Research Of Transcription Regulation And Gene Expression Correlated Network Of Cotton

Gene expression was the process of transformation from genetic information to functional productby transcription，splicing and translation，which was the foundational and key process of all life activity.Gene expression was a complex process，gene transcription of which was the key and critial step.Transcriptional factors （TFs） promoted to transcription or regulate the gene expression by bindingwith Cis-regulatory elements （CRE），and regulated the expression level of genes and adjusted theactivity of life. Determination of the binding relationship between TFs and CRE could thus determinethe regulation relationship between TFs and promoters of functional genes，which was one of the mostimportant component of gene regulation network study.In this study，we attempted to construct the gene regulation network by bioinformation methodsbased on the draft genome of Gossypium raimondii （Gr） and published RNA-seq data in cotton.Firstly，we genome-widely scanned the types and distribution of CRE in the1000base pair （bp）upstream of annotated genes of Gr and Arabidopsis thaliana （At） by using the data of Plant Cis-actingRegulatory DNA Elements database （PLACE），10bp length was used as a frame to statistic thedistribution of CRE in all gene promoters. The result indicated that some typical CRE such asTATABOX and ACGTG could generate obvious “peak-like” distribution at specific position of all genepromoters，which demonstrated that there was distribution law existed in all gene promoters for CRE.Along with this study，we also constructed the basic database of cotton.The CRE analysis could only reflect the static distribution and for further determination of therelationship between TFs and regulated genes need to analyze more data. Therefore，based on the bigdata theory，we downloaded all available RNA-seq raw read （~100bp length） from internet andconstructed the mega cotton RNA-seq database，which was mainly integrated and categorized into twotissues，fiber and seed. All the raw reads from different tissues were mapped to annotated gene library（which constructed by only CDS region） of Gossypium raimondii. Different samples from twocategories were separately calculated the value of RPKM（Reads Per Kilo bases per Million reads） andintegrated them together as “expression cluster” for each gene. All the expression cluster were used toseparately calculate the correlation coefficient （Pearson’s） in two tissues to determine the interrelationamong all genes. Two genes were considered as significantly active correlation if the expression clusterwere significantly correlated between them.In48samples of fiber and72samples of seed，thecorrelation coefficient were calculated separately，significantly correlated genes were screened toconstruct the expression network，by controlling the sample size of all screened genes，the networkcould be divided into several sub-networks，the gene ontology （GO） analysis of each sub-networkrevealed that the gene enrichment agreed with their biological function，which indicated that the geneswere significantly correlated in expression level could also be significantly correlated in function.According to the theory of TFs regulation，the increasing of expression level of TFs could obviously effected the target genes. Therefore，expression cluster correlation possibly existed betweenTFs and their target genes. In this study，all the target genes regulated by different TF families inGossypium raimondii were screened for enrichment analysis，three TF families （MYB，NAC andbHLH） with clear function enrichment were screened for further study，the result indicated that TFs inNAC family were significantly enriched in fiber related biosynthesis pathways and some of pathwayenriched by other families were also matched previous studies，which demonstrated that potentialregulation relationship existed between TFs and their expression significantly correlated genes.This study firstly reported the correlationship by using abundant of RNA-seq data，and proved thatthe correlation among genes could reflect their function relationshiop. Meanwhile，through analyzingthe TFs and their significantly correlated genes，the regulation network was revealed. This studyprovided new ideas for predication of TF targeted genes and construction of gene regulation network incotton.