Next Article in Journal
Nucleolar Expression and Chromosomal Associations in Robertsonian Spermatocytes of Mus musculus domesticus
Previous Article in Journal
At the Beginning of the End and in the Middle of the Beginning: Structure and Maintenance of Telomeric DNA Repeats and Interstitial Telomeric Sequences
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Genes
Volume 10
Issue 2
10.3390/genes10020119
genes-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Co-Expression Network Analysis and Hub Gene Selection for High-Quality Fiber in Upland Cotton (Gossypium hirsutum) Using RNA Sequencing Analysis

by
Xianyan Zou
1,2,†,
Aiying Liu
1,†,
Zhen Zhang
1,
Qun Ge
1,
Senmiao Fan
1,
Wankui Gong
1,
Junwen Li
1,
Juwu Gong
1,
Yuzhen Shi
1,
Baoming Tian
2,
Yanling Wang
1,
Ruixian Liu
1,
Kang Lei
1,
Qi Zhang
1,
Xiao Jiang
1,
Yulong Feng
1,2,
Shuya Zhang
1,
Tingting Jia
1,
Lipeng Zhang
1,
Youlu Yuan
1,2,* and
Haihong Shang
1,2,*add Show full author list remove Hide full author list
1
State Key Laboratory of Cotton Biology, Key Laboratory of Biological and Genetic Breeding of Cotton, The Ministry of Agriculture, Institute of Cotton Research, Chinese Academy of Agricultural Sciences, Anyang 455000, Henan, China
2
School of Agricultural Science, Zhengzhou University, Zhengzhou 450001, Henan, China
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Genes 2019, 10(2), 119; https://doi.org/10.3390/genes10020119
Submission received: 28 November 2018 / Revised: 27 January 2019 / Accepted: 30 January 2019 / Published: 6 February 2019
(This article belongs to the Section Plant Genetics and Genomics)
Browse Figures
Review Reports Versions Notes

Abstract

:
Upland cotton (Gossypium hirsutum) is grown for its elite fiber. Understanding differential gene expression patterns during fiber development will help to identify genes associated with fiber quality. In this study, we used two recombinant inbred lines (RILs) differing in fiber quality derived from an intra-hirsutum population to explore expression profiling differences and identify genes associated with high-quality fiber or specific fiber-development stages using RNA sequencing. Overall, 72/27, 1137/1584, 437/393, 1019/184, and 2555/1479 differentially expressed genes were up-/down-regulated in an elite fiber line (L1) relative to a poor-quality fiber line (L2) at 10, 15, 20, 25, and 30 days post-anthesis, respectively. Three-hundred sixty-three differentially expressed genes (DEGs) between two lines were colocalized in fiber strength (FS) quantitative trait loci (QTL). Short Time-series Expression Miner (STEM) analysis discriminated seven expression profiles; gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation were performed to identify difference in function between genes unique to L1 and L2. Co-expression network analysis detected five modules highly associated with specific fiber-development stages, especially for high-quality fiber tissues. The hub genes in each module were identified by weighted gene co-expression network analysis. Hub genes encoding actin 1, Rho GTPase-activating protein with PAK-box, TPX2 protein, bHLH transcription factor, and leucine-rich repeat receptor-like protein kinase were identified. Correlation networks revealed considerable interaction among the hub genes, transcription factors, and other genes.

1. Introduction

Cotton (Gossypium spp.) is among the most important crops worldwide and is grown for its natural fiber for the textile industry. The global area planted with Gossypium hirsutum (upland cotton) comprises 95% of all planted cotton because of its long, strong, and white fibers and high yield [1]. With the dramatic changes in climate and environment and the reason of the often crosspollinated plant, breeders are focused on the development of new cultivars to improve the fiber quality of upland cotton, which will be highly beneficial for the textile industry.
A thorough understanding of the underlying mechanisms of cotton fiber development is crucial to improve fiber quality. Fiber development can be classified into five stages: Fiber initiation (0–3 days post-anthesis (DPA)), elongation (3–15 DPA), transitional cell-wall remodeling (15–20 DPA), secondary wall biosynthesis (20–40 DPA), and maturity (40–50 DPA) [2]. The first two stages determine the number and length of the fibers. The third and fourth stages are associated with fiber strength and fineness as a result of the development of secondary cell wall thickening.
Comparative transcriptome sequencing is an efficient technology useful for differential expression analysis to identify candidate genes associated with target traits. Cotton fibers can be sampled from cotton bolls at each development stage. RNA sequencing (RNA-seq) provides a means of analyzing individual gene transcription and the entire transcriptome profile throughout fiber development. In the past decade, numerous comparative transcriptome analyses of upland cotton utilized chromosome segment substitution lines derived from G. hirsutum × Gossypium barbadense [3,4,5], wild and domesticated G. hirsutum [1,6], and near-isogenic lines [7,8], which discovered many candidate genes associated with high-quality fiber. However, there are few studies of gene expression patterns using a recombinant inbred line (RIL) population in which the progeny lines are approximately homozygous. In our previous study, 16 stable quantitative trait loci (QTLs) for fiber strength were identified from a RIL population consisting of 196 lines [9], in which certain individuals showed exceptionally high-quality fiber strength. To explore and identify genes responsible for the difference in fiber quality among the progeny and to compare gene expression profiles between high- and low-quality fiber lines, RNA-seq analysis was performed in the present study using two RILs that differed in fiber quality.

2. Methods

2.1. Plant Materials

Two upland cotton lines designated 69307 (L1) and 69362 (L2) from an intraspecific F6:8 RIL population were used. The RIL population was developed from a cross between the strains ‘0-153’ and ‘sGK9708’, which are characterized by excellent fiber quality and high yield, respectively. The procedure for development of the RIL population was detailed previously [9,10,11]. The high-quality fiber line L1 is a positive extreme-parent line, whereas L2 is a negative extreme-parent line for fiber quality [12].
The two lines were used as plant materials and grown under standard field conditions at the Experimental Station, Institute of Cotton Research of the Chinese Academy of Agricultural Sciences located in Anyang, Henan, China in 2015. The first day of flowering was recorded as 0 DPA and flower buds were tagged. At 10, 15, 20, 25, and 30 DPA cotton fibers were collected from bolls, with three biological repeats, using a sterile medical scalpel and were frozen in liquid nitrogen for RNA-seq analysis. The corresponding samples for the L1 and L2 lines were designated L1D10, L1D15, L1D20, L1D25, L1D30, L2D10, L2D15, L2D20, L2D25, and L2D30, where D is days post-anthesis. RNA-Seq raw data accession number SRP172799 in NCBI sequence read archive (http://www.ncbi.nlm.nih.gov/sra/) and bioproject archive (http://www.ncbi.nlm.nih.gov/bioproject/) PRJNA508480.

2.2. RNA Isolation, Library Construction, and RNA-Seq Analysis

Total RNA was extracted from each fiber sample using the RNAprep Pure Plant Kit (Polysaccharides & Polyphenolics-rich, Tiangen, Beijing, China) following the manufacturer’s protocol. The RNA was quantified using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). RNA degradation and contamination were assessed using 1% agarose gel electrophoresis. The RNA integrity was confirmed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, California, USA). A total amount of 2 μg RNA for each sample was used for transcriptome library sequencing. Libraries were constructed using the Illumina TruSeq™ RNA Sample Preparation Kit (Illumina, San Diego, CA, USA) following the manufacturer’s recommendations. A total of 30 libraries for the two cotton lines at five developmental stages (10, 15, 20, 25, and 30 DPA) with three biological repeats were sequenced using an Illumina HiSeq 2500 sequencing platform with 125 base pair (bp) paired-end (PE) raw reads (Berry Genomics Co., Ltd., Beijing, China).
Raw data in Fastq format were processed using Trimmomatic software [13]. Clean data were obtained by removing reads that contained the adapter, poly-N, and low-quality reads, which were reads with ≥10% unidentified nucleotides (N), >50% bases with Phred quality < 5, >10 nt aligned to the adapter, and 0 nt aligned to the adapter. Finally, the percentage GC content and Q30 were determined for the clean data, which were then used for downstream analysis. Bowtie v2.0.6 [14] was used to build an index of the reference genome, which was downloaded from the CottonGen database (http://www.cottongen.org) [15]. The clean data were mapped to the G. hirsutum genomes [16,17] with default parameters using TopHat2 [18]. The fragments per kilobase of exon per million reads (FPKM) values of genes were calculated using Cufflinks [19]. Pearson correlation analysis was performed to calculate correlation coefficients between samples. Samples with correlation coefficients less than 0.8 among the three biological replicates were removed from the data set.

2.3. Differentially Expressed Genes Analysis and Comparison of Differentially Expressed Genes Expression Patterns

Differentially expressed genes (DEGs) were identified using the DESeq2 R package using the count number of each gene [20,21]. The DEGs were identified with an expression level FPKM > 0.5, false discovery rate < 0.01, and log2 fold change value > 1 or <−1 between each pairwise comparison.
Given that different gene expression patterns might result in different phenotypes, the Short Time-series Expression Miner (STEM) software was used to explore the expression patterns of DEGs during fiber development [22,23]. To understand function of the DEGs, the gene ontology (GO) enrichment was performed using BLASTX program [24] and GO databases (http://archive.geneontology.org/latest-lite/) and (ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/); and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was carried out using KOBAS 3.0 software [25,26].

