© (2002) Damian Counsell.

Fugu rubripes, a small South-East Asian fish, was proposed ten years ago as a genomic model vertebrate whose small genome could aid the understanding of vertebrate genome evolution and the functional study of complex genomes. Aparicio et al. now report the sequence of ∼95% of the Fugu genome. Their analysis hints at the mechanisms that might be responsible for the compactness of Fugu's genome, whereas comparative genomics in the human and Fugu reveals aspects of genome and protein evolution.

At 365 Mb, Fugu's genome — sequenced to around sixfold coverage using the shotgun method — is one-eighth of that of humans. This difference is mainly due to smaller intergenic regions and introns, which seem to be kept trim by frequent deletions. In the absence of experimental data, annotation of the genome was homology based, and ∼33,000 genes have so far been predicted — closely matching the number estimated for humans. As in the human, gene density varies across the Fugu genome and, because of smaller introns, most Fugu genes are smaller than their human orthologues, although the intron–exon structure of most genes is preserved. However, some 'giant' genes have also been found, which the authors speculate might have a role in the balance between gain and loss of DNA and therefore in genome evolution. Although many segments are conserved between Fugu and human genomes, gene order has been considerably scrambled.

Differences between Fugu and human proteomes reflect differences in physiology, but also highlight systems that are rapidly evolving in humans, such as T-cell-mediated immunity. Although there are more similarities than differences, ∼25% of the human proteome does not have a Fugu counterpart.

So, Fugu has passed the test as a genomic model organism, and this first analysis of the whole genome sequence has already provided a wealth of information on vertebrate genome and proteome evolution.

Fugu's genome is the first publicly available vertebrate genome sequence after that of humans, and it is the first large genome to be sequenced that didn't rely on a physical map. But maps aren't only useful for genome sequencing, they are a valuable resource in experimental models. Although the sequence of the mouse genome isn't yet complete, Gregory et al. report the generation of a mouse physical map.

The authors assembled 296 contigs of overlapping BACs (∼9.3 Mb long), thereby providing a nearly complete coverage of the estimated 2.9-Gb mouse genome. When 97% of the mouse BAC map was aligned to the human genome, 88% of mouse clones were collinear. The authors used SSLP and radiation hybrid data to assign 203 clones to individual chromosomes, generating an invaluable resource for the mouse community. Even before the mouse genome is completed in 2005, we are bound to hear of important insights that come from the analysis of this physical map and its alignment to the human genome.