Meta-Analysis of Hypoxic Transcriptomes from Public Databases.

Hidemasa Bono,Kiichi Hirota

doi:10.3390/biomedicines8010010

Hidemasa Bono, Kiichi Hirota

Open Access

https://doi.org/10.3390/biomedicines8010010

Copy DOI

Abstract

Hypoxia is the insufficiency of oxygen in the cell, and hypoxia-inducible factors (HIFs) are central regulators of oxygen homeostasis. In order to obtain functional insights into the hypoxic response in a data-driven way, we attempted a meta-analysis of the RNA-seq data from the hypoxic transcriptomes archived in public databases. In view of methodological variability of archived data in the databases, we first manually curated RNA-seq data from appropriate pairs of transcriptomes before and after hypoxic stress. These included 128 human and 52 murine transcriptome pairs. We classified the results of experiments for each gene into three categories: upregulated, downregulated, and unchanged. Hypoxic transcriptomes were then compared between humans and mice to identify common hypoxia-responsive genes. In addition, meta-analyzed hypoxic transcriptome data were integrated with public ChIP-seq data on the known human HIFs, HIF-1 and HIF-2, to provide insights into hypoxia-responsive pathways involving direct transcription factor binding. This study provides a useful resource for hypoxia research. It also demonstrates the potential of a meta-analysis approach to public gene expression databases for selecting candidate genes from gene expression profiles generated under various experimental conditions.

Highlights

The development of high-throughput sequencing technology has enabled cost-effective reading of tens of millions of base pairs in a single run
Using the conventional keyword search by All Of gene Expression (AOE), we integrated graphical web tool for gene expression data called AOE, which has been maintained as an index of public gene expression
While over two million transcriptome samples have been archived in public databases

Summary

Introduction

The development of high-throughput sequencing technology has enabled cost-effective reading of tens of millions of base pairs in a single run. RNA-seq takes advantage of this technology to elucidate the expression profiles of genes and transcriptomes assayed under particular conditions by producing counts of sequences corresponding to genes of interest. Published transcriptome data have been archived to two large public databases (DBs), the Gene. DBs is over two million in samples and near a hundred thousand in data series. They are freely accessible and ready to be reused for data-driven research. Large-scale comparison among archived data has not often been carried out because of the technical challenges presented by the magnitude, complexity, and cumbersome nature of the data. Even if very large amounts of data can be successfully downloaded, it can be difficult to interpret data from different laboratories, as experimental protocols used are not uniform

Objectives

Methods

Results

Conclusion