It is still early days for the field of comparative genomics. Only around a dozen species of animal have so far had their complete DNA sequence determined, even to draft coverage. These are predominantly the widely studied model species, such as mice, fruitflies and nematode worms, or species of particular interest to humans, such as the malaria-carrying mosquito.

It may come as a surprise, therefore, to find that the list now includes not one, but two species of Tetraodontiformes, a relatively obscure group of fish also known as puffers. Following on from the publication two years ago of the genome sequence of the Japanese pufferfish Takifugu rubripes1, Jaillon and colleagues2 report, on page 946 of this issue, the near-complete sequence of the spotted green pufferfish Tetraodon nigroviridis. Takifugu is a poisonous marine fish best known to connoisseurs of sushi restaurants, whereas Tetraodon is a small, brackish-water pufferfish commonly kept in aquaria. But, like all Tetraodontiformes, the two species share a feature of great convenience for genomics: their cells possess less DNA than those of any other group of backboned animals — about eight or nine times less than human cells.

Although the Tetraodon genome is small compared with that of other vertebrates, sequencing it was still a hugely formidable task. The research reported in this issue2 was performed in a collaboration between Genoscope in France and the Broad Institute of the Massachusetts Institute of Technology and Harvard University in the United States. Together they have generated a genome sequence of impressive accuracy and coverage, with 64% of the DNA sequence mapped to specific chromosomes3.

By comparing the Tetraodon genome sequence with that of humans, Jaillon et al. even allow us to peer into the genome of the last common ancestor of pufferfish and humans — a primitive bony fish that lived hundreds of millions of years ago. The descendants of this long-extinct ancestor split into two distinct evolutionary lineages: the actinopterygians (ray-finned fish), which include teleosts such as pufferfish and zebrafish, and the sarcopterygians (lobe-fins), which include lungfish, coelacanths and ourselves (Fig. 1). By matching up the genes on each pufferfish chromosome to the related genes on human chromosomes, Jaillon et al. deduce that the extinct ancestor had 12 pairs of chromosomes. Work on partially completed genome sequences had suggested this number4,5, but the new analyses add fascinating detail to the picture. For example, it is now possible to say which genes were on which chromosomes, despite this unknown animal having been extinct for more than 400 million years.

Figure 1: Evolutionary relationships between humans, pufferfish and other vertebrates.
figure 1

Jaillon et al.2 have sequenced the genome of the pufferfish Tetraodon nigroviridis. By comparing this genome sequence with that of humans, the authors deduce that the extinct ancestor of actinopterygians (ray-finned fish, including pufferfish) and sarcopterygians (lobe-finned fish, the lineage that gave rise to humans) had 12 pairs of chromosomes (n = 12). They also show that a whole-genome duplication (WGD) occurred during the evolution of ray-finned fish.

One puzzling observation concerns the apparent stability of the genomes of ray-finned fish. It seems that the ancestral genome underwent as few as ten large interchromosomal rearrangements (exchanges, fissions or fusions) to give rise to the present-day Tetraodon genome. Indeed, 11 Tetraodon chromosomes have not experienced any such rearrangements. Only one human chromosome (14) can make the same claim, despite the timescale being identical.

Although the genomes of ray-finned fish may have been slowly evolving in terms of chromosome breakages and fusions, they have experienced a cataclysmic event in their history. Jaillon and colleagues' analyses of the complete Tetraodon genome sequence show clearly that a duplication of the whole genome occurred sometime within the ray-finned-fish lineage. This inference is not new, having been previously suggested from analyses of the Hox-gene clusters and other gene families in zebrafish, Takifugu and other teleosts4,5,6,7, but the conclusion has remained controversial8.

Two new analyses should now settle the issue, however. First, Jaillon and colleagues plotted the chromosome positions for about 750 pairs of ‘ancient’ duplicated genes within the Tetraodon genome, revealing related pairs of chromosomes or chromosomal regions. Every chromosome is involved, consistent with an ancient whole-genome duplication. In the second test, chromosome positions for more than 6,000 pufferfish genes were compared with the positions of related genes in the human genome. This revealed a striking pattern of ‘double conserved synteny’, meaning that one chromosomal region in humans matches two in pufferfish, across the entire genome. This is a clear echo of whole-genome duplication in the ray-finned-fish lineage. Every gene, on every chromosome, was duplicated, although there has since been a massive degree of gene loss and local gene shuffling.

When did this whole-genome duplication occur? Analysis of zebrafish genetic maps strongly suggests that this species also underwent such an event in its history4. Pufferfish and zebrafish belong to distinct taxonomic orders of fish, so the duplication must have occurred early in teleost evolution. As previously pointed out7, this implies that traces of the ancient whole-genome duplication should be found in more than 20,000 species of living teleost fish. But teleosts do not make up the whole of the ray-finned fish. Significantly, a study of one of the Hox-gene clusters of an earlier (more ‘basally’) branching ray-finned fish, Polypterus (Fig. 1), found no evidence of a genome duplication9. Together with data from other basal actinopterygians10, this suggests that the genome duplication occurred close to the origin of the teleost fish themselves, perhaps 230 million years ago.

Less clear are the biological consequences. It is tempting to suggest that the species richness of the teleosts is somehow related to the whole-genome duplication, either because natural selection has ‘exploited’ the extra genes, or because differential mutation of duplicate genes caused reproductive isolation, facilitating speciation11. However, much of teleost diversity is found in just one group, the acanthopterygians (‘spiny fins’), which underwent a massive increase in diversity only around 55 million years ago. So if the whole-genome duplication did affect species richness, it was not immediate, and further studies of morphological and genetic evolution in teleosts will be needed to resolve the mechanisms involved.

A final lesson from the Tetraodon study2 concerns the power of comparative genomics, both for gaining insights into mechanisms of genome evolution and for deducing genome organization in extinct species. But we have a long lineage of extinct ancestors, which means that a wide range of genomes will need to be compared if we want to look at each node in our evolutionary history (Fig. 1). Particularly useful will be complete genome sequences for a shark, a lamprey and amphioxus, as each will provide insight into yet more ancient ancestral states. We may not have long to wait: this year the Joint Genome Institute in California began sequencing the amphioxus genome, while the National Human Genome Research Institute has announced plans to sequence that of the sea lamprey.