As I have mentioned before, de novo sequencing of whole eukaryotic genomes may be a thing of the past (or, at least, these whole genome projects won't be getting very much more common). Instead, I proposed that people would use the new high-throughput technologies to sequence parts of the genome they found interesting. What did I propose they'd sequence (based on discussion I'd had with various folks)? The transcriptome, or, essentially, a whole genome cDNA library. This allows for a much higher coverage of the genes in the genome, but sacrifices information regarding non-transcribed sequences and gene order.
It turns out that sacrifice is one worth making. In the April issue of Molecular Ecology, Vera et al. report that they used 454 to sequence the genome of the Glanville fritillary butterfly (doi:10.1111/j.1365-294X.2008.03666.x). This organism is a popular model in population biology, but had no genetic resources. That's changed thanks to this project. Here's the punchline from the abstract:
We conclude that 454 sequencing, when performed to provide sufficient coverage depth, allows de novo transcriptome assembly and a fast, cost-effective, and reliable method for development of functional genomic tools for nonmodel species. This development narrows the gap between approaches based on model organisms with rich genetic resources vs. species that are most tractable for ecological and evolutionary studies.
Molecular Ecology also has a review of the paper (doi:10.1111/j.1365-294X.2008.03699.x).
nice post. I was thinking about commenting on this paper myself. I suspect that it is among the first of many such efforts. You are right that sequencing and SNP discovery are definitely going to move in this direction in many species. Though working out where these sequences lie on the physical and genetic map (if that knowledge is needed) is going to potentially be a serious bottleneck in the process.