An Embarrassment of Switches

What makes a human different from a chimpanzee or a mouse? Of course, we know the answer in broad outline. Mutations in the genome, sifted by natural selection, cause changes in appearance, physiology, and behavior—what geneticists call the phenotype. But we have only a vague picture of a more detailed answer. Precisely which mutations generate phenotypic evolution? It's not that we can't find the mutations. Today's DNA sequencing technology readily identifies all differences between two genomes. There are simply too many differences—tens of millions between human and chimp, for example (1). An unknown fraction of these mutations alter the phenotype. Nonetheless, the molecular effects of mutations provide a rough guide to their phenotypic effects. Some mutations change the amino acid sequence of proteins, thereby altering their functions, and some change so-called cis-regulatory regions, altering when and where proteins are produced. We know a lot about the first class, but much less about the second. Several recent papers, including one by Borneman et al. on page 815 of this issue (2), demonstrate a surprising abundance of cis-regulatory changes between closely related species.