Free & Open Environment for NGS analysis: Galaxy (http://usegalaxy.org)

[nekrut]

The Galaxy team is announcing the launch of the first free public resource for NGS analysis at http://usegalaxy.org. This service is the beginning of our campaign to provide free web-based utilities for NGS analysis that later in the year will take advantage of Cloud resources (see http://bit.ly/aMUkpo).

At present there are three main groups of tools including (you can find them in the left pane of http://usegalaxy.org):

1. NGS QC and manipulation - contains a variety of tools for dealing with all flavors of fastq datasets as well as outputs of SOLiD and 454 instruments.
2. NGS Mapping - currently includes bowtie (Illumina & SOLiD), BWA (Illumina), and lastz (454) mappers. PerM (SOLiD) is on the way and more will be added in the coming months. Transcriptome tools (e.g., top-hat) are also in the final stages of development.
3. NGS SAMTools - includes a variety of utilities for SAM/BAM manipulation. Some are based on the samtools library, some are written by the Galaxy team.

The Galaxy team does not like to read documentation and expects that others don't either. This is why we make short movies called quickies. To see what Galaxy can do, see these:

Example 1 - mapping mate-paired SOLiD data
Example 2 - mapping SOLiD (or Illumina) data against a custom genome
Example 3 - mapping paired-end Illumina run and visualizing results at the UCSC Browser

Enjoy and send us feedback!

Those who want to contribute tools, brains, or coding skills should consider attending the Galaxy Developer Conference (http://www.galaxyproject.org/dev2010; Cold Spring Harbor Lab; immediately after the Biology of Genomes) by e-mailing us ([email protected]). We might sponsor your participation!

This free service is brought you by NIH (NHGRI), NSF, the Huck Institutes for the Life Sciences and the Institute for CyberScience at Penn State University, Emory University and the Pennsylvania Department of Public Health.

[ECO]

This is amazing by the way. If there is anything the site or community can do to support this effort, don't hesitate to let me/us know.

Galaxy =

[thinkRNA]

This is simply mind-blowing and must be a lot of work. I love the videos. Now, I can think of at least 5 companies that are going to be put out of business because they are charging 1000s of $ for this service.
Are you all thinking about including other analysis tools into Galaxy like cufflinks software suite, Degseq and Eland/Erange, other R statistical packages in the near future (if so,by when?)

[dawe]

Those who want to contribute tools, brains, or coding skills should consider attending the Galaxy Developer Conference (http://www.galaxyproject.org/dev2010) by e-mailing us ([email protected]). We might sponsor your participation!

See you there!

[KevinLam]

Is it a typo that BWA doesn't support SOLID or does Galaxy have no support for BWA on SOLID reads?

forgot to add I am blown away by the movies too.

[maubp]

Looks very impressive

I have one suggestion for enhancement: Right now, users would have to provide their Roche 454 data already as FASTA+QUAL or merged into FASTQ. One feature that should be fairly straightforward to add would be SFF to Sanger FASTQ (or SFF to FASTA, SFF to QUAL). I'd suggest looking at sff_extract for this (also in Python), which can handle paired end 454 data too.

Peter

[lh3]

I guess bwa's solid support was too buggy at the time of developing galaxy. It should become better now. For solid, I think it is important to include bfast as perm/bowtie do not do gapped alignment. Bwa does gapped alignment for SOLiD, but not as good as bfast. Now I think gapped alignment is crucial to accurate variant discovery, more important than I thought before. Several other publications have already emphasized this point.

[thinkRNA]

I guess bwa's solid support was too buggy at the time of developing galaxy. It should become better now. For solid, I think it is important to include bfast as perm/bowtie do not do gapped alignment. Bwa does gapped alignment for SOLiD, but not as good as bfast. Now I think gapped alignment is crucial to accurate variant discovery, more important than I thought before. Several other publications have already emphasized this point.

Are you talking about splice-variant discovery or snp discovery? I am guessing its the latter.

[nekrut]

Looks very impressive

I have one suggestion for enhancement: Right now, users would have to provide their Roche 454 data already as FASTA+QUAL or merged into FASTQ. One feature that should be fairly straightforward to add would be SFF to Sanger FASTQ (or SFF to FASTA, SFF to QUAL). I'd suggest looking at sff_extract for this (also in Python), which can handle paired end 454 data too.

Peter

Dear Peter:

Yes, this is on the to do list, and since we have 454s here this will be happening soon.

[nekrut]

I guess bwa's solid support was too buggy at the time of developing galaxy. It should become better now. For solid, I think it is important to include bfast as perm/bowtie do not do gapped alignment. Bwa does gapped alignment for SOLiD, but not as good as bfast. Now I think gapped alignment is crucial to accurate variant discovery, more important than I thought before. Several other publications have already emphasized this point.

At the time the challenge was the interpretation of SAM output produced by BWA on solid reads. It will be quite simple to enable BWA for SOLiD in Galaxy, since all cs indices are alreay built. Do you think it's time to enable BWA for SOLiD again? Speaking of bfast, we really need Nils' input on making index generation a bit more user friendly.

[nekrut]

Are you all thinking about including other analysis tools into Galaxy like cufflinks software suite, Degseq and Eland/Erange, other R statistical packages in the near future (if so,by when?)

Cufflinks by early summer. No plans for DegSeq yet (but integrating tools into Galaxy is easy, so anyone can do this). Speaking of Eland/Erange - we give priority to Open Source tools.

[KevinLam]

I guess bwa's solid support was too buggy at the time of developing galaxy. It should become better now. For solid, I think it is important to include bfast as perm/bowtie do not do gapped alignment. Bwa does gapped alignment for SOLiD, but not as good as bfast. Now I think gapped alignment is crucial to accurate variant discovery, more important than I thought before. Several other publications have already emphasized this point.

interesting revelation!
lh3: slightly OT. How would you compare bfast mapping vs bioscope's mapreads then?

Galaxy: I think you should ask ABI to write wrappers for their binaries in Galaxy. I think they should be more than happy to have better support for their platform. and Bioscope is still ^%*^% propriety &^(&*^ software. This seriously limits their widespread adaptation

[nekrut]

interesting revelation!
Galaxy: I think you should ask ABI to write wrappers for their binaries in Galaxy. I think they should be more than happy to have better support for their platform. and Bioscope is still ^%*^% propriety &^(&*^ software. This seriously limits their widespread adaptation

We prefer to stay on the open source side of the world = how else do you know what software actually does?

[nilshomer]

Speaking of bfast, we really need Nils' input on making index generation a bit more user friendly.

I believe before, we planned to design a script that would find the "optimal" parameters for building the indexes. The indexes in the manual work very well and support most types of sequence data (lengths and technologies) and genomes (long and short). Given this information, it should be trivial to get BFAST support up and running. Feel free to PM me or email me.

[KevinLam]

We prefer to stay on the open source side of the world = how else do you know what software actually does?

I think we can lobby for ABI to release as open source.
I mean they ain't selling the software anyway.

for ppl stuck with service providers its either you use corona lite which has very little documentation or u go with open source tools.

if no one uses ABI tools then what's the point?