN and HFFBN with those identified in previous GWASs

To validate our GWAS results, we compared the significant SNPLDBs for FFBN and HFFBN with the significant SNP loci for the two target traits identified in previous studies. Previous studies have identified numerous SNP loci for FFBN and HFFBN via GWAS, and several genome regions for FFBN and HFFBN have been identified. According to the reference genome v2.1 of upland cotton (Hu et al., 2019), the four genome regions related to the FFBN and HFFBN were A05: 15.88–16.24 Mb (Fu et al., 2019), A06: 95.78–98.30 Mb (Shen et al., 2019), D03: 11.36–18.19 Mb (Shen et al., 2019), and D03: 39.34–41.84 Mb (Su et al., 2018; Shen et al., 2019; Li et al., 2021) (Table S8). In the present study, four stable and true major SNPLDB loci, namely, LDB_5_15144433, LDB_11_112414092_112414298, LDB_16_37952328 and LDB_26_48647858, were not distributed in the four genome regions related to FFBN and HFFBN identified in previous studies. The results demonstrated that the four stable and true major SNPLDB loci identified via the RTM-GWAS procedure were novel genetic loci for cotton FFBN and HFFBN. Additionally, we observed that the number of significant SNPLDB loci identified via the RTM-GWAS approach in the present study was greater than the number of significant SNPs identified by MLM-GWAS in previous studies.

Interestingly, we found that LDB_16_37952328 was simultaneously related to FFBN, HFFBN, PH, FT and WGP (Wang et al., 2022, 2023). Moreover, striking differences in FFBN and HFFBN were found among the accessions from the five distinct cotton cultivation areas, and the HFFBN of the NIR lines was significantly greater than that of the YZRR and YRR lines (Fig. 5c, d). Similarly, we also observed that the frequencies of promising haplotypes/alleles of the four stable and true major SNPLDBs exhibited visible differences among the five cultivated regions (Fig. 5e). These results imply that promising haplotypes/alleles might undergo artificial selection and/or natural selection in the process of cotton breeding.

Elucidation of genomic regions and candidate genes related to FFBN and HFFBN

During the past five years, several candidate genes controlling FFBN and HFFBN in upland cotton have been predicted. Two and 22 SNPs were identified for FFBN and HFFBN, respectively, in at least two environments and/or by two GWAS models. The authors of one study speculated that Gh_A05G1482 might be a potential candidate gene for cotton FFBN and HFFNB (Fu et al., 2019). Via QTL-seq, another study showed that 17 QTLs associated with FFBN were mapped, and two of them (qNFFB-D_t2–1 and qNFFB-D_t3–3) were located in two hotspots. In the two QTL intervals, two possible genes (GhAPL and GhHDA5) controlling FFBN were predicted by qRT‒PCR, and compared with the flowering period and the FFBN of empty vector-transformed plants, the flowering period was delayed and the FFBN increased in cotton plants with the two candidate genes silenced (Zhang et al., 2021). In another study, two genomic regions related to FFBN were found via BSA-seq on D03 (41,779,195–41,836,120 bp and 41,836,768–41,872,287 bp), and two probable genes (Ghir_D03G012430 and Ghir_D03G012390) controlling FFBN were identified in these two regions by qRT‒PCR (Jia et al., 2023). In this study, two genomic regions (D03:37.45–38.45 Mb and A05:14.64–15.64 Mb) associated with FFBN and HFFBN were detected by integrating MQTLs and the SNPLDB loci for the target traits, and three candidate genes (GhE6, GhDRM1 and GhGES) were identified in the region of A05:14.64–15.64 Mb by using RNA-Seq and qRT‒PCR. Furthermore, the VIGS results showed that GhE6 is a key modulator related to HFFBN and PH and that GhDRM1 and GhGES are key genes associated with early flowering in cotton.

Acknowledgements

We especially thank Professor Shuxun Yu and his team (at the Institute of Cotton Research of CAAS), who provided us with good guidance for the experiment. This research was funded by Natural Science Foundation of Xinjiang Uygur Autonomous Region (2022D01E103), the Chinese National Natural Science Foundation (32260522) and the technological Innovation 2030- Major Project Topic (2023ZD0404103).