2.4. Construction of Gene Co-Expression Networks and Screening of Hub Genes

The WGCNA (weighted gene co-expression network analysis) R package [27] was used to analyze co-expression of genes and select highly correlated (or hub) genes that may be strongly associated with fiber quality. The genes clustered in modules that contained a coefficient (>0.6) with each sample, especially the elite fiber tissues, were used to further analysis. After completion of the co-expression analysis, the edge files were sorted by weight value and the first 200 pairs of network connections were used to establish interaction networks among the DEGs. The hub genes were screened on the basis of module membership (KME) values. The interaction networks among were drawn using Cytoscape 3.6.1 [28].

3. Results

3.1. Transcriptome Sequencing Analysis and Correlation of Replicate Samples

To identify genes that play crucial roles in fiber development, fibers were sampled at five developmental stages (10, 15, 20, 25, and 30 DPA) and transcriptome sequencing was conducted. A total of 1543.602 million clean reads were obtained from 30 libraries, with an average of 51.45 million reads per sample. In the reads mapping analysis, 91.56% to 95.98% reads could be mapped to the reference genome. The Q30 was 89.45% to 94.65%, with a mean value of 93.45%, and the GC content range was 44.22% to 47.98%, which indicated that the RNA-seq data were reliable (Figure S1a,b, Table S1). To evaluate the distribution density of clean data reads, reads were mapped to the 26 chromosomes of G. hirsutum using Bowtie 1 and visualized using the Circos program [29] (Figure S1c,d), which indicated the difference of levels of transcripts among the fiber development stages.
After reads mapping and FPKM calling, the correlation between biological replicates of samples was assessed (Figure S2, Table S2). In this analysis, genes with FPKM > 0.5 were considered to be expressed genes. Consequently, 41,583, 40,429, 40,236, 39,544, and 39,032 genes were expressed at 10, 15, 20, 25, and 30 DPA in the L1 fiber tissues, respectively. Similarly, 41,711, 40,463, 40,532, 40,929, and 39,561 expressed genes were detected at the five respective stages in L2 fiber tissues. Genes with FPKM values of 0.5–5, 5–100, and >100 accounted for 45.4%, 49.7%, and 4.9% of all gene models, respectively (Figure 1).

3.2. Differential Gene Expression Analysis

To explore the genes associated with fiber quality in fiber tissues, the DEGs were identified in vertical and horizontal comparisons, i.e. between the same developmental stage of the two lines and between different developmental stages of each line. The numbers of DEGs detected are shown in Figure 2. After removal of duplicate genes, a total of 13,881 DEGs were identified and their function descriptions annotated in vertical (Table S3) and horizontal (Table S4) comparisons, respectively. The FPKM values of the 13,881 DEGs in each sample are shown in Table S5.
At 10 DPA, 99 genes were significantly differentially expressed, including 72 upregulated and 27 downregulated genes. Representative DEGs that not only showed high |log2(FC)| between L1 and L2, but also significantly (corrected p-value < 0.01)) enriched in the KEGG pathways are listed in Table 1 and Table 2, respectively. In upregulated genes, two brassinosteroid biosynthesis pathway genes encoding cytochrome P450 proteins exhibited a five-fold change in expression in the high-quality fiber line L1 compared with that of L2, which may enhance plant performance and productivity [30]. A gene of the auxin-responsive GH3 protein family, which participates in auxin homeostasis by catalyzing auxin conjugation and binding free indole-3-acetic acid to amino acids [31], showed an approximately three-fold increase in expression in L1. Additional genes, such as genes encoding a glycosylic hydrolase protein, a 10-formyltetrahydrofolate synthetase, and a UDP-glucose 6-dehydrogenase protein, also showed several-fold higher expression levels in L1 compared with those in L2. In the downregulated genes, only six genes were enriched into four pathways listed in Table 2. Interestingly, the expression levels of two transcription factors encoding WRKY11, which has been characterized to be involved in the response to dehydration stress [32] and plays key role in plant defense [33], and C3H proteins in L1D10 were four times than that of L2D10.
At 15 DPA, 1137 and 1584 genes were significantly upregulated and downregulated, respectively, in fiber tissue of the high-quality L1 line compared with their expression levels in the L2 line. A total of 830 DEGs were detected between the L1D20 and L2D20 fiber samples, of which 437 genes were upregulated and 393 genes were downregulated in the L1. Representative upregulated and downregulated genes are listed in Table 3 and Table 4, respectively. The expression levels of two acyl-CoA oxidase genes and a beta-ketoacyl reductase 1 gene involved in the biosynthesis pathway for unsaturated fatty acids exhibited three-fold changes in expression in the L1D20 sample compared with their expression levels in the L2D20 sample. Seven genes associated with microtubule development in the phagosome pathway showed 2.4- to 8.9-fold changes in expression level in L1 compared with those of L2 at 20 DPA. Genes encoding ketoacyl-CoA synthase, 3-ketoacyl-acyl carrier protein synthase I, and acetyl-CoA carboxylase carboxyl transferase subunit beta in the fatty acid elongation pathway, which are essential for membrane biosynthesis [34], were also enriched. In addition, transcription factors, such as basic helix-loop-helix (bHLH), bZIP and CCCH-type zinc finger protein, MYB-like Ap2/B3 protein, and WRKY protein, showed their highest expression level in the L1D20 fiber sample. In the downregulated genes, the expression of four genes encoding cytokinin oxidase in L2D20 exhibited 3.2~4.6-fold changes compared with those of L1D20, in which the high expression of cytokinin oxidase may lead to the degradation of cytokinin, thus reducing seed yield [35]. In addition, transcription factors, such as GRAS that might promote the stress-resistant [36] bZIP, myb-like .etc were also expressed higher in L2D20 than that of L1D20.
After the transitional cell-wall remodeling stage, 1019 and 184 genes were significantly upregulated and downregulated at 25 DPA, respectively. At the fiber maturity stage, 4034 genes were significantly differentially expressed in pairwise comparisons between L1 and L2 at 30 DPA, of which 2555 and 1479 genes were upregulated and downregulated at 30 DPA, respectively.

3.3. Congruence Analysis with the Previous Report

In a previous study, Zhang et al. [9] identified 16 stable QTLs for fiber strength using the RILs of present study, in which 3364 genes were identified as candidate genes by comparison with the reference G. hirsutum genome. To identify genes that were not only associated with fiber strength in the entire RILs population, but also were differentially expressed in progeny of the high-quality and poor-quality fiber lines, we compared these candidate genes and the DEGs between L1 and L2. Of the 3364 genes in total, 363 DEGs comprising 4, 75, 39, 62, and 183 genes at 10, 15, 20, 25, and 30 DPA were detected, of which 228 (62.6%) genes were upregulated in L1. These genes that were not only located in QTLs related to FS trait, but also identified to be differentially expressed in the two lines might contribute to the phenotypic differences in FS. In these genes, the gene Gh_A07G1758 encoding RAB GTPase homolog B1C is required for starch and sugar homeostasis in A. thaliana. Meanwhile, Gh_A07G1758 was also detected to be associated with fiber FL and FS by GWAS method [37,38]. The genes Gh_D06G1939 encoding WRKY DNA-binding protein could enhance the drought and salt stress tolerance of transgenic Nicotiana benthamiana [38]. Another gene, Gh_D06G1944, encodes a putative leucine-rich repeat receptor protein kinase, which is characterized to control pollen production of the Gossypium anther [39]. The functional annotation of these genes in Arabidopsis is shown in Table S6.

