The Comparative Transcriptome and Co-Expression of Hub Genes Analysis for Fiber Development in RIL Populations of Upland Cotton

Upland cotton (Gossypium hirsutum) is one of the important crops to produce natural fiber, and fiber length (FL) and fiber strength (FS) are important fiber quality traits. In this study, we developed a population of recombinant inbred lines (RILs) with the parents 0–153 (P1, high fiber quality) and sGK9708 (P2, low fiber quality), which were constructed for detection of elite alleles. According to the RNA-seq data of P1, P2, 69307 (L1, high fiber quality) and 69362 (L2, low fiber quality), 350/90, 71/132, 214/120, 199/201, and 287/229 genes were differentially expressed (up-/down-regulated) in high quality fiber lines (P1 and L1) in contrary to low quality fiber lines (P2 and L2) at 5, 10, 15, 20 and 25 days post-anthesis (DPA), respectively. Short Time-series Expression Miner (STEM) analysis detected the presence of five expression profiles, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation were investigated to determine the functional differences between specific genes of profile 7 and 26. A total of 234 differentially expressed genes (DEGs) belong to quantitative trait locus loci (QTLs) controlling FL and FS. Co-expression network analysis revealed the presence of two modules which were associated with various stages of fiber development. The weighted gene co-expression network analysis (WGCNA) was also carried out to distinguish 13 hub genes in blue module and 24 hub genes in turquoise module. The alignment between quantitative real-time PCR and RNA-seq data gave the clue for guaranteeing reliability of selected candidate genes for further analysis. In this study, we combine the identified genes from FL and FS QTLs with the results of STEM, WGCNA and time-course analysis. In addition, six candidate genes (GH_A04G0817, GH_A05G0644, GH_A08G0267, GH_A11G0850, GH_D02G1011 and GH_D11G3494) were selected as candidate genes to improve cotton fiber quality.