Author Contribution statement

JS designed the research program; WY, JJ and PL analyzed the data; JS, KF and JJ performed the field trial; WY, NW and JS conducted the RTM-GWAS and meta‑QTL analyses; DL, YL and CXW performed the qRT-PCR and VIGS; and JJS and CXW wrote the manuscript. All of the authors read and approved the manuscript.

Conflict of interest All authors declare that they have no conflicts of interests.

Ethical standards We declare that these experiments complied with the ethical standards in China.

Arcade A, Labourdette A, Falque M, Mangin B, Chardon F, Charcosset A, Joets J (2004) BioMercator: integrating genetic maps and QTL towards discovery of candidate genes. Bioinformatics 20:2324-2326
Badji A, Otim M, Machida L, Odong T, Kwemoi DB, Okii D, Agbahoungba S, Mwila N, Kumi F, Ibanda A, Mugo S, Kyamanywa S, Rubaihayo P (2018) Maize combined insect resistance genomic regions and their co-localization with cell wall constituents revealed by tissue-specifc QTL meta-analyses. Front Plant Sci 9:895
Cai J, Wang S, Su Z, Li T, Zhang X, Bai G (2019) Meta-analysis of QTL for Fusarium head blight resistance in Chinese wheat landraces. Crop J 7:784-798
Chen H, Khan MK, Zhou Z, Wang X, Cai X, Ilyas MK, Wang C, Wang Y, Li Y, Liu F, Wang K (2015) A high-density SSR genetic map constructed from a F₂ population of Gossypium hirsutum and Gossypium darwinii. Gene 2574:273-286
Chen M, Yang Y, Wang Y, Tian J, Xu S, Liu N, Dang K, Zhang W (2019) Plant type characteristics and evolution of main economic characters in early maturing upland cotton cultivar replacement in Xinjiang. Scientia Agricultura Sinica 52:3279-3290.
Chen ZJ, Scheffler BE, Dennis E, Triplett BA, Zhang T, Guo W, Chen X, Stelly DM, Rabinowicz PD, Town CD, Arioli T, Brubaker C, Cantrell RG, Lacape JM, Ulloa M, Chee P, Gingle AR, Haigler CH, Percy R, Saha S, Wilkins T, Wright RJ, Van Deynze A, Zhu Y, Yu S, Abdurakhmonov I, Katageri I, Kumar PA, Mehboob-Ur-Rahman, Zafar Y, Yu JZ, Kohel RJ, Wendel JF, Paterson AH (2007) Toward sequencing cotton (Gossypium) genomes. Plant Physiol 145:1303-10
Cheng S, Chen P, Su Z, Ma L, Hao P, Zhang J, Ma Q, Liu G, Liu J, Wang H, Wei H, Yu S (2021) High-resolution temporal dynamic transcriptome landscape reveals a GhCAL-mediated flowering regulatory pathway in cotton (Gossypium hirsutum L.). Plant Biotechnol J 19:153-166
Darvasi A, Soller M (1997) A simple method to calculate resolving power and confidence interval of QTL map location. Behav Genet 27:125-132
Daryani P, Ramandi HD, Dezhsetan S, Mansuri RM, Salekdeh GH, Shobbar ZS (2022) Pinpointing genomic regions associated with root system architecture in rice through an integrative meta‑analysis approach. Theor Appl Genet 135:81-106
Fan S, Yu S, Yuan R, Song M (2006) Genetic effects and environmental interactions of early maturity in short-season cotton. Acta Bot Bor Occid Sin 26(11):2270-2275
Fang D, Yu J (2016) Addition of 455 Microsatellite Marker Loci to the High-Density Gossypium hirsutum TM-1 × G. barbadense 3-79 Genetic Map. J Cotton Sci 16:229-248
Fu M, Wang Y, Ren H, Du W, Wang D, Bao R, Yang X, Tian Z, Fu L, Cheng Y, Su J, Sun B, Zhao J, Gai J (2020a) Genetic dynamics of earlier maturity group emergence in south-to-north extension of Northeast China soybeans. Theor Appl Genet 133:1839-1857
Fu M, Wang Y, Ren H, Du W, Yang X, Wang D, Cheng Y, Zhao J, Gai J (2020b) Exploring the QTL-allele constitution of main stem node number and its differentiation among maturity groups in a Northeast China soybean population. Crop Sci 60:1223-1238
