Shotgun Microarrays: High-throughput Validation of Gene Expression Changes Using Multiple Platforms

Roderick V. Jensen
Wesleyan University and Brigham and Womens Hospital
Harvard Medical School

DNA Microarrays have emerged as important new tools for the
large-scale measurement of differential gene expression in a wide
variety of biomedical experiments, comparing normal and diseased
tissues, different states of biological processes, and the response to
drugs or toxins. A single cDNA or oligonucleotide microarray can
provide detailed quantitative information on the levels of expression
of thousands of genes. However, as with any laboratory measurement,
this data is subject to errors. To identify real changes in gene
expression it is essential to understand and model the sources of
errors in these measurements. Our laboratory has performed extensive
studies of microarray data from multiple microarray platforms
including homemade cDNA arrays, Amersham/Motorola Codelink arrays, and
the popular Affymetrix gene chips. The comparison of differential gene
expression measurements both within and across platforms provides the
basis for statistical error models essential for the useful
interpretation of the data. Our "shotgun microarray" approach has
successfully identified gene expression signatures in a variety of
different applications ranging from the distinction of normal human
tissues, to the classification of cancer, and the identification of
key genes in involved in differentiation and development.