Gene expression studies with microarrays and SAGE: biological and analytical principles1

Abstract

AbstractThe present article continues a series of publications and internet presentations (www.dgkl.de/bioninformatik) on analytical and postanalytical aspects of gene expression studies. We present a critical laboratory diagnostics view on the two most important analytical techniques for genome-wide expression profiles, i.e., DNA microarrays (biochips) and Serial Analysis of Gene Expression (SAGE). A gene expression profile is a list of mRNA species (transcripts) with corresponding quantities in a given tissue. Of the 10–20 pg total RNA per cell, mRNA constitutes only 1–5% and is dispersed among about 30,000 individual transcripts. Therefore, most of them are present in extremely low amounts reaching one molecule per cell, whereas a few show 10,000-fold higher concentrations. The analytical challenge of whole genome expression profiling is therefore substantial.Both techniques have specific strengths and weaknesses: Microarray experiments yield semi-quantitative results with less effort in a relatively short time, but they can analyze only known gene products. SAGE is extremely costly, time-consuming and difficult to interpret, but it counts well-defined sequence tags and detects – at least in principle – any mRNA in the sample. Due to the large number of hardly standardized preanalytical and analytical procedures, both techniques suffer from accuracy and precision problems and may lead to contradictory conclusions. A signal highly expressed with microarrays may be undetectable with SAGE and vice versa. High throughput sequencing techniques could help to close the gap between both techniques.