Fu Y, Dong C, Wang J, Wang Y, Li CQ (2019) Genome-wide association study reveals the genetic control underlying node of the first fruiting branch and its height in upland cotton (Gossypium hirsutum L.). Euphytica 215:35
Godoy A, Palomo G (1999) Genetic analysis of earliness in upland cotton (Gossypium hirsutum L.). i. morphological and phenological variables. Euphytica 105: 155-160.
Gofnet B, Gerber S (2000) Quantitative trait loci: a meta-analysis. Genetics 155:463-473
Guo B, Sleper DA, Lu P, Shannon JG, Nguyen HT, Arelli PR (2006) QTL associated with resistance to soybean cyst nematode in soybean: meta-analysis of QTL locations. Crop Sci 46:595-602
Guo Y, McCarty JC, Jenkins JN, An C, Saha S (2009) Genetic detection of node of first fruiting branch in crosses of a cultivar with two exotic accessions of upland cotton. Euphytica 166(3):317-329
Guo Y, McCarty JC, Jenkins JN, Saha S (2008) QTL for node of first fruiting branch in a cross of an upland cotton, Gossypium hirsutum L., cultivar with primitive accession Texas 701. Euphytica 163:113-122
Hanson CH, Robinson HF, Comstock RE (1956) Biometrical studies of yield in segregating populations of Korean Lespedeza. Agron J 48:267-282
He J, Meng S, Zhao T, Xing G, Yang S, Li Y, Guan R, Lu J, Wang Y, Xia Q (2017) An innovative procedure of genome-wide association analysis fits studies on germplasm population and plant breeding. Theor Appl Genet 130:2327-2343
Hu Y, Chen J, Fang L, Zhang Z, Ma W, Niu Y, Ju L, Deng J, Zhao T, Lian J, Baruch K, Fang D, Liu X, Ruan Y, Rahman MU, Han J, Wang K, Wang Q, Wu H, Mei G, Zang Y, Han Z, Xu C, Shen W, Yang D, Si Z, Dai F, Zou L, Huang F, Bai Y, Zhang Y, Brodt A, Ben-Hamo H, Zhu X, Zhou B, Guan X, Zhu S, Chen X, Zhang T (2019) Gossypium barbadense and Gossypium hirsutum genomes provide insights into the origin and evolution of allotetraploid cotton. Nat Genet 51:739-748.
Huo W, Zhang Z, Ren Z, Zhao J, Song C, Wang X, Pei X, Liu Y, He K, Zhang F, Li X, Li W, Yang D, Ma X (2023) Unraveling genomic regions and candidate genes for multiple disease resistance in upland cotton using meta-QTL analysis. Heliyon 9:e18731
Jia X, Pang C, Wei H, Wang H, Ma Q, Yang J, Cheng S, Su J, Fan S, Song M, Wusiman N, Yu S (2016) High-density linkage map construction and QTL analysis for earliness-related traits in Gossypium hirsutum L. BMC Genom 17:909
Jia X, Wang, Zhao H, Zhu J, Li M, Wang G (2023) QTL mapping and BSA-seq map a major QTL for the node of the first fruiting branch in cotton. Front Plant Sci 14:1113059
Khahani B, Tavakol E, Shariati V, Fornara F (2020) Genome wide screening and comparative genome analysis for Meta-QTLs, ortho-MQTLs and candidate genes controlling yield and yield-related traits in rice. BMC Genom 21:294
Khan M, Tong F, Wang W, He J, Zhao T, Gai J (2018) Analysis of QTL-allele system conferring drought tolerance at seedling stage in a nested association mapping population of soybean [Glycine max (L.) Merr.] using a novel GWAS procedure. Planta 248:947-962
Khowaja F, Norton G, Courtois B, Price A (2009) Improved resolution in the position of drought-related QTLs in a single mapping population of rice by meta-analysis. BMC Genom 10:276
Kumar A, Saripalli G, Jan I, Kumar K, Sharma PK, Balyan HS, Gupta PK (2020) Meta-QTL analysis and identifcation of candidate genes for drought tolerance in bread wheat (Triticum aestivum L.). Physiol Mol Biol Plants 26:1713-1725
