Will the Wikipedia model work for genome annotations?

The annotation of most genomes becomes outdated over time, owing in part to our ever-improving knowledge of genomes and in part to improvements in bioinformatics software. Unfortunately, annotation is rarely if ever updated and resources to support routine reannotation are scarce. Wiki software, which would allow many scientists to edit each genome's annotation, offers one possible solution.

(this abstract and full article can be found on the Genome Biology website.

Comments? Discussion?
This is, after all, a discussion board.

Like it or not, wikis are coming to biology (and everywhere else) in a big way. For a preview of whats coming, have a look at the as-yet un-announced new version of UniProt at http://beta.uniprot.org . For each protein entry, a link is established to a page in the (not yet operational) Wiki Proteins semantic wiki. If you watch the demo on wiki proteins, note the extensive interconnection with the major players in bioinformatics (e.g. genbank).

It will be interesting to see how they handle control over their users (who gets to say what..) which I think is a fundamental question right now all over the social web.

It's true. And the wiki mentality didn't beat Genomics to this concept. The idea of open annotation has been working in the field for years. My guess is that the new tools coming out that make the web "2.0" will just make this practice more advanced and widespread.

I have to disagree with you on the idea that open annotation has really been with the community or whether it ever really will be (or should..). In the article you site below, the annotation strategy (for the honeybee genome) was quite carefully orchestrated and limited to a (Web-scale) very small number of people. A quote from the article:

OK, here is a citation that might better model Wikipedia, though without the hired editors that Wikipedia employs. In a perfect world, peer review by the community would be enough... in a perfect world.

PeerGAD article

It still doesn't model wikipedia very well - but is closer ;). I like their peer-review pattern, though calling it a 'peer-review' pattern rather than a templated run through a 'circle of trust' or something makes it seem a little antiquated. Pointing out that Wikipedia does in fact have 'hired editors', while true, is misleading. They are clearly not the primary content generators nor content editors.. they simply coudn't possibly keep up.

I'm not interested in perfect worlds, I believe that, if they are designed appropriately, open systems will succeed in our own exceedingly imperfect world. However, I certainly don't think that we've figured out how to design such systems very well just yet. Trust is the fundamental thing to figure out... (incentive and interface are next on the list).