3.4. Temporal Gene Expression Patterns

To explore temporal differences in gene expression profiles in the two lines, STEM analysis of the DEGs was performed. A total of 8506 and 8415 genes were identified and organized into seven expression profiles (Figure 3a) with e-values less than 0.001 in L1 and L2. Each profile contained a set of genes that showed similar expression patterns at all fiber development stages (Table S7). The largest two gene sets were profile ID 4 and 22, which represented a continuous upregulated and downregulated trend, respectively, as fiber development progressed. The expression profile 22 contained 2308 genes in L1, which was 19.7% more than the number of genes for L2 (1928 genes). Comparison of the genes of the two profiles between L1 and L2 was visualized by means of a Venn diagram (Figure 3b,c), which showed that approximately half of the genes were different in the expression pattern 4, even though the number of genes was similar. A total of 1112 and 1060 genes were specific to L1 and L2, respectively, for expression profile 4, with 1298 genes in common. Similarly, 1180 genes were common to L1 and L2 for expression profile 22, with 1128 and 748 genes respectively unique to each line.
These differences in gene expression may contribute to phenotypic variation. Hence, we performed GO enrichment analysis of the genes in profiles 4 and 22 for each of L1 and L2. The 30 most frequent GO terms are shown in Figure 4 and Figure 5. The terms were categorized into biological process, cellular component, and molecular function, which showed differences between the two RIL lines. In L1_profile 4, genes were annotated with the GO terms “regulation of transcription, DNA-templated” (112 genes, 10.1%), “transcription, DNA-templated” (94 genes, 8.5%), “protein phosphorylation” (92 genes, 8.3%), and “response to abscisic acid” (91 genes, 8.2%), which differed from that of L2_profile 4. In the molecular function category, the GO terms “transcription factor activity, sequence-specific DNA binding” (98 genes in L1, 8.8%) and “protein serine/threonine kinase activity” (91 genes in L1, 8.2%) showed difference between the two lines (Figure 4). The GO terms for the L1_profile 22 were mainly of the cellular component category, differing from the L2_profile 22, which was manifested as “endosome” (113 genes, 10%), “trans-Golgi network” (110 genes, 9.8%), “Golgi membrane” (98 genes, 8.7%), and “vacuolar membrane” (77 genes, 6.8%) (Figure 5).
To investigate differences in the gene pathways involved in the profiles 22 and 4, KEGG enrichment analysis was performed under the background of Arabidopsis thaliana and pathways are shown in a bubble graph (Figure 3d,e). The genes in expression profile 22 were mainly enriched in metabolic pathways and biosynthesis of secondary metabolites, followed by plant hormone signal transduction, phenylpropanoid biosynthesis, and cysteine and methionine metabolism. In contrast, in expression profile 4, in addition to enrichment in metabolic pathways and biosynthesis of secondary metabolites, genes were enriched in ribosome, phagosome, oxidative phosphorylation, and carbon metabolism pathways.
Other expression profiles, such as profiles 23 and 10, showed specific expression patterns at 20 DPA and 25–30 DPA, which may indicate that the expressed genes were associated with cell-wall remodeling and secondary cell-wall synthesis. The enriched KEGG pathways for these profiles are shown in Table S8. The genes enriched in profile 23 were annotated with the terms “amino sugar and nucleotide sugar metabolism”, including 10 genes that encode galacturonosyltransferases, and “starch and sucrose metabolism” consisting of sucrose synthases (Gh_D07G0139 and Gh_D06G0832), which are expressed in response to cotton fiber growth and development [40], and UDP-D-glucuronate 4-epimerase (Gh_D07G1467), which plays a role in providing UDP-D-galacturonate to glycosyltransferases in pectic polysaccharide synthesis [41]. The enriched pathways “plant hormone signal transduction”, including a GRAS family transcription factor that is valuable for stress-resistance breeding in G. hirsutum [36], “cutin, suberin, and wax biosynthesis” (7 genes), “metabolic pathways” (121 genes) were identified by genes of expression profile 10, which indicated that these regulatory pathways played key roles in secondary cell wall growth of fiber tissues (Table S8).
In addition, the functional differences of genes unique to L1 and L2 in the same expression trend of profile 10 and 23, respectively, were also performed using the GO enrichment analysis. In profile 10, there are 148 common genes with 628 and 571 genes were unique to L1 and L2, respectively (Figure S3a). The GO terms of the unique genes were shown in Figure S3b,c, which showed the differences in terms of “oxidation-reduction process” (40 genes, 6.37%), “response to water deprivation” (38 genes, 6%), “defense response to bacterium” (32 genes, 5.1%) in biological process category of L1. Three terms in cellular component such as “Golgi apparatus” (45 genes, 7.17%), “endoplasmic reticulum” (38 genes, 6.05%) and “chloroplast stroma” (31 genes, 4.94%) in L1 were enriched. Totally 189 common genes showed the same expression profile 23 with 723 and 689 genes unique to L1 and L2, respectively (Figure S4a). The top 30 enrichment terms of these genes were showed in Figure 4b,c. For example, terms “response to salt stress” (60 genes, 8.3%), “transcription, DNA-templated” (48 genes, 4.64%) and “oxidation-reduction process” (43 genes, 5.95%) in category of biological process were different between L1 and L2.

3.5. Gene Co-Expression Network Analysis

To gain further insight into the relationship between gene expression and progressive fiber development and to identify genes that were specifically associated with fiber quality, the co-expression networks for a total of 13,881 differentially expressed genes (DEGs) was generated using weighted gene co-expression network analysis (WGCNA).
The topology overlap matrix was built using the hierarchical clustering method and the dynamic cut module that characterized similar expression patterns were merged. In total, 13 gene modules were identified based on genes expression profiles and lines, of which five modules (indicated in red in Figure 6) were highly associated with an individual developmental stage of both L1 and L2 or were specifically associated with certain developmental stage of the high-quality fiber line on the basis of their coefficient between modules and tissues (Figure 6, Table S9).
During the fiber elongation stage, the dark red module contained 1370 genes, including 84 transcription factors associated with the sample at 10 DPA in both L1 and L2. A total of 657 genes (48 transcription factors) were clustered in the dark turquoise module, which was specific to the high-quality fiber L1 line at 15 DPA. The violet module (178 genes) was significantly specific to the L1 phenotype at 20 DPA, which was the period of transitional cell-wall remodeling or synthesis of the winding layer [42]. The pink module harbored 443 genes that were not only highly associated with L1 at 30 DPA, but also associated with L1 at 20 DPA, which indicated that these genes were involved in both transitional cell-wall remodeling and fiber maturity. The dark grey module contained 1593 genes and was also significantly associated with L1 at 30 DPA. Eighty-four of the 1593 genes were transcription factors, which suggested that these genes played key roles in the shaping of fiber quality.
A KEGG pathway analysis of the genes in each module was performed. The expression patterns of representative genes in key pathways and the FPKM values are shown in Figure 7 and Table S10. Genes in pathways for ubiquinone and other terpenoid-quinone biosynthesis, fatty acid metabolism, and phenylpropanoid biosynthesis were highly expressed at 10 DPA in L1 and L2. Four genes, namely Gh_D05G0184, Gh_A05G0122, Gh_A02G1611, and Gh_D06G1792, were annotated with the brassinosteroid biosynthesis pathway, which is required for fiber initiation as well as fiber elongation [43]. The genes in the dark turquoise module were mainly associated with photosynthesis, photosynthesis-antenna proteins, and galactose metabolism, which indicated their involvement in the accumulation of organic materials. In addition, genes in the dark grey module were associated with biosynthesis of secondary metabolites, cutin, suberin and wax biosynthesis, and fatty acid metabolism.

3.6. Identification of Hub Genes for Elite Fiber Quality and Visualization of Correlation Networks

In WGCNA, KME values indicate the eigengene connectivity, based on the assumption that hub genes are identified by sorting KME values. In the present study, the several top-ranked genes with the highest KME values in each of the five modules were selected as hub genes and are listed in Table 5. The hub genes in the dark red module encode actin proteins, glycosyl hydrolase proteins, and Rho GTPase-activating proteins with a PAK-box, which may significantly affect fiber elongation. Hub genes highly associated with L1 at 15 DPA were annotated as targeting protein for the Xklp2 (TPX2) protein family, and calcium-dependent phosphotriesterase superfamily protein.
Given that 10 DPA and 20 DPA were the critical fiber developmental stages of primary cell wall (PCW) and the transition to secondary cell wall (SCW) growth [1,2,44], respectively, the correlation networks of the high-weight values pairs were visualized using Cytoscape software. The candidate hub genes in the dark red module at 10 DPA were highlighted in red in the networks (Figure 8a); similarly, the hub genes in the violet module that were highly associated with 20 DPA of the high-quality fiber line (L1) are shown in Figure 8b. The gene Gh_D10G0295 is both a hub gene and a bHLH DNA-binding protein 30 [45].

4. Discussion

4.1. Transcriptome Sequencing of Elite and Poor-Quality Fiber Lines Mapped to the Reference Genome of Upland Cotton

RNA-seq could be a useful tool to discover whole-genome expression profile and DEGs for screening candidate genes related to fiber development. In the past decade, application of RNA-seq were reported in fiber development [1,3,4,7,57], in biotic stress [58,59,60], and in abiotic stress [61,62]. In the present study, five time points (10, 15, 20, 25, and 30 DPA) were chosen to investigate the genetic difference in fiber development between the progenies of two extreme-parent RILs. The 10 and 15 DPA time points coincided with critical stage for fiber elongation, which may continue up to 25 DPA [2,63]. Cotton fiber tissues start transitional cell wall remodeling at 20 DPA, which represents the onset of the secondary wall synthesis [42]. The 25 DPA and 30 DPA time points represented the period of secondary wall thickening, which affects fiber strength and fineness [44]. Therefore, the DEGs identified between the two lines during fiber development, and mapped to the published reference genome and annotated transcripts for G. hirsutum [16,17], may reflect the genetic differences associated with fiber quality. The average number of clean reads and Q30 were 51.45 million and 93.8% per sample, respectively, which were indicators of the reliable quality of the RNA-seq data. However, five samples in the third biological replicate showed low correlation (<0.8), which may be a result of weather conditions or another unknown reason.