Li C, Wang C, Dong N, Wang X, Zhao H, Converse R, Xia Z, Wang R, Wang Q (2012) QTL detection for node of first fruiting branch and its height in upland cotton (Gossypium hirsutum L.). Euphytica 188:441-451
Li C, Wang Y, Ai N, Li Y, Song J (2018) A genome-wide association study of early-maturation traits in upland cotton based on the CottonSNP80K array. J Integr Plant Biol 60:970-985
Li C, Zhang J, Hu G, Fu Y, Wang Q (2016) Association mapping and favorable allele mining for node of first fruiting/sympodial branch and its height in upland cotton (Gossypium hirsutum L.). Euphytica 210:57-68
Li L, Zhang C, Huang J, Liu Q, Wei H, Wang H, Liu G, Gu L, Yu S (2021) Genomic analyses reveal the genetic basis of early maturity and identification of loci and candidate genes in upland cotton (Gossypium hirsutum L.). Plant Biotechnol J 19: 109-123
Li L, Zhao S, Su J, Fan S, Pang C, Wei H, Wang H, Gu L, Zhang C, Liu G, Yu D, Qi Liu, Zhang X, Yu S (2017) High-density genetic linkage map construction by F₂ populations and QTL analysis of early-maturity traits in upland cotton (Gossypium hirsutum L.). PLoS ONE 12:e0182918
Li S, Cao Y, He J, Zhao T, Gai J (2017) Detecting the QTL-allele system conferring fowering date in a nested association mapping population of soybean using a novel procedure. Theor Appl Genet 130:2297-2314
Liu F, He J, Wang W, Xing G, Zhao J, Li Y, Gai J (2021a) Genetic dynamics of fowering date evolved from later to earlier in annual wild and cultivated soybean in china. Crop Sci 61:2336-2354
Liu H, Mullan D, Zhang C, Zhao S, Li X, Zhang A, Lu Z, Wang Y, Yan G (2020a) Major genomic regions responsible for wheat yield and its components as revealed by meta-QTL and genotype-phenotype association analyses. Planta 252:65
Liu X, Li C, Cao J, Zhang X, Wang C, He J, Xing G, Wang W, Zhao J, Gai J (2021b) Growth period QTL-allele constitution of global soybeans and its differential evolution changes in geographic adaptation versus maturity group extension. Plant J 108:1624-1643
Liu Y, Salsman E, Wang R, Galagedara N, Zhang Q, Fiedler J, Liu Z, Xu S, Faris J, Li X (2020b) Meta-QTL analysis of tan spot resistance in wheat. Theor Appl Genet 133:2363-2375
Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2^-^ΔΔ^CT method. Methods 25:402-8.
Mahuku G, Chen J, Shrestha R, Narro LA, Guerrero KVO, Arcos AL, Xu Y (2016) Combined linkage and association mapping identifies a major QTL (qRtsc8-1), conferring tar spot complex resistance in maize. Theor Appl Genet 129:1217-1229
Mao L, Zhang L, Evers JB, van der Werf W, Liu S, Zhang S, Wang B, Li Z (2015) Yield components and quality of intercropped cotton in response to mepiquat chloride and plant density. Field Crops Res 179:63-71
Mao L, Zhang L, Zhao X, Liu S, van der Werf W, Zhang SP, Spiertz H, Li Z (2014) Crop growth, light utilization and yield of relay intercropped cotton as affected by plant density and a plant growth regulator. Field Crops Res 155:67-76
Meng S, He J, Zhao T, Xing G, Li Y, Yang S, Lu J, Wang Y, Gai J (2016) Detecting the QTL-allele system of seed isofavone content in Chinese soybean landrace population for optimal cross design and gene system exploration. Theor Appl Genet 129:1557-1576
Nusirat·Osman, Yu S, Fan S, Mei Y, Yuan R (2012) Analysis of genetic contribution of mechanical harvesting traits to lint yield in upland cotton. Cotton Sci 24: 10-17.
Pan L, He J, Zhao T, Xing G, Wang Y, Yu D, Chen S, Gai J (2018) Efficient QTL detection of flowering date in a soybean RIL population using the novel restricted two-stage multi-locus GWAS procedure. Theor Appl Genet 131:2581-2599
