Predicting the basis of convergent evolution.

R epeated evolution of similar traits in organisms facing the same ecological challenges has long captured the interest of evolutionary biologists (1–4). Naturally occurring examples of “convergent evolution” offer new opportunities to ask about predictability in evolution. Do complex genomes mean that there are endless possibilities for adapting to an ecological challenge? Or must evolution target the same genes, or even the same amino acids in the same proteins, in order to increase the fitness and therefore survival of different species facing similar challenges? Natarajan et al. (5), on page 336 of this issue, provide an example of an integrated approach to answer these questions. By using a combination of genetic data with experimental tests, they show that evolution of a new protein function in response to low-oxygen, high-altitude conditions can occur through different genetic mechanisms across a wide diversity of avian species living at high and low elevations. Hemoglobin proteins found in different vertebrate species living either at low or high altitudes provide classic examples of convergent protein evolution. Natarajan et al. sequenced, expressed, and measured oxygen affinity for hemoglobins isolated from 56 species of birds clustered into related lowand high-altitude–species pairs. They found convergent evolution of high-oxygen affinity in proteins from high-elevation species. They then determined the genetic basis for this functional shift in hemoglobin oxygen affinity. Evolution of new protein functions is thought to be under biophysical constraint, limiting genetic changes to certain portions of proteins, or to specific amino acid substitutions (6, 7), potentially making evolution largely predictable. On the other hand, evolutionary trajectories have been shown to be historically contingent, with possible evolutionary paths being dependent on the ancestral starting point (8, 9), resulting in relatively unpredictable, distinct genetic changes in different lineages subjected to similar evolutionary pressures. Were there common genetic changes in the hemoglobin from high-elevation–avian species? Amino acids were identified that diverged between lowand high-elevation– species pairs and were associated only with high-altitude environments. Although many amino acid differences were found between species pairs, only four were replicated in different lineages. Two of these positions are known to play key roles in protein function, yet these changes occurred in only two or three species pairs. The single most common substitution was found primarily in hummingbird species pairs but was also seen in one species of flowerpiercers. Was this change responsible for the convergent evolution of high-oxygen affinity? Natara-