It's the Sequence, Stupid!

One of the surprises revealed by comparative genome sequencing is that closely related species share remarkably similar complements of genes. For example, a recent evaluation of the human gene catalog found at most 168 genes without close homologs in mouse or dog, with perhaps as few as 12 representing newly evolved protein-coding regions (1). Moreover, the corresponding genes tend not to differ much in their coding sequences: Nearly 80% of amino acids are identical between orthologous human and mouse proteins (2). Although this leaves many potentially functional coding changes, these observations lend further credence to the proposal, first made more than 30 years ago, that many of the observed differences between species likely stem from when and where the products of the genes are made (3). But what governs these changes in gene expression? There is no shortage of possible explanations—differences in external cues and cellular milieus, in how genomes are packaged, in the proteins that control transcription, and in the regulatory sequences to which they bind. Strikingly, on page 434 of this issue (4), Wilson et al. show that in human and mouse liver cells, the differences in regulatory sequences dominate all other factors.