|Toward global gene expression profiles:
The RIKEN 19K mouse arrays
Edward R. Winstead
March 19, 2001
Japanese researchers and their American colleagues characterized patterns of gene expression in 49 embryonic and adult mouse tissues using DNA microarrays with 13,600 genes. They found striking similarities among expression patterns for certain tissues in early and adult life. But there were also tissue-specific differences in the expression of key regulatory enzymes in metabolic pathways.
It may soon be feasible to study the expression of all protein-coding mouse genes simultaneously, but assembling cDNAs for large arrays has proved difficult. Researchers at Japan's Institute of Physical and Chemical Research (RIKEN) in Tsukuba constructed libraries of mouse cDNAs before building the '19K' arrays used in the study. The arrays consist of nearly 19,000 cDNAs, but because of redundancy the total number of unique genes on the arrays is about 13,600.
"We have established an expression database for a large set of genes in 49 different tissues," the researchers write in Proceedings of the National Academy of Sciences. Yoshihide Hayashizaki, of the RIKEN Genomic Sciences Center, led the study. The researchers suggest that in the future there will be databases containing expression profiles of every mouse gene in every tissue at different stages of development, providing a picture of normal gene expression and patterns associated with disease.
To make sense of 1.8 million pieces of gene expression data, the researchers grouped, or 'clustered,' both genes and tissues according to expression patterns. As expected, the clustering revealed genes that are expressed ubiquitously, such as 'housekeeping' genes and several types of ribosomal proteins. The study's US team included Patrick O. Brown, of the Stanford University School of Medicine, who has pioneered clustering techniques in array studies.
The clustering of tissues on the basis of expression patterns revealed a "remarkably ordered pattern" among functionally related tissues. The researchers classified the following tissues in discrete groups based on similarity in expression pattern: the muscular tissues (tongue, heart, and skeletal muscle), adult brain regions (cortex, cerebellum, and brain) and eye, and digestive organs (pancreas, small intestine, stomach, cecum, and colon).
Enzymes in all classes of metabolic pathways were represented on the RIKEN arrays, and the researchers focused on gene expression in 78 metabolic pathways. They found "striking" similarities between genes coding for enzymes in the same metabolic pathway. "Intriguingly, genes were clustered into two major groups in all of the metabolic pathways," the researchers write. "One group was the ubiquitously expressed gene set and the other was a tissue-specifically expressed gene set."
See related GNN article
. . .