Steven Salzberg has an opinion piece in Genome Biology that talks about Wikis for managing genome annotation and the problem of bit rot in gene annotation (would that be annotation rot?).
I wrote a quick post about it. This harkens to the other gene wiki discussions on nodalpoint as well as some of the things Ian Holmes and his group are thinking about on the GbrowseAJAX mailing list. I know others are also talking about this. I guess there is a wiki for gene function and perhaps another for genome annotation. Tracking it all seems like it will be a big job. Just trying to track automated gene names across different assemblies and gene builds like the Ensembl team has to do is hard enough!
Comments
Good ideas, now we need details
I think the idea of a wiki for collaborative annotation appeals to a lot of people. The discussion around your post highlights the challenge: we have to go beyond the idea to the details of the implementation, i.e. the software engineering.
It would be hopelessly inefficient to rely entirely on manual annotation and the efforts of users. A wiki would somehow need to capture data from existing sources (sequence databases + automated annotation) and present it to the user for editing. There would have to be rules (ontologies) and a way to agree on a final annotation (voting? approval by an expert?)
I guess we're looking for a combination of the functionality of the IMG, the ideas of manatee (but not the implementation) and the philosophy of wiki.