From: Mitch S. <mit...@be...> - 2007-08-24 02:27:06
|
Ian Holmes wrote: > So I keep posting this sort of thing even though I really don't know how > relevant these things are, and of course we do have a long list of more > practical immediate things to do (like, um, getting something that we > can seamlessly use in place of an existing GBrowse instantiation, > ideally before modENCODE is history...) > I like the game angle; it seems like it could be a pretty rich vein of new UI ideas, which I get a big kick out of. However, if at all possible I'd like to start moving things in a more concrete/applied direction. Like collecting, for example. We've talked about collecting things like tracks, but I think a lot of biologists would also be interested in collecting individual features (genes). That's something that the biologists that I've worked with liked to do a lot (then they'd send me Excel files full of gene symbols that weren't necessarily the standardized symbols). Then I can also imagine them wanting to see SNPs near their genes, regulatory elements near their genes, etc. I've been thinking that I'd like to make track definitions just be queries on a genomic data source (plus some information on how to render the result). That's pretty much how the gbrowse config works right now; each track is defined in terms of the type of the feature to be displayed in it. But I'd like to also allow for queries that are more general than just by type, for tracks like "the genes I'm interested in" or "SNPs near the genes I'm interested in", "SNPs near the genes that are active during development phase foo", "genes that are under GO term such-and-such", etc. Mitch |