4.2. RNA-Seq Provides Potentia information by Comparison of Expression Profiles between the Two Recombinant Inbred Lines

Vertical comparison of DEGs showed that a large number of genes were significantly expressed in L1 and L2 at each stage of fiber development, which revealed that gene expression patterns differed considerably between the two lines. DEGs analysis of lines showing different phenotypes can uncover functional genes related to target traits. Previous studies for exploring fiber quality and yield and resistance to Verticillium wilt were also performed using RNA-seq. Islam et al. performed RNA-seq of fiber tissues on 15 and 20 DPA, which showed that two signaling pathways including ethylene and the interconnected phytohormonal pathways and receptor-like kinases (RLKs) signaling pathways might contribute to the high fiber strength [7]. RNA-seq focusing on fiber length was also performed using the backcross inbred lines on 5, 10, 15, and 20 DPA, which identified that eight DEGs were located in the FL QTLs and that three SNP markers were associated with FL trait [3]. A SKIP35 protein involved in ubiquitin-mediated signal was identified using RNA-seq and validated to enhance the tolerance to Verticillium wilt [58].
In the present study, upregulated and downregulated genes between two lines were identified in each stage, and representative DEGs were partly characterized in A. thaliana and Gossypium. The functions of these genes that played key roles during plant development in A. thaliana still needed to be solved in Gossypium. For instances, an auxin-responsive GH3 family protein enriched in plant hormone signal transduction pathway participates in auxin homeostasis [31]. A WRKY transcription factor plays roles in dehydration stress and plant defense in upland cotton [32]. Interestingly, at 10 DPA, the genes Gh_D05G0184 and Gh_A05G0122 were significantly upregulated in L1. These genes are involved in the brassinosteroid pathway and encode cytochrome P450 proteins, playing roles in catabolizing active BRs during fiber elongation by modulating ethylene and cadmium signaling [64]. The tubulin proteins enriched in the “phagosome” pathway were highly upregulated in L1 compared with the expression level in L2 at 20 DPA (Table 3), and are involved in cotton fiber development [65]. In each stage, part of DEGs were colocalized in stable FS QTLs, meanwhile, several genes listed in Section 3.3 could be identified in previous study [37,38], which suggests that these DEGs are more valuable for influencing fiber strength.
The similar expression trends but different gene sets (Table S7) and further GO analysis of genes unique to L1 and L2 in expression profile 4, 22, 10, and 23, as shown in Figure 4 and Figure 5, revealed the difference between the two lines, which contribute to the difference in FS. It is noteworthy that genes clustered in expression profile 23 and 10 were specific expressed in secondary cell wall thickening and cell wall remodeling. For instance, genes participate in “amino sugar and nucleotide sugar metabolism” pathway can affect the cotton fiber growth and development [40], as detailed in Section 3.4. The specific function of these genes in Gossypium remains to be studied.

4.3. Hub Genes Identified by WGCNA May Play Critical Roles in Improvement of Fiber Quality

In the present study, we used WGCNA to identify significant modules of genes associated with specific developmental stages, especially for the high-quality fiber line L1, and to select hub genes. Five modules were highly associated with high-quality fiber, hence further analysis of these modules was performed in the present study, although several modules also showed high correlation with samples of the poor-quality fiber line L2. Transcription factors play key roles in plant growth and development. For instance, the transcription factors in elongation stage at 10 DPA contained 11 HD-ZIP that play an important role in epidermal development [66], nine bHLH that are involved in brassinosteroid (BR) signaling in fiber development of upland cotton [67], six ERF responded to salt/drought and ABA [68], and each five of MYB, C2H2, C3H, and ARF (Table S9). On 30 DPA, the transcription factors of TALE, MYB, GRAS, NAC, and Trihelix GATA were specific expressed in fiber cell development. Several KEGG pathways were highly specific to fiber developmental stages such as “ubiquinone and other terpenoid-quinone biosynthesis” at 10 DPA, “brasssinosteroid biosynthesis” and “plant hormone signal transduction at 15 DPA of L1, photosynthesis at 20 DPA of L1 .etc played are important factors in affecting fiber development. The functions of hub genes were partly characterized in A. thaliana (Table 5). A TPX2 protein was identified as hub gene. In A. thaliana, TPX2 protein performs multiple roles in microtubule organization such as reinforcing microtubule formation in the vicinity of chromatin and the nuclear envelope [69]. A WDL3 protein belonging to TPX2 family was in response to light to modulate hypocotyl cell elongation [47]. The interactions between hub genes and transcription factors could be a kind of transcriptional regulation. An actin 1 gene (Gh_A10G1961) was predicted to interact with five transcription factors (2 SBP, NAC, GRF and GATA). Therefore, these hub genes were also worthy of further research.

5. Conclusions

In summary, to understand the differences in gene expression patterns between two extreme-parent lines derived from an intra-hirsutum RIL population of upland cotton and to identify genes associated with elite fiber quality, RNA-seq was performed by constructing 30 libraries from the two lines with three biological replications at five time points that coincided with the stages of fiber initiation, elongation, cell-wall remodeling, and secondary cell wall growth. Hence, the DEGs were identified by multiple comparisons, especially the vertical comparison between L1 and L2, and the representative genes showed key functions in fiber elongation and thickening. The 363 genes identified by RNA-seq and colocalized in FS QTL were valuable for affecting fiber development. Gene ontology analysis underlined the difference in gene function categories between two key expression patterns (profiles 4 and 22; continuous upregulation and downregulation, respectively), and another two expression profiles (profiles 10 and 23) that were highly associated with fiber cell wall remodeling and thickening. KEGG analysis revealed that the genes that showed contrasting expression trends were involved in different pathways, such as plant hormone signal transduction, glycerolipid metabolism, and phenylpropanoid biosynthesis. The genes enriched in expression profiles 23 and 10 were involved in starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, and galactose metabolism, and thus may also contribute to improvement of fiber quality. Co-expression network analysis identified five modules that were highly associated with developmental stages of elite-quality fiber. These findings have revealed promising candidate genes for improvement of fiber quality in breeding of upland cotton.

Supplementary Materials

The following are available online at https://www.mdpi.com/2073-4425/10/2/119/s1. Figure S1. Quality statistics for the filtered reads, Figure S2. Pearson correlation coefficients among the 30 samples. R1, R2, and R3 represent the three biological replications, Figure S3. Further differences analysis to expression profile 10 between the two lines, Figure S4: Further differences analysis to expression profile 23 between the two lines.

Author Contributions

H.S. and Y.Y. designed the experiments and provided the resources for all experiments. X.Z. and A.L. coordinated the experiments, data analysis, and wrote the manuscript. Z.Z., Q.G. and S.F. planted the material and collected the fiber. W.G., W.L. and J.G. extracted the total RNA. Y.S., B.T. and Y.W. performed the data quality control and reads mapping. R.L., K.L. and Q.Z. performed the DEGs analysis and congruence analysis. X.J. and Y.F. analyzed the expression patterns. S.Z. and T.J. performed gene annotation. L.Z. performed WGCNA.

Funding

This work was funded by the China National Science Foundation Innovative Engineering (1610162016004, 31471538); and the central level of the scientific research institutes for basic R & D special fund business (161012018017).

