[Bioperl-l] slippy map for genomic features.
mitch_skinner at berkeley.edu
Thu Jul 5 17:22:38 EDT 2007
FWIW, we've been working on something similar:
scratch (with the prototype library). When our project was starting up
(fall 05) Andrew looked but didn't find openlayers; I'm not sure if it
was public back then but their current svn only goes back to 2006.
I think that things like layout (bumping) ought to be done in advance on
a chromosome-wide basis; otherwise it's difficult to keep features from
ending up at different heights on neighboring tiles. And it would be
difficult for the server to know what was being clicked on. So we've
been doing some up-front work to either do layout or to just render all
the tiles in advance:
which is driven by this script:
Or you could just not bump at all, I guess. I think of that as
important functionality but I'd be interested in hearing about use cases
where it's not necessary. It's not just bumping, though; things like
text labels also make it difficult to predict exactly what pixels a
feature will span if you only have its genomic coordinates.
To make features clickable we've been using imagemaps; it simplifies the
server code but it bogs down the client quite a bit.
I'd certainly be interested in seeing if there are ways we could work
together; if you're at Berkeley maybe we could meet.
Brent Pedersen wrote:
> here's a side project i've been tinkering on in googlecode svn that
> may be useful to some.
> it's a simple hack on top of OpenLayers (openlayers.org) to provide a
> features. It can be used with any image generation program that can
> accept &xmin= and &xmax= parameters through the url. -- though i
> havent had it working it bioperl as bioperl generates images of
> different height depending on the number of tracks.
> there's a live example of the code in SVN here:
> with images generated by a colleague's modules on first request. those
> images are then cached by a simple perl script included in the SVN
> repo. all subsequent requests are returned from the cache.
> but any implementation need only implement xmin and xmax. all other
> parameters will be used for caching but are not required.
> if anyone is interested in getting this going with bioperl image
> generation--or improving the project in any way, let me know and i'll
> tar ball download:
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
More information about the Bioperl-l