Serial Analysis of Gene Expression

Abstract

Molecular medical research has traditionally required hypothesis-driven strategies, often focusing on identification of linear signaling pathways that mediate disease pathogenesis. This approach does not identify novel pathways, or genes that act together to produce pathology or a disease-permissive milieu. To generate new hypotheses that require no a priori assumptions about disease pathogenesis, gene-expression profiles of normal and diseased tissues or cell types can be compared. Although determining protein-expression patterns would be preferable, necessary methodologies are not readily available, and protein diversity generated by post-translation modifications adds technical complexities to this strategy. On the other hand, techniques for the study of RNA-expression patterns are available, and represent a reasonable “data mining” alternative. Complex relationships between multiple molecules that regulate disease pathogenesis should be identified, at least in part, by analyzing transcript profiles generated from normal and diseased states. Differential display (1,2), subtractive hybridization (3), and subtraction libraries (4) are semiquantitative, comparative tools that have been utilized previously. However, molecular pathway discovery has been revolutionized by development of high-throughput, quantitative techniques, including serial analysis of gene expression (SAGE) and hybridization array (5,6). Because both of these methods generate transcript libraries or transcriptomes, which catalog thousands of simultaneously expressed genes, a small laboratory can comprehensively determine differential gene-expression profiles within months.