As whole-genome expression profiling has become increasingly affordable, <br>
publicly available repositories of microarray data are expanding rapidly. <br>
Still, a large fraction of the enormous information content in these databases<br>
 remains unexplored. In this talk, I will discuss data integration techniques <br>
allowing us to take advantage of this data to discover new components of metabolic<br>
 pathways. Specifically, we have developed a technique that systematically searches<br>
 data repositories for genes that are consistently and specifically co-expressed <br>
with a pathway of interest. Using this method, we have successfully identified five<br>
 proteins as required for functional heme biosynthesis, and a novel regulator of <br>
oxidative phosphorylation acting on mitochondrial mRNA.