[galaxy-dev] New tools: bigWig & bigBed
gordon at cshl.edu
Mon Oct 5 21:53:21 EDT 2009
I'd like to suggest couple of tools to handle bigBed and bigWig formats.
The bigBed and BigWig formats are new to the UCSC genome browser,
and allow efficient uploading of user tracks.
They are efficient not only because they are binary, but also because they don't need to be uploaded to the genome browser itself - the binary format support partial HTTP gets, so the bigWig/bigWig file can stay on your server, and the genome browser fetches only the relevant portion of the file (for the current viewing coordinates).
The tools are:
I provide only the wrappers, the tools themselves are available for download from the UCSC genome browser.
You can see example of the running tools here:
(Bottom category - "bigBed/bigWig tools")
You can download the XML wrappers from here:
I'd like to point three advantages to the bigWig/bigBed formats:
1. They go hand-in-hand with next-gen sequencing files:
A user uploads a huge FASTQ file, maps it with BWA/Bowtie (or other)
and wants to visualize the results. Uploading the resulting textual BED/WIG file is too wasteful.
2. With some minor Galaxy modifications, adding custom tracks to the Genome Browser can be done easily, even faster than clicking on the "main" link of a BED file:
Add to the dataset controller a method that returns something like:
track name="User Track" type=bigWig bigDataURL=http://main.g2.bx.psu.edu/datasets/837890/display/index
And point the genome browser to this URL - it will take the data directly from Galaxy - no need to upload anything.
3. The BedGraph format (which consists of chrom,start,end,value) is exactly what the 'fastx-collapser' was written for: assigning a multiplicity count for an interval...
But many other mappers/aligners programs also provided their own "collapsed" or multiplicity-value for intervals - this makes presenting data one the genome browser very easy.
More information about bigWig & bigBed:
Comments are welcomed,
More information about the galaxy-dev