Arguments have raged about how many times entire-genome duplications have occurred in vertebrate evolution (and indeed whether they have occurred at all!). “At least two” seems to be the answer based on the thorough comparative analysis of the human and Fugu genomes by Klaas Vandepoele, Wouter De Vos and their colleagues.

The sequencing of the Fugu genome was always going to be a key step in unravelling the twists and turns of vertebrate genome evolution. Fugu is a member of the sister group of land vertebrates, the ray-finned fish. In addition to the two genome duplications that are long-suspected to have occurred at the base of the vertebrate tree, some studies indicate that there has been another genome duplication in this group. However, preliminary analyses of the Fugu genome did not turn up evidence of an entire-genome duplication.

The authors realized that the most powerful way to address the question of vertebrate genome duplications was to do a comparative analysis. Using BLASTP, they searched for gene families that are shared by Fugu and human, and then added in homologous sequences from other fish, the mouse and the outgroups Ciona and Drosophila. They then built phylogenetic trees for each of the 3,077 gene families identified that had between two and ten Fugu genes.

The divergences between duplicated genes were dated relative to the ray-finned fish/land vertebrate divergence event within the 752 gene families in which there was strong statistical support for the relevant branches. The authors then estimated the absolute date of the divergences of gene duplicates in the 488 gene families and found evidence that a molecular clock was operating. A total of 166 (30%) of these divergences seem to have occurred between 225 and 425 million years ago (mya) — that is, after the ray-finned fish/land vertebrate split. Further analyses clearly indicated that a suite of duplicate genes arose approximately 320 mya. The implication is clear: an entire genome duplication event in an ancestor of ray-finned fish probably gave rise to many paralogues in the present-day Fugu genome.

However, the distribution of divergence times of Fugu gene paralogues is bimodal, with approximately 70% falling in the 500–900 mya window. So, it seems that there was probably at least one, and possibly two (roughly contemporaneous), genome-duplication events much earlier in vertebrate evolution — pre-dating the ray-finned fish/land vertebrate split and even the jawed/jawless vertebrate split (∼575 mya). These data indicate that genome duplication has almost certainly had a large role in the evolution of all vertebrates.

These few ancient duplications of vertebrate genomes are fascinating; however, it will take a much more intensive sampling of vertebrate lineages before we will be able to effectively identify and analyse enough similar duplications to start making some generalizations about these events and their consequences.