Acknowledgments

We thank Robert McKenzie, PhD, from Liwen Bianji, Edanz Group China (www.liwenbianji.cn/ac), for editing the English text of a draft of this manuscript. The first author would like to thank his girlfriend, JieLi, for accompanying him all the way, past and future.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Yoo, M.-J.; Wendel, J.F. Comparative evolutionary and developmental dynamics of the cotton (Gossypium hirsutum) fiber transcriptome. PLoS Genet. 2014, 10, e1004073. [Google Scholar] [CrossRef] [PubMed]
  2. Haigler, C.H.; Betancur, L.; Stiff, M.R.; Tuttle, J.R. Cotton fiber: A powerful single-cell model for cell wall and cellulose research. Front. Plant Sci. 2012, 3, 104. [Google Scholar] [CrossRef] [PubMed]
  3. Li, X.; Wu, M.; Liu, G.; Pei, W.; Zhai, H.; Yu, J.; Zhang, J.; Yu, S. Identification of candidate genes for fiber length quantitative trait loci through RNA-Seq and linkage and physical mapping in cotton. BMC Genom. 2017, 18, 427. [Google Scholar] [CrossRef] [PubMed]
  4. Li, P.-T.; Wang, M.; Lu, Q.-W.; Ge, Q.; Rashid, M.H.O.; Liu, A.-Y.; Gong, J.-W.; Shang, H.-H.; Gong, W.-K.; Li, J.-W.; et al. Comparative transcriptome analysis of cotton fiber development of Upland cotton (Gossypium hirsutum) and Chromosome Segment Substitution Lines from G. hirsutum × G. barbadense. BMC Genom. 2017, 18, 705. [Google Scholar] [CrossRef] [PubMed]
  5. Lu, Q.; Shi, Y.; Xiao, X.; Li, P.; Gong, J.; Gong, W.; Liu, A.; Shang, H.; Li, J.; Ge, Q.; et al. Transcriptome analysis suggests that chromosome introgression fragments from sea island cotton (Gossypium barbadense) increase fiber strength in upland cotton (Gossypium hirsutum). G3 Genes|Genomes|Genet. 2017, 7, 3469–3479. [Google Scholar] [CrossRef] [PubMed]
  6. Applequist, W.L.; Cronn, R.; Wendel, J.F. Comparative development of fiber in wild and cultivated cotton. Evol. Dev. 2001, 3, 3–17. [Google Scholar] [CrossRef] [PubMed]
  7. Islam, M.S.; Fang, D.D.; Thyssen, G.N.; Delhom, C.D.; Liu, Y.; Kim, H.J. Comparative fiber property and transcriptome analyses reveal key genes potentially related to high fiber strength in cotton (Gossypium hirsutum L.) line MD52ne. BMC Plant Biol. 2016, 16, 36. [Google Scholar] [CrossRef] [PubMed]
  8. Gilbert, M.K.; Kim, H.J.; Tang, Y.; Naoumkina, M.; Fang, D.D. Comparative Transcriptome Analysis of Short Fiber Mutants Ligon-Lintless 1 And 2 Reveals Common Mechanisms Pertinent to Fiber Elongation in Cotton (Gossypium hirsutum L.). PLoS ONE 2014, 9, e95554. [Google Scholar] [CrossRef]
  9. Zhang, Z.; Ge, Q.; Liu, A.; Li, J.; Gong, J.; Shang, H.; Shi, Y.; Chen, T.; Wang, Y.; Palanga, K.K.; et al. Construction of a High-Density Genetic Map and Its Application to QTL Identification for Fiber Strength in Upland Cotton. Crop Sci. 2017, 57, 774. [Google Scholar] [CrossRef]
  10. Sun, F.D.; Zhang, J.H.; Wang, S.F.; Gong, W.K.; Shi, Y.Z.; Liu, A.Y.; Li, J.W.; Gong, J.W.; Shang, H.H.; Yuan, Y.L. QTL mapping for fiber quality traits across multiple generations and environments in upland cotton. Mol. Breed. 2012, 30, 569–582. [Google Scholar] [CrossRef]
  11. Zhang, Z.; Shang, H.; Shi, Y.; Huang, L.; Li, J.; Ge, Q.; Gong, J.; Liu, A.; Chen, T.; Wang, D. Construction of a high-density genetic map by specific locus amplified fragment sequencing (SLAF-seq) and its application to Quantitative Trait Loci (QTL) analysis for boll weight in upland cotton (Gossypium hirsutum.). BMC Plant Biol. 2016, 16, 79. [Google Scholar] [CrossRef] [PubMed]
  12. Zhang, Z.; Li, J.; Muhammad, J.; Cai, J.; Jia, F.; Shi, Y.; Gong, J.; Shang, H.; Liu, A.; Chen, T.; et al. High Resolution Consensus Mapping of Quantitative Trait Loci for Fiber Strength, Length and Micronaire on Chromosome 25 of the Upland Cotton (Gossypium hirsutum L.). PLoS ONE 2015, 10, e0135430. [Google Scholar] [CrossRef] [PubMed]
  13. Bolger, A.M.; Lohse, M.; Usadel, B. Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics 2014, 30, 2114–2120. [Google Scholar] [CrossRef] [PubMed]
  14. Langmead, B.; Salzberg, S.L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 2012, 9, 357–359. [Google Scholar] [CrossRef] [PubMed]
  15. Yu, J.; Jung, S.; Cheng, C.H.; Ficklin, S.P.; Lee, T.; Zheng, P.; Jones, D.; Percy, R.G.; Main, D. CottonGen: A genomics, genetics and breeding database for cotton research. Nucleic Acids Res. 2014, 42, 1229–1236. [Google Scholar] [CrossRef]
  16. Zhang, T.; Hu, Y.; Jiang, W.; Fang, L.; Guan, X.; Chen, J.; Zhang, J.; Saski, C.A.; Scheffler, B.E.; Stelly, D.M.; et al. Sequencing of allotetraploid cotton (Gossypium hirsutum L. acc. TM-1) provides a resource for fiber improvement. Nat. Biotechnol. 2015, 33, 531. [Google Scholar] [CrossRef] [PubMed]
  17. Li, F.; Fan, G.; Lu, C.; Xiao, G.; Zou, C.; Kohel, R.J.; Ma, Z.; Shang, H.; Ma, X.; Wu, J.; et al. Genome sequence of cultivated Upland cotton (Gossypium hirsutum TM-1) provides insights into genome evolution. Nat. Biotechnol. 2015, 33, 524–530. [Google Scholar] [CrossRef]
  18. Kim, D.; Pertea, G.; Trapnell, C.; Pimentel, H.; Kelley, R.; Salzberg, S.L. TopHat2: Accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 2013, 14, R36. [Google Scholar] [CrossRef]
  19. Trapnell, C.; Williams, B.A.; Pertea, G.; Mortazavi, A.; Kwan, G.; van Baren, M.J.; Salzberg, S.L.; Wold, B.J.; Pachter, L. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 2010, 28, 511. [Google Scholar] [CrossRef]
  20. Anders, S.; Huber, W. Differential expression analysis for sequence count data. Genome Biol. 2010, 11, R106. [Google Scholar] [CrossRef]
  21. Love, M.I.; Huber, W.; Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014, 15, 550. [Google Scholar] [CrossRef] [PubMed]
  22. Ernst, J.; Nau, G.J.; Bar-Joseph, Z. Clustering Short Time Series Gene Expression Data; Oxford University Press: Oxford, UK, 2005; pp. 159–168. [Google Scholar]
  23. Ernst, J.; Barjoseph, Z. STEM: A tool for the analysis of short time series gene expression data. BMC Bioinform. 2006, 7, 191. [Google Scholar] [CrossRef] [PubMed]
  24. Altschul, S.F.; Gish, W.; Miller, W.; Myers, E.W.; Lipman, D.J. Basic local alignment search tool. J. Mol. Biol. 1990, 215, 403–410. [Google Scholar] [CrossRef]
  25. Xie, C.; 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, W316–W322. [Google Scholar] [CrossRef] [PubMed]
  26. Wu, J.; Mao, X.; Cai, T.; Luo, J.; Wei, L. KOBAS server: A web-based platform for automated annotation and pathway identification. Nucleic Acids Res. 2006, 34, W720–W724. [Google Scholar] [CrossRef] [PubMed]
  27. Langfelder, P.; Horvath, S. WGCNA: An R package for weighted correlation network analysis. BMC Bioinform. 2008, 9, 559. [Google Scholar] [CrossRef] [PubMed]
  28. Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N.S.; Wang, J.T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13, 2498–2504. [Google Scholar] [CrossRef]
  29. Krzywinski, M.; Schein, J.; Birol, I.; Connors, J.; Gascoyne, R.; Horsman, D.; Jones, S.J.; Marra, M.A. Circos: An information aesthetic for comparative genomics. Genome Res. 2009, 19, 1639–1645. [Google Scholar] [CrossRef]
  30. Tong, H.; Chu, C. Functional Specificities of Brassinosteroid and Potential Utilization for Crop Improvement. Trends Plant Sci. 2018. [Google Scholar] [CrossRef]
  31. Feng, S.; Yue, R.; Tao, S.; Yang, Y.; Zhang, L.; Xu, M.; Wang, H.; Shen, C. Genome-wide identification, expression analysis of auxin-responsive GH3 family genes in maize (Zea mays L.) under abiotic stresses. J. Integr. Plant Biol. 2015, 57, 783–795. [Google Scholar] [CrossRef]
  32. Liu, H.; Yang, W.; Liu, D.; Han, Y.; Zhang, A.; Li, S. Ectopic expression of a grapevine transcription factor VvWRKY11 contributes to osmotic stress tolerance in Arabidopsis. Mol. Biol. Rep. 2011, 38, 417–427. [Google Scholar] [CrossRef] [PubMed]
  33. Journot-Catalino, N.; Somssich, I.E.; Roby, D.; Kroj, T. The transcription factors WRKY11 and WRKY17 act as negative regulators of basal resistance in Arabidopsis thaliana. Plant Cell 2006, 18, 3289–3302. [Google Scholar] [CrossRef] [PubMed]
  34. Qin, Y.M.; Hu, C.Y.; Pang, Y.; Kastaniotis, A.J.; Hiltunen, J.K.; Zhu, Y.X. Saturated very-long-chain fatty acids promote cotton fiber and Arabidopsis cell elongation by activating ethylene biosynthesis. Plant Cell 2007, 19, 3692–3704. [Google Scholar] [CrossRef] [PubMed]
  35. Bartrina, I.; Otto, E.; Strnad, M.; Werner, T.; Schmulling, T. Cytokinin regulates the activity of reproductive meristems, flower organ size, ovule formation, and thus seed yield in Arabidopsis thaliana. Plant Cell 2011, 23, 69–80. [Google Scholar] [CrossRef] [PubMed]
  36. Zhang, B.; Liu, J.; Yang, Z.E.; Chen, E.Y.; Zhang, C.J.; Zhang, X.Y.; Li, F.G. Genome-wide analysis of GRAS transcription factor gene family in Gossypium hirsutum L. BMC Genom. 2018, 19, 348. [Google Scholar] [CrossRef]
  37. Ma, Z.; He, S.; Wang, X.; Sun, J.; Zhang, Y.; Zhang, G.; Wu, L.; Li, Z.; Liu, Z.; Sun, G.; et al. Resequencing a core collection of upland cotton identifies genomic variation and loci influencing fiber quality and yield. Nat. Genet. 2018, 50, 803–813. [Google Scholar] [CrossRef]
  38. Sun, Z.; Wang, X.; Liu, Z.; Gu, Q.; Zhang, Y.; Li, Z.; Ke, H.; Yang, J.; Wu, J.; Wu, L.; et al. Genome-wide association study discovered genetic variation and candidate genes of fibre quality traits in Gossypium hirsutum L. Plant Biotechnol. J. 2017, 15, 982–996. [Google Scholar] [CrossRef]
  39. Shi, Y.L.; Guo, S.D.; Zhang, R.; Meng, Z.G.; Ren, M.Z. The role of Somatic embryogenesis receptor-like kinase 1 in controlling pollen production of the Gossypium anther. Mol. Biol. Rep. 2014, 41, 411–422. [Google Scholar] [CrossRef]
  40. Brill, E.; van Thournout, M.; White, R.G.; Llewellyn, D.; Campbell, P.M.; Engelen, S.; Ruan, Y.L.; Arioli, T.; Furbank, R.T. A novel isoform of sucrose synthase is targeted to the cell wall during secondary cell wall synthesis in cotton fiber. Plant Physiol. 2011, 157, 40–54. [Google Scholar] [CrossRef]
  41. Molhoj, M.; Verma, R.; Reiter, W.D. The biosynthesis of D-Galacturonate in plants. functional cloning and characterization of a membrane-anchored UDP-D-Glucuronate 4-epimerase from Arabidopsis. Plant Physiol. 2004, 135, 1221–1230. [Google Scholar] [CrossRef]
  42. Tuttle, J.R.; Nah, G.; Duke, M.V.; Alexander, D.C.; Guan, X.; Song, Q.; Chen, Z.J.; Scheffler, B.E.; Haigler, C.H. Metabolomic and transcriptomic insights into how cotton fiber transitions to secondary wall synthesis, represses lignification, and prolongs elongation. BMC Genom. 2015, 16, 477. [Google Scholar] [CrossRef] [PubMed]
  43. Sun, Y.; Veerabomma, S.; Abdel-Mageed, H.A.; Fokar, M.; Asami, T.; Yoshida, S.; Allen, R.D. Brassinosteroid regulates fiber development on cultured cotton ovules. Plant Cell Physiol. 2005, 46, 1384–1391. [Google Scholar] [CrossRef] [PubMed]