Said J, Song M, Wang H, Lin Z, Zhang X, Fang D, Zhang J (2015) A comparative meta-analysis of QTL between intraspecific Gossypium hirsutum and interspecifc G. hirsutum × G. barbadense populations. Mol Genet Genomics 290:1003-1025
Said JI, Lin Z, Zhang X, Song M, Zhang J (2013) A comprehensive meta QTL analysis for fiber quality, yield, yield related and morphological traits, drought tolerance, and disease resistance in tetraploid cotton. BMC Genom 14:776
Shen C, Wang N, Huang C, Wang M, Zhang X, Lin Z (2019) Population genomics reveals a fine‐scale recombination land scape for genetic improvement of cotton. Plant J 99: 494-505
Sonah H, O’Donoughue L, Cober E, Rajcan I, Belzile F (2015) Identifcation of loci governing eight agronomic traits using a GBS-GWAS approach and validation by QTL mapping in soybean. Plant Biotechnol J 3:211-221
Soriano JM, Royo C (2015) Dissecting the genetic architecture of leaf rust resistance in wheat by QTL meta-analysis. Phytopathology 105:1585-1593
Sosnowski O, Charcosset A, Joets J (2012) Biomercator V3: An upgrade of genetic map compilation and quantitative trait loci meta-analysis algorithms. Bioinformatics 28:2082-2083
Su J, Li L, Zhang C, Wang C, Gu L, Wang H, Wei H, Liu Q, Huang L, Yu S (2018) Genome-wide association study identifed genetic variations and candidate genes for plant architecture component traits in Chinese upland cotton. Theor Appl Genet 131:1299-1314
Su J, Pang C, Wei H, Li L, Liang B, Wang C, Song M, Wang H, Zhao S, Jia X (2016) Identifcation of favorable SNP alleles and candidate genes for traits related to early maturity via GWAS in upland cotton. BMC Genom 17:687
Su J, Wang C, Ma Q, Zhang A, Shi C, Liu J, Zhang X, Yang D, Ma X (2020a) An RTM-GWAS procedure reveals the QTL alleles and candidate genes for three yield-related traits in upland cotton. BMC Plant Biol 20:416
Su J, Wang C, Yang D, Shi C, Zhang A, Ma Q, Liu J, Zhang X, Huang L, Ma X (2020b). Decryption of favourable haplotypes and potential candidate genes for five fibre quality properties using a relatively novel genome-wide association study procedure in upland cotton. Ind Crop Prod 158:113004
Su Y, Zhang Z, He J, Zeng W, Cai Z, Lai Z, Pan Y, Hao X, Xing G, Wang W, Zhang J, Li Y, Sun Z, Gai J (2023) Gene-allele system of shade tolerance in southern China soybean germplasm revealed by genome-wide association study using gene-allele sequence as markers. Theor Appl Genet 136:152
Tai L, Wang H, Xu X, Sun W, Ju L, Liu W, Li W, Sun J, Chen K (2021) Cereal pre-harvest sprouting: a global agricultural disaster regulated by complex genetic and biochemical mechanisms. J Exp Bot 72:2857-2876
Tao Y, Jiang L, Liu Q, Zhang Y, Zhang R, Ingvardsen CR, Frei UK, Wang B, Lai J, Lübberstedt T, Xu M (2013) Combined linkage and association mapping reveals candidates for Scmv1, a major locus involved in resistance to sugarcane mosaic virus (SCMV) in maize. BMC Plant Biol 13:162
Venske E, Dos Santos RS, Farias DD, Rother V, Maia LC, Pegoraro C, Costad e Oliveira A (2019) Meta-analysis of the QTL ome of Fusarium head blight resistance in bread wheat: refning the current puzzle. Front Plant Sci 10:727
Veyrieras JB, Goffinet B, Charcosset A (2007) MetaQTL: a package of new computational methods for the meta-analysis of QTL mapping experiments. BMC Bioinform 8:49
Wang C, Liu J, Xie X, Wang J, Ma Q, Chen P, Yang D, Ma X, Hao F, Su J (2023) GhAP1-D3 positively regulates flowering time and early maturity with no yield and fiber quality penalties in upland cotton. J Integr Plant Biol 65:985-1002
