[biomcgary]

This approach appears to work well for E. coli and other organisms with streamlined genomes and large population sizes. One of the challenges with studying gene expression in humans is that it is quite noisy with regard to identifiable function (i.e., contribution to reproduction, however indirectly).

Biologists tend to treat gene expression as functional. In E. coli, this is much more likely to be true than for humans. First, E. coli is a single-cell organism so there is not conflict in differential regulation across cell types. Second, E. coli has a large population relative to its information content, which allows natural selection to operate more efficiently.

mRNA transcription (i.e., gene expression) by itself is relatively inexpensive, so natural selection only weakly optimizes expression in multicellular organisms with small populations. Protein production is much more expensive and is under stronger selection.