  44. Kim, H.J.; Triplett, B.A. Cotton fiber growth in planta and in vitro. Models for plant cell elongation and cell wall biogenesis. Plant Physiol. 2001, 127, 1361–1366. [Google Scholar] [CrossRef] [PubMed]
  45. Bailey, P.C.; Martin, C.; Toledo-Ortiz, G.; Quail, P.H.; Huq, E.; Heim, M.A.; Jakoby, M.; Werber, M.; Weisshaar, B. Update on the basic helix-loop-helix transcription factor gene family in Arabidopsis thaliana. Plant Cell 2003, 15, 2497–2502. [Google Scholar] [CrossRef] [PubMed]
  46. Schaefer, A.; Hohner, K.; Berken, A.; Wittinghofer, A. The unique plant RhoGAPs are dimeric and contain a CRIB motif required for affinity and specificity towards cognate small G proteins. Biopolymers 2011, 95, 420–433. [Google Scholar] [CrossRef] [PubMed]
  47. Liu, X.; Qin, T.; Ma, Q.; Sun, J.; Liu, Z.; Yuan, M.; Mao, T. Light-Regulated Hypocotyl Elongation Involves Proteasome-Dependent Degradation of the Microtubule Regulatory Protein WDL3 in Arabidopsis. Plant Cell 2013, 25, 1740–1755. [Google Scholar] [CrossRef] [PubMed]
  48. Sun, J.; Ma, Q.; Mao, T. Ethylene Regulates the Arabidopsis Microtubule-Associated Protein WAVE-DAMPENED2-LIKE5 in Etiolated Hypocotyl Elongation. Plant Physiol. 2015, 169, 325–337. [Google Scholar] [CrossRef]
  49. An, R.; Liu, X.; Wang, R.; Wu, H.; Liang, S.; Shao, J.; Qi, Y.; An, L.; Yu, F. The over-expression of two transcription factors, ABS5/bHLH30 and ABS7/MYB101, leads to upwardly curly leaves. PLoS ONE 2014, 9, e107637. [Google Scholar] [CrossRef]
  50. Louvet, R.; Cavel, E.; Gutierrez, L.; Guénin, S.; Roger, D.; Gillet, F.; Guerineau, F.; Pelloux, J. Comprehensive expression profiling of the pectin methylesterase gene family during silique development in Arabidopsis thaliana. Planta 2006, 224, 782–791. [Google Scholar] [CrossRef]
  51. Gonzalez, A.; Mendenhall, J.; Huo, Y.; Lloyd, A. TTG1 complex MYBs, MYB5 and TT2, control outer seed coat differentiation. Dev. Biol. 2009, 325, 412–421. [Google Scholar] [CrossRef]
  52. Li, S.F.; Milliken, O.N.; Pham, H.; Seyit, R.; Napoli, R.; Preston, J.; Koltunow, A.M.; Parish, R.W. The Arabidopsis MYB5 transcription factor regulates mucilage synthesis, seed coat development, and trichome morphogenesis. Plant Cell 2009, 21, 72–89. [Google Scholar] [CrossRef] [PubMed]
  53. Saez-Aguayo, S.; Rautengarten, C.; Temple, H.; Sanhueza, D.; Ejsmentewicz, T.; Sandoval-Ibañez, O.; Doñas-Cofré, D.A.; Parra-Rojas, J.P.; Ebert, B.; Lehner, A. UUAT1 Is a Golgi-Localized UDP-Uronic Acid Transporter that Modulates the Polysaccharide Composition of Arabidopsis Seed Mucilage. Plant Cell 2017, 29, tpc.00465.02016. [Google Scholar] [CrossRef] [PubMed]
  54. Matija, D.; Anne-Ruxandra, C.; Benoit, C.; Mary, G.; Samuel, J.P.; Murat, T. Evidence for Network Evolution in an Arabidopsis Interactome Map. Science 2011, 333, 601–607. [Google Scholar] [CrossRef]
  55. Yamaguchi, Y.; Huffaker, A.; Bryan, A.C.; Tax, F.E.; Ryan, C.A. PEPR2 is a second receptor for the Pep1 and Pep2 peptides and contributes to defense responses in Arabidopsis. Plant Cell 2010, 22, 508–522. [Google Scholar] [CrossRef] [PubMed]
  56. Krupkova, E.; Immerzeel, P.; Pauly, M.; Schmulling, T. The TUMOROUS SHOOT DEVELOPMENT2 gene of Arabidopsis encoding a putative methyltransferase is required for cell adhesion and co-ordinated plant development. Plant J. Cell Mol. Biol. 2007, 50, 735–750. [Google Scholar] [CrossRef] [PubMed]
  57. Naoumkina, M.; Thyssen, G.N.; Fang, D.D. RNA-seq analysis of short fiber mutants Ligon-lintless-1 (Li 1) and -2 (Li 2) revealed important role of aquaporins in cotton (Gossypium hirsutum L.) fiber elongation. BMC Plant Biol. 2015, 15, 65. [Google Scholar] [CrossRef] [PubMed]
  58. Zhang, W.; Zhang, H.; Liu, K.; Jian, G.; Qi, F.; Si, N. Large-scale identification of Gossypium hirsutum genes associated with Verticillium dahliae by comparative transcriptomic and reverse genetics analysis. PLoS ONE 2017, 12, e0181609. [Google Scholar] [CrossRef] [PubMed]
  59. Xu, L.; Zhu, L.; Tu, L.; Liu, L.; Yuan, D.; Jin, L.; Long, L.; Zhang, X. Lignin metabolism has a central role in the resistance of cotton to the wilt fungus Verticillium dahliae as revealed by RNA-Seq-dependent transcriptional analysis and histochemistry. J. Exp. Bot. 2011, 62, 5607–5621. [Google Scholar] [CrossRef]
  60. Artico, S.; Ribeiro-Alves, M.; Oliveira-Neto, O.B.; de Macedo, L.L.; Silveira, S.; Grossi-de-Sa, M.F.; Martinelli, A.P.; Alves-Ferreira, M. Transcriptome analysis of Gossypium hirsutum flower buds infested by cotton boll weevil (Anthonomus grandis) larvae. BMC Genom. 2014, 15, 854. [Google Scholar] [CrossRef]
  61. Bowman, M.J.; Park, W.; Bauer, P.J.; Udall, J.A.; Page, J.T.; Raney, J.; Scheffler, B.E.; Jones, D.C.; Campbell, B.T. RNA-Seq transcriptome profiling of upland cotton (Gossypium hirsutum L.) root tissue under water-deficit stress. PLoS ONE 2013, 8, e82634. [Google Scholar] [CrossRef]
  62. Zhang, F.; Zhu, G.; Du, L.; Shang, X.; Cheng, C.; Yang, B.; Hu, Y.; Cai, C.; Guo, W. Genetic regulation of salt stress tolerance revealed by RNA-Seq in cotton diploid wild species, Gossypium davidsonii. Sci. Rep. 2016, 6, 20582. [Google Scholar] [CrossRef] [PubMed]
  63. Stiff, M.R.; Haigler, C.H. Recent Advances in Cotton Fiber Development; Flowering and Fruiting in Cotton Tennessee; The Cotton Foundation: Cordova, TN, USA, 2012; pp. 163–192. [Google Scholar]
  64. Yang, Z.; Zhang, C.; Yang, X.; Liu, K.; Wu, Z.; Zhang, X.; Zheng, W.; Xun, Q.; Liu, C.; Lu, L. PAG1, a cotton brassinosteroid catabolism gene, modulates fiber elongation. New Phytol. 2014, 203, 437–448. [Google Scholar] [CrossRef]
  65. He, X.C.; Qin, Y.M.; Xu, Y.; Hu, C.Y.; Zhu, Y.X. Molecular cloning, expression profiling, and yeast complementation of 19 beta-tubulin cDNAs from developing cotton ovules. J. Exp. Bot. 2008, 59, 2687–2695. [Google Scholar] [CrossRef] [PubMed]
  66. Javelle, M.; Vernoud, V.; Rogowsky, P.M.; Ingram, G.C. Epidermis: The formation and functions of a fundamental plant tissue. New Phytol. 2011, 189, 17–39. [Google Scholar] [CrossRef]
  67. Lu, R.; Zhang, J.; Liu, D.; Wei, Y.L.; Wang, Y.; Li, X.B. Characterization of bHLH/HLH genes that are involved in brassinosteroid (BR) signaling in fiber development of cotton (Gossypium hirsutum). BMC Plant Biol. 2018, 18, 304. [Google Scholar] [CrossRef] [PubMed]
  68. Ma, L.; Hu, L.; Fan, J.; Amombo, E.; Khaldun, A.B.M.; Zheng, Y.; Chen, L. Cotton GhERF38 gene is involved in plant response to salt/drought and ABA. Ecotoxicology 2017, 26, 841–854. [Google Scholar] [CrossRef] [PubMed]
  69. Petrovska, B.; Jerabkova, H.; Kohoutova, L.; Cenklova, V.; Pochylova, Z.; Gelova, Z.; Kocarova, G.; Vachova, L.; Kurejova, M.; Tomastikova, E.; et al. Overexpressed TPX2 causes ectopic formation of microtubular arrays in the nuclei of acentrosomal plant cells. J. Exp. Bot. 2013, 64, 4575–4587. [Google Scholar] [CrossRef]
Genes 10 00119 g001 550
Figure 1. Statistics for transcript levels at each development stage, the numbers of genes and classification were divided by 0.5 < FPKM < 5, 5 < FPKM < 100, and 100 < FPKM in each sample. FPKM: fragments per kilobase of exon per million reads.
Figure 1. Statistics for transcript levels at each development stage, the numbers of genes and classification were divided by 0.5 < FPKM < 5, 5 < FPKM < 100, and 100 < FPKM in each sample. FPKM: fragments per kilobase of exon per million reads.
Genes 10 00119 g001
Genes 10 00119 g002 550
Figure 2. Vertical of L1 relative to L2 and horizontal comparisons of genes that showed differential expression levels between the two lines at the same developmental stage and between different stages of an individual line. Red numbers represent upregulated genes; black numbers indicate down-regulated genes. The differentially expressed genes were identified by the criteria FPKM > 0.5, |log2(FC)| > 1.
Figure 2. Vertical of L1 relative to L2 and horizontal comparisons of genes that showed differential expression levels between the two lines at the same developmental stage and between different stages of an individual line. Red numbers represent upregulated genes; black numbers indicate down-regulated genes. The differentially expressed genes were identified by the criteria FPKM > 0.5, |log2(FC)| > 1.
Genes 10 00119 g002
Genes 10 00119 g003 550
Figure 3. (a) Different gene expression profiles in the two recombinant inbred lines (RILs). Each square represents a profile of gene expression trend. The profile ID and gene number in the corresponding profile are shown in the top and bottom left corners, respectively. (b) and (c) Venn diagram showed the different genes set between L1 and L2 in profile 4 and profile 22, respectively. (d) and (e) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of common genes between L1 and L2 in expression profile 4 and 22. The size of the ball represents the genes number enriched in the pathway; the depth of the color represents the size of the p-value.
Figure 3. (a) Different gene expression profiles in the two recombinant inbred lines (RILs). Each square represents a profile of gene expression trend. The profile ID and gene number in the corresponding profile are shown in the top and bottom left corners, respectively. (b) and (c) Venn diagram showed the different genes set between L1 and L2 in profile 4 and profile 22, respectively. (d) and (e) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of common genes between L1 and L2 in expression profile 4 and 22. The size of the ball represents the genes number enriched in the pathway; the depth of the color represents the size of the p-value.
Genes 10 00119 g003
Genes 10 00119 g004 550
Figure 4. Gene ontology (GO) enrichment of genes listed in expression profile 4. (a) and (b) show the top 30 terms of GO enrichment for 1112 and 1060 genes unique to L1 and L2, respectively. The main different terms are highlighted in red.
Figure 4. Gene ontology (GO) enrichment of genes listed in expression profile 4. (a) and (b) show the top 30 terms of GO enrichment for 1112 and 1060 genes unique to L1 and L2, respectively. The main different terms are highlighted in red.
Genes 10 00119 g004
Genes 10 00119 g005 550
Figure 5. GO enrichment of genes listed in expression profile 22. (a) and (b) show the top 30 terms of GO enrichment for 1128 and 748 genes unique to L1 and L2, respectively. The main different terms names are highlighted in red.
Figure 5. GO enrichment of genes listed in expression profile 22. (a) and (b) show the top 30 terms of GO enrichment for 1128 and 748 genes unique to L1 and L2, respectively. The main different terms names are highlighted in red.
Genes 10 00119 g005
Genes 10 00119 g006 550