Wang C, Ma Q, Xie X, Zhang X, Yang D, Su J, Ma X, Lin H (2022) Identification of favorable haplotypes/alleles and candidate genes for three plant architecture-related traits via a restricted two-stage multilocus genome-wide association study in upland cotton. Ind Crop Prod 177, 114458
Wang Y, Huang Z, Deng D, Ding H, Zhang R, Wang S, Bian Y, Yin Z, Xu X (2013) Meta-analysis combined with syntenic meta-QTL mining dissects candidate loci for maize yield. Mol Breeding 31:601-614
Wang Y, Wang Y, Wang X, Deng D (2020) Integrated meta-QTL and genome-wide association study analyses reveal candidate genes for maize yield. J Plant Growth Regul 39:229-238
Welcker C, Sadok W, Dignat G, Renault M, Salvi S, Charcosset A, Tardieu F (2011) A common genetic determinism for sensitivities to soil water defcit and evaporative demand: meta-analysis of quantitative trait loci and introgression lines of maize. Plant Physiol 157:718-729
Yang Y, Amo A, Wei D, Chai Y, Zheng J, Qiao P, Cui C, Lu S, Chen L, Hu Y (2021) Large-scale integration of meta-QTL and genome-wide association study discovers the genomic regions and candidate genes for yield and yield-related traits in bread wheat. Theor Appl Genet 134:3083-3109
Yang Y, Chai Y, Zhang X, Lu S, Zhao Z, Wei D, Chen L, Hu YG (2020) Multi-Locus GWAS of quality traits in bread wheat: mining more candidate genes and possible regulatory network. Front Plant Sci 11:1091
Yang Z, Gao C, Zhang Y, Yan Q, Hu W, Yang L, Wang Z and Li F (2023) Recent progression and future perspectives in cotton genomic breeding. J Integr Plant Biol 65:548-569.
Yu J, Pressoir G, Briggs WH, Vroh BI, Yamasaki M, Doebley JF, Mcmullen MD, Gaut BS, Nielsen DM, Holland JB (2006) A unifed mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet 38:203-208
Yu JZ, Kohel RJ, Fang DD, Cho J, Van Deynze A, Ulloa M, Hoffman SM, Pepper AE, Stelly DM, Jenkins JN, Saha S, Kumpatla SP, Shah MR, Hugie WV, Percy RG (2012) A high-density simple sequence repeat and single nucleotide polymorphism genetic map of the tetraploid cotton genome. G3 (Bethesda, Md.) 2:43-58
Yu L, Barbier H, Rouse MN, Singh S, Singh RP, Bhavani S, Huerta-Espino J, Sorrells ME (2014) A consensus map for Ug99 stem rust resistance loci in wheat. Theor Appl Genet 127:1561-1581
Yu S, Wang H, Wei H, Su J (2017) Research Progress and Application of Early Maturity in Upland Cotton. Cotton Sci 29:1-10
Zhang J, Jia X, Guo X, Wei H, Zhang M, Wu A, Cheng S, Cheng X, Yu S, Wang H (2021) QTL and candidate gene identification of the node of the frst fruiting branch (NFFB) by QTL-seq in upland cotton (Gossypium hirsutum L.). BMC Genom 22:882.
Zhang J, Yu J, Pei W, Li X, Said J, Song M, Sanogo S (2015) Genetic analysis of Verticillium wilt resistance in a backcross inbred line population and a meta-analysis of quantitative trait loci for disease resistance in cotton. BMC Genom 16:577
Zhang W, Liu Q, Zeng Q, Wang Z, Feng Y, Lu T (2022) Effects of different row spacing on the growth and photosynthetic characteristics of machine picked cotton. Agric. Res. Arid. Areas 40:155-164.
Zhao L, Chen S, Xu S (1974) An estimation of heritability of the indicative characters of earliness in upland cotton. Acta Genet Sin 1:107-115
Zheng T, Hua C, Li L, Sun Z, Yuan M, Bai G, Humphreys G, Li T (2020) Integration of meta-QTL discovery with omics: towards a molecular breeding platform for improving wheat resistance to Fusarium head blight. Crop J 9:739-749

