#158 implement genomespace specs for IGB and genopub

[ido_m_tamir]

For the long term viability of the IGB it would be good to integrate with the gneomespace initiative.

www.genomespace.org

source code is here:
https://bitbucket.org/GenomeSpace/combined

maybe genopub could implement the data serving API.

[David Nix]

Possibly... Hopefully they are taking advantage of the work that went into DAS/2 and not reinventing the wheel.

There is definitely a need for the community to adopt a common protocol for data exchange. So far I see little effort along these lines by the major genome browser groups.

People are still building islands of data instead of bridges.

[Ann E. Loraine]

Thanks for the pointer!

Have you looked much at the data serving API? Is there an aspect that seems particularly promising or useful?

I would be particularly interested in whether the API is going to support alternative formats for data delivery in response to SOAP or ReST style queries.

I mention this because one aspect of DAS2 that has worked well for us is is that the DAS2 servelet/Genopub code can send data in compact file formats (e.g., psl, bed) that IGB can read easily. By contrast, DAS1 data sources send data in XML, which for massive data sets like RNA-Seq, is not practical. For sending meta-data about data sets and genomes, XML works great, but for sending the raw data itself, I don't know if XML or JSON is a good idea. Does it seem like the genomespace API is heading toward using JSON or other verbose formats to send data, or will they allow other formats? I looked at some of the code and it's looked like JSON is being used as the data exchange format, but I could be wrong on that.

Since the project is just starting, we may have a chance to influence the direction of development.

Best,

Ann