Identification of uniformly expressed genes in Escherichia coli K-12 by using RNA-Seq data

Abstract

Escherichia coli have been one of the famous model organisms in numerous studies by advantages of being easy to culture compared to other organisms of bacterial species. The fundamental principle of transcriptome analysis is a comparative study on interested case samples and control samples to define differentially expressed genes (DEGs). Reference genes, genes that are expressed at a steady level for any circumstances of cells, are required to perform a comparative study on those two different types of samples. By the definition of housekeeping genes (HKGs), essential genes in survival of living organisms, they should be existed in any conditions at a steady expression level theoretically. Thus, HKGs are often used as reference genes. There are numerous studies about defining HKGs by microarray analysis but few by RNA-Seq analysis. The objective of this study is to define Uniformly Expressed Genes (UEGs) in E. coli by RNA-Seq analysis. Ninety-six of public RNA-Seq data were analyzed to define UEGs in E. coli using HTSeq-DESeq and RSEM-DESeq pipelines. During the process of analyzing UEGs, comparative studies of count algorithms, normalization algorithms, and statistical analysis were carried out to examine the consistency of the expression levels by each algorithm. Afterwards, 120 of candidate UEGs by each pipeline were selected. Based on the intersection of candidate UEGs by both pipelines, a comparative study with defined reference genes by microarray analysis was performed. Genes that were candidate UEGs in both HTSeq-DESeq and RSEM-DESeq pipelines and known reference genes from microarray were defined as UEGs. Only 2 out of 15 known reference genes, cca and ssrA, were confirmed as UEGs by RNA-Seq analysis. Furthermore, transcripts of candidate UEGs and known reference genes were quantified by quantitative real-time PCR (qRT-PCR) under various growth conditions to analyze difference between UEGs selection derived from analysis of microarray and RNA-Seq. Depending on analyzing tools, the result of RNA-Seq analysis can be different. The result of UEGs by RNA-Seq analysis did not have a strong correlation with the defined reference genes by microarray analysis. This phenomenon could be due to the different process of experiments. Thus, careful consideration should be taken in selecting known HKGs that are defined by microarray analysis for applying to RNA-Seq analysis.