Standard file format for profiles?

[dawe]

Hi all,
I would like to share an issue I'm facing... I'm working with users who like IGV (from broad institute), others who like IGB (genoviz) and other simply use the UCSC Genome Browser to visualize profiles for ChIP-seq experiments.
It happens that there's no standard file format for all three platform.
Well, they all support bedgraph but this is definitely unsuitable for NGS data. IGV has TDF, IGB has Useq, UCSC has bigWig... is there anybody working to propose a standard (in the same way SAM has been introduced) ?

d

[malcook]

bam bam bam bam

BAM format is supported by IGV as well as by UCSC and is just coming out in recent version of IGB

[lh3]

All these formats (tdf/bigwig/useq) have their own advantages. I guess it is hard to unify them. If I were asked to design a format, I would probably write a program and a C+Java library to index and achieve random access to a bgzip'ed wig file. It will have the following advantages: a) wig is the most widely used; b) bgzip'ed wig is small in size; c) the C bgzf library supports remote access (essential to UCSC); d) the indexing scheme should be easy enough to be implemented in different languages.

In addition, I know bigWig support is on the todo list of IGV. Once this is done, we will have a java library and bigwig will be quickly adopted by those java based viewers. BigWig is both the most technically advanced and the most technically challenging.

[dawe]

BAM format is supported by IGV as well as by UCSC and is just coming out in recent version of IGB

BAM reports the alignments, not the profiles on a chromosome. You may say that BAM->Coverage is straightforward (and this function is included in IGV) but one (me) can obtain profiles after a huge processing of alignment data...

[lh3]

Have you noticed this paper? Is it good?

http://bioinformatics.oxfordjournals.org/cgi/content/short/btq164v1

[dawe]

Have you noticed this paper? Is it good?

http://bioinformatics.oxfordjournals...short/btq164v1

Nice, thanks! I'll give a try, will the genome browsers do? :-)

d

[quinlana]

Have you noticed this paper? Is it good?

http://bioinformatics.oxfordjournals...short/btq164v1

Thanks for pointing this paper, Heng. Looks pretty useful.

[dawe]

It's hard to build if one has hdf5 and pytables already installed...