Semantic resources project/CandidateResources/Resources/CMAP

Lamb et al. |"The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease." (Science, 2006)

The "connectivity map" is a large resource of microarray data experiments, indexed by the perturbation and cell line that were used to produce them; it's |hosted by the Broad Institute (registration required) and comes with a web-service that they provide -- you input a "signature" of gene expression, and it returns the perturbations or other experimental parameters that match that input profile.

It's a pretty basic example of a large-scale "join-through-data" using simple gene-based annotations and expression data. But it's basically the most straightforward thing you would want to do with a large, well-annotated set of expression experiments, and so it's worthwhile studying it (or considering importing it, or something like it, as a resource).

The CMAP project also makes use of the mSigDB, which is part of (or derived from) the GSEA system at the Broad.