A Novel Strategy for Comprehensive Prediction of Gene Function by Integration of Transcriptomics and Metabolomics

Abstract

Gene expression profiles of Arabidopsis under various conditions and in various organs are available in public databases. Analysis of co-expressed genes using such microarray data enables us to predict gene function comprehensively. In our previous study on integration of transcriptome and metabolome data of sulfur-starved Arabidopsis, we were able to identify the genes encoding the enzymes involved in particular metabolic pathways. In a similar way, we can predict the genes encoding transcriptional factors involved in particular metabolic pathways. The genes which co-express in all experimental conditions are considered to be regulated by a single regulatory mechanism, while the genes which co-express under particular conditions are considered to be regulated by several different mechanisms. Such a comprehensive analysis of all available microarray data enables us to predict regulatory networks. Collection of metabolite profiles under various conditions and integrated analysis with gene expression profiles is useful for precise prediction of gene function.