Integrating database information in microarray expression analyses: Application to melanoma cell lines profiled in the NCI60 data set.

A series of successful microarray experiments typically results in a data matrix containing a list of a few thousand genes and their measured expression values across various conditions tested. The understanding of why some genes are clustered in their expression pattern can be greatly facilitated by gathering all prior knowledge about all the genes in a given cluster. Here we demonstrate a strategy using the NCI60 data set (Scherf U, Ross DT,Waltham M, et al., Nature Genet 2000;24(3):236-244) on a cluster of genes highly expressed in the melanoma cell lines. By traversing public databases using cross-references between the databases,we show that the Gene Ontology (GO) classifications of some of the genes are consistent with their products' involvement in melanin biosynthesis.