Meta-analysis of hypoxic transcriptomes from public databases

Abstract Hypoxia is an insufficient level of oxygen supply in the cell, and hypoxia-inducible factor is a central regulator of oxygen homeostasis. In order to elucidate functional insights in hypoxic response in data-driven way, we attempted meta-analysis of hypoxic transcriptome for public expression data which have been archived as microarray and RNA-seq data in public databases, NCBI Gene Expression Omnibus (GEO) and EBI ArrayExpress. While various hypoxic conditions (oxygen concentration and duration of hypoxia) and cell lines are taken in the stored data, we manually curated possible pairs of transcriptome before and after hypoxic stress from microarray and RNA-seq data. As a result, we got 37 pairs in human and 53 pairs in mouse from microarray and 23 pairs in human from RNA-seq. We counted the number of experiments for all genes and classified into three categories, which are up-regulated, down-regulated, and unchanged. We then compared human and mouse in microarray, and microarray and RNA-seq in human. Genes up-regulated in all records contained well-studied hypoxia responsive genes, while the number of down-regulated genes were smaller and most of them included unfamiliar ones in hypoxia. Meta-analysis approach to public gene expression database is useful for selecting candidate genes from gene expression profiles of various experimental conditions. The data produced in this study can be a good resource for hypoxic researches.