Figure 6. Weighted gene co-expression network analysis (WGCNA) of differentially expressed genes (DEGs) in L1 and L2 at five time points of fiber development. (a) Hierarchical dendrogram showing co-expression modules identified by WGCNA. Each leaf in the tree represents one gene. The major tree was divided into 13 modules based on calculation of eigengenes; each module is highlighted in a different color. (b) Module–sample relationships. Each row represents a module, and the correlation coefficient and the e-value are shown in each square. Each column corresponds to a sample tissue and replication (R). The number of genes and transcription factors enriched in the module are shown in the left module squares and parentheses, respectively. The modules names in red were highly associated with the high-quality fiber line (L1).
Figure 6. Weighted gene co-expression network analysis (WGCNA) of differentially expressed genes (DEGs) in L1 and L2 at five time points of fiber development. (a) Hierarchical dendrogram showing co-expression modules identified by WGCNA. Each leaf in the tree represents one gene. The major tree was divided into 13 modules based on calculation of eigengenes; each module is highlighted in a different color. (b) Module–sample relationships. Each row represents a module, and the correlation coefficient and the e-value are shown in each square. Each column corresponds to a sample tissue and replication (R). The number of genes and transcription factors enriched in the module are shown in the left module squares and parentheses, respectively. The modules names in red were highly associated with the high-quality fiber line (L1).
Genes 10 00119 g006
Genes 10 00119 g007 550
Figure 7. Heatmap comparison DEGs associated with fiber developmental stages. (a), (b), (c), and (d) show the DEGs associated to the overrepresented functional categories and transcription factors at 10, 15, 20, and 30 days post-anthesis (DPA), respectively.
Figure 7. Heatmap comparison DEGs associated with fiber developmental stages. (a), (b), (c), and (d) show the DEGs associated to the overrepresented functional categories and transcription factors at 10, 15, 20, and 30 days post-anthesis (DPA), respectively.
Genes 10 00119 g007
Genes 10 00119 g008 550
Figure 8. Correlation networks for the dark red and violet modules. Hub genes are indicated by red circles; green circles indicate transcription factors; yellow circles indicate both a hub gene and a transcription factor.
Figure 8. Correlation networks for the dark red and violet modules. Hub genes are indicated by red circles; green circles indicate transcription factors; yellow circles indicate both a hub gene and a transcription factor.
Genes 10 00119 g008
Table
Table 1. Fragments per kilobase of exon per million reads (FPKMs) and functional categories of genes significantly up-regulated in L1 on 10DPA.
Table 1. Fragments per kilobase of exon per million reads (FPKMs) and functional categories of genes significantly up-regulated in L1 on 10DPA.
Gene NameL1D10L2D10Description in Arabidopsis thaliana
Transcription factors
Gh_A11G06926.490.9homeobox-leucine zipper protein 4
Gh_D07G03628.391.6Integrase-type DNA-binding superfamily protein
Gh_D10G027049.3111.65homeobox-3
Arginine biosynthesis
Gh_A05G214311.592arginase
Gh_A11G23546.861.47glutamate dehydrogenase 1
Brassinosteroid biosynthesis
Gh_D05G018419.252.36Cytochrome P450 superfamily protein
Gh_A05G012241.0611.02Cytochrome P450 superfamily protein
Metabolic pathways
Gh_A03G050719.781.02Glycosyl hydrolases family 32 protein
Gh_A05G347366.675.2110-formyltetrahydrofolate synthetase
Gh_A05G399754.475.124-coumarate:CoA ligase 2
Gh_A03G193810.111.62N-acetylglucosamine-1-phosphate uridylyltransferase 2
Gh_A05G04795.741.09aldehyde dehydrogenase 11A3
Gh_D07G069213.392.57UDP-glucose 6-dehydrogenase family protein
Gh_D02G07619.21.95Thiamin diphosphate-binding fold (THDP-binding) protein
Gh_A10G23277.371.75ACT domain-containing small subunit of acetolactate synthase protein
Gh_A09G244920.785.87NAD(P)-binding Rossmann-fold superfamily protein
mRNA surveillance pathway
Gh_A05G02386.550.21homolog of CFIM-25
Gh_A03G1654133.1739.45poly(A) binding protein 2
Gh_D02G2070108.0233.65poly(A) binding protein 2
Plant hormone signal transduction
Gh_D11G198915.55.38Auxin-responsive GH3 family protein
Gh_D09G158514.321.17PYR1-like 4
Gh_A09G242111.061.06PYR1-like 4
Table
Table 2. FPKMs and functional categories of genes significantly down-regulated in L1 on 10DPA.
Table 2. FPKMs and functional categories of genes significantly down-regulated in L1 on 10DPA.
Gene NameL1D10L2D10Description in A. thaliana
Phenylalanine, tyrosine and tryptophan biosynthesis
Gh_A13G02581.5912.58aspartate aminotransferase
Gh_A13G06031.367.683-deoxy-d-arabino-heptulosonate 7-phosphate synthase
Biosynthesis of amino acids
Gh_A07G06190.429.83-phosphoserine phosphatase
Nicotinate and nicotinamide metabolism
Gh_D01G14781.616.54quinolinate phoshoribosyltransferase
Gh_A06G01285.3148.44GA requiring 3
Phagosome
Gh_D13G188311.9796.74calreticulin 1a
Transcription factors
Gh_A05G05282.7311.38WRKY DNA-binding protein 11
Gh_A13G17289.9146.36Zinc finger C-x8-C-x5-C-x3-H type family protein
Table
Table 3. FPKMs and functional categories of genes significantly upregulated in L1 on 20DPA.
Table 3. FPKMs and functional categories of genes significantly upregulated in L1 on 20DPA.
Genes idL1D20L2D20Description
Biosynthesis of unsaturated fatty acids
Gh_Sca007938G0110.651.5beta-ketoacyl reductase 1
Gh_D01G0186139.5529.59acyl-CoA oxidase 3
Gh_A13G2171112.4136.72acyl-CoA oxidase 4
Butanoate metabolism
Gh_D12G043613.791.9short-chain dehydrogenase-reductase B
Gh_A12G141456.9716.99glutamate decarboxylase 4
Gh_A07G0810110.446.02glyoxylate reductase 1
Cutin, suberine and wax biosynthesis
Gh_A07G09915.360.72Jojoba acyl CoA reductase-related male sterility protein
Gh_D05G129467.3911.9Glucose-methanol-choline (GMC) oxidoreductase family protein
Gh_D01G1400172.149Caleosin-related family protein
Fatty acid elongation
Gh_A01G1563150.2629.073-ketoacyl-CoA synthase 6
Gh_D01G1810213.8254.53-ketoacyl-CoA synthase 6
Gh_A09G074919.877.063-ketoacyl-acyl carrier protein synthase I
Gh_Sca006141G0123.738.81acetyl-CoA carboxylase carboxyl transferase subunit β
Phagosome
Gh_A08G2381635.5971.34β-6 tubulin
Gh_A11G20951198.32173.46tubulin α-3
Gh_Sca012883G012274.54359.05tubulin β 8
Gh_D08G1960460.8975.32β-6 tubulin
Gh_D05G1052572.74176.55tubulin β 8
Gh_A02G0819357.99136.45tubulin α-2 chain
Gh_D13G1047259.25106.17tubulin α-3
Transcription factors
Gh_D12G060716.714.38basic helix-loop-helix (bHLH) DNA-binding superfamily protein
Gh_A07G117182.8217.27basic region/leucine zipper motif 53
Gh_D05G046337.913.35CCCH-type zinc finger family protein
Gh_A12G056140.1615.17B-box zinc finger family protein
Gh_D05G2596267.05106.76K-box region and MADS-box transcription factor family protein
Gh_A12G124421.194.2MYB-like 102
Gh_D08G015710.982.96AP2/B3 transcription factor family protein
Gh_D02G004340.6911.26WRKY family transcription factor family protein
Table
Table 4. FPKMs and functional categories of genes significantly downregulated in L1 on 20DPA.
Table 4. FPKMs and functional categories of genes significantly downregulated in L1 on 20DPA.
Genes NameL1D20L2D20Description in A. thaliana
Biosynthesis of secondary metabolites
Gh_D06G00960.1212.6cytochrome P450, family 82, subfamily C, polypeptide 4
Gh_D04G06050.389.89glycerol-3-phosphate acyltransferase 5
Gh_A13G06031.4710.653-deoxy-d-arabino-heptulosonate
7-phosphate synthase
Gh_D05G25544.2622.71acetyl-CoA carboxylase 1
Nitrogen metabolism
Gh_Sca069862G0118.38189.16glutamine synthase clone R1
Gh_D07G177354.01333.88glutamine synthase clone R1
Gh_A01G15861.47.56nitrate reductase 2
Gh_D01G18727.8326.87nitrate reductase 2
Phenylpropanoid biosynthesis
Gh_D10G047311.363.264-coumarate:CoA ligase 1
Gh_D08G113515.6262.86O-methyltransferase 1
Gh_A08G071132.62105.85Peroxidase superfamily protein
Gh_A04G103229.2980.06S-adenosyl-L-methionine-dependent methyltransferases
Zeatin biosynthesis
Gh_D05G18139.2343.01cytokinin oxidase 5
Gh_A05G02906.6425.84cytokinin oxidase 7
Gh_A05G16317.626.27cytokinin oxidase 5
Gh_D05G039112.2739.85cytokinin oxidase 7
Transcription factors
Gh_A02G15759.1228.68GRAS family transcription factor
Gh_D01G0974102.38371.7basic leucine-zipper 44
Gh_A11G28753.4215.35myb-like transcription factor family protein
Gh_A11G25223.5219.11myb domain protein 120
Gh_A05G37782.524.64LOB domain-containing protein 13
Gh_D07G13300.535.69NAC domain transcriptional regulator protein
Gh_D06G02540.526.37GATA-type zinc finger protein with TIFY domain
Gh_A10G05162.0626.94myb domain protein 26
Gh_D08G14240.116.98WRKY DNA-binding protein 3
Table
Table 5. Candidate hub genes in 10, 15, 20, and 30 DPA modules.
Table 5. Candidate hub genes in 10, 15, 20, and 30 DPA modules.
Gene NameKMEArabidopsis IDFunction Description in Arabidopsis thaliana Gene
10 DPA specific darkred module
Gh_D12G21720.991AT1G12240Glycosyl hydrolases family 32 protein
Gh_A10G19610.991AT2G37620actin 1
Gh_A11G10870.997AT4G03100Rho GTPase activating protein with PAK-box [46]
Gh_D06G06180.991AT4G28950RHO-related protein from plants 9
Gh_D05G05910.992AT4G31890ARM repeat superfamily protein
15 DPA of L1 specific darkturquoise module
Gh_A02G16920.968AT3G51430Calcium-dependent phosphotriesterase superfamily protein
Gh_A07G02250.983AT4G32330TPX2 (targeting protein for Xklp2) protein family [47,48]
Gh_A12G02260.971AT2G15780Cupredoxin superfamily protein
Gh_D03G10540.973AT1G01630Sec14p-like phosphatidylinositol transfer family protein
Gh_D12G00260.964AT3G51895sulfate transporter 3;1
20 DPA of L1 specific violet module
Gh_A03G03470.967AT2G28790Pathogenesis-related thaumatin superfamily protein
Gh_A05G18790.981AT4G33580beta carbonic anhydrase 5
Gh_A13G10120.965AT4G09510cytosolic invertase 2
Gh_D06G11510.968AT4G16130arabinose kinase
Gh_D10G02950.974AT1G68810basic helix-loop-helix (bHLH) DNA-binding superfamily protein 30 [49]
Gh_D02G07400.97AT1G21460Nodulin MtN3 family protein
Gh_D07G01440.969AT5G20860Plant invertase/pectin methylesterase inhibitor superfamily [50]
Gh_D08G27250.973AT3G62690AtL5
20&30 DPA of L1 specific pink module
Gh_A08G14860.95AT5G47120BAX inhibitor 1
Gh_A08G25080.944AT4G27000RNA-binding (RRM/RBD/RNP motifs) family protein
Gh_A11G27260.961AT3G13540myb domain protein 5 [51,52]
Gh_D03G18430.958AT1G32050SCAMP family protein
Gh_D09G03470.962AT5G14240Thioredoxin superfamily protein
30 DPA of L1 specific darkgrey module
Gh_A09G11370.981AT5G04160UDP-URONIC acid transporter 1 [53,54];
Gh_A10G17540.989AT5G06390FASCICLIN-like arabinogalactan protein 17 precursor
Gh_D01G16970.982AT1G20550O-fucosyltransferase family protein
Gh_D08G21520.985AT2G33170Leucine-rich repeat receptor-like protein kinase family protein [55]
Gh_D10G10560.99AT5G64030S-adenosyl-L-methionine-dependent methyltransferases superfamily protein [56]