Table 1 The significant SNPLDB loci associated with the FFBN and HFFBN

Numbers	Traits	Loci	Chr.	Positions	-lg (P) values	Single environments
1	FFBN	LDB_16_37952328	16	37952328	21.21	E1(6.15), E2(6.04)
2		LDB_13_75038211_75038228	13	75038211	15.09	E1(19.42)
3		LDB_12_103278546	12	103278546	13.85	E1(5.03)
4		LDB_5_15144433	5	15144433	12.97	E1(6.47)
5	HFFBN	LDB_11_112414092_112414298	11	112414092	22.41	E3(15.79)
6		LDB_18_37434283	18	37434283	12.51	E1(11.50)
7		LDB_16_37952328	16	37952328	11.77	E1(10.26), E2(9.73),
8		LDB_26_48647858	26	48647858	10.10	E1(3.18)

Fig.S1.pdf
Fig.S1 Manhattan plot for RTM-GWAS using the single-environment phenotypic value of FFBN in E₁(a), E₂(b) and E₃(c).
Fig.S2.pdf
Fig.S2 Manhattan plot for RTM-GWAS using the single-environment phenotypic value of HFFBN in E₁(a), E₂(b) and E₃(c)
Fig.S3.pdf
Fig.S3 Three highly significant DEGs (GhE6, GhDRM1 and GhGES) for the VIGS experiments. (a) The leaves of the TRV:GhCLA1 plants became albino; (b–d) the expression levels of the three candidate genes in the gene-silenced and TRV:00 plants.
TableS1.xlsx
Table S1 Information of the 315 upland cotton accessions
Table.S2.docx
Table S2 Sequences of the primer pairs used in the qRT-PCR of candidate genes.
Table.S3.docx
Table S3 Sequences of the primer pairs of constructing VIGS vectors.
Table.S4.docx
Table S4 Analysis of variance of cotton FFBN and HFFBN.
Table.S5.xlsx
Table S5 Totals of 52 and 54 significant associations with the total phenotypic values of FFBN and HFFBN.
Table.S6.xlsx
Table S6 Original QTL information of cotton FFBN and HFFBN for Meta-QTL analysis.
Table.S7.xlsx
Table S7 A total of 142 genes annotated in the two adjacent genome fragments of the two stable and true SNPLDB loci.
Table.S8.xlsx
Table S8 Comparison of the significant SNPLDBs for FFBN and HFFBN with the significant SNP loci for the two target traits identified in previous studies.

Download PDF

Journal Publication

published 22 Aug, 2024

Read the published version in Theoretical and Applied Genetics →

Editorial decision: Minor revisions
02 Jul, 2024
Reviewers agreed at journal
01 May, 2024
Reviewers invited by journal
30 Apr, 2024
Editor assigned by journal
23 Apr, 2024
First submitted to journal
22 Apr, 2024

You are reading this latest preprint version

Integrating RTM-GWAS and meta‑QTLs uncovers the genomic regions and candidate genes for the first fruit branch node and its height in upland cotton

Status:

Journal Publication

Version 1

Abstract

Figures

Key message

Introduction

Materials and Methods

Results

Discussion

Declarations

References

Table

Supplementary Files

Status:

Journal Publication

Version 1