Share and Cite

MDPI and ACS Style

Zou, X.; Liu, A.; Zhang, Z.; Ge, Q.; Fan, S.; Gong, W.; Li, J.; Gong, J.; Shi, Y.; Tian, B.; et al. Co-Expression Network Analysis and Hub Gene Selection for High-Quality Fiber in Upland Cotton (Gossypium hirsutum) Using RNA Sequencing Analysis. Genes 2019, 10, 119. https://doi.org/10.3390/genes10020119

AMA Style

Zou X, Liu A, Zhang Z, Ge Q, Fan S, Gong W, Li J, Gong J, Shi Y, Tian B, et al. Co-Expression Network Analysis and Hub Gene Selection for High-Quality Fiber in Upland Cotton (Gossypium hirsutum) Using RNA Sequencing Analysis. Genes. 2019; 10(2):119. https://doi.org/10.3390/genes10020119

Chicago/Turabian Style

Zou, Xianyan, Aiying Liu, Zhen Zhang, Qun Ge, Senmiao Fan, Wankui Gong, Junwen Li, Juwu Gong, Yuzhen Shi, Baoming Tian, and et al. 2019. "Co-Expression Network Analysis and Hub Gene Selection for High-Quality Fiber in Upland Cotton (Gossypium hirsutum) Using RNA Sequencing Analysis" Genes 10, no. 2: 119. https://doi.org/10.3390/genes10020119

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop