[Gmod-help] GMOD component contribution

[Joachim Baran]

Hello Hilmar,

  Actually, we met at the BioHackathon 2011 in Kyoto.

On 2012-05-06, at 5:15 PM, Hilmar Lapp wrote:
> Nice stuff! However, there are actual criteria to be met (and a process) for software to become a GMOD member project:
> 
> http://gmod.org/wiki/GMOD_Membership#Software
> 
> Could you elaborate on how these prospective components would meet the requirements?
  - Meets a Common Need
    Yoctogi is a data agnostic backend that can expose a wide variety of data formats through a number of generic interfaces. An example is provided for loading and accessing VEGA's GFF files, but Yoctogi is not limited to this data format. Users/developers can also make use of Yoctogi's plug-in system to extend the software with additional query interfaces or database backends.
    opacmo is a text-mining pipeline that is tailored to work with the complete open access subset of PubMed Central and a range of gene/species databases and ontologies. This can be interpreted as a common need already, but opacmo's small and commented scripts can also be easily customised and extended for other needs. A plug-in system for opacmo is planned, but not implemented yet.

  - Useful Over Time
    Yoctogi's plug-in system makes the software flexible enough to deal with emerging technologies. It also includes Semantic Web support, which will be a major driving factor in bioinformatics over the next decade.
    opacmo's design purpose was to have a reproducible and scalable text-mining pipeline. The opacmo web-site provides three or four updates of its PubMed Central text-mining dataset per year, where the actual pipeline has been extended to work with Oracle Grid Engine clusters and its Amazon EC2 implementation is currently in the making (see the forked GitHub repo).

  - Configurable and Extensible
    Yoctogi can be configured in many ways and an example configuration is provided that shows how the many interfaces of Yoctogi can be set up. It also has a plug-in architecture that permits custom components to be added to the system, where most currently implemented interfaces and the complete backend are already written as plug-ins (and hence, can server as implementation templates).
    opacmo's scripts offer a wide range of documented command line parameters for running the text-mining pipeline in various "configurations". Other settings can be carried out by modifying the commented parameters within the header of the scripts. A plug-in system is going to be implemented for opacmo, which will allow users to include other text-mining tools (Textpresso perhaps).

  - Open Source License for All Users
    Yoctogi's and opacmo's software is released under the MIT License.

  - Interoperable With Other GMOD Components
    Yoctogi's example shows how GFF data can be read into a Yoctogi database. Yoctogi supports asynchronous Galaxy requests, where documentation is provided on how to configure a Galaxy installation in order to work with Yoctogi. The live opacmo web-site has already been included in some Galaxy installations.
    It would be possible to provide further interfaces to other GMOD components through Yoctogi's plug-in system, but that would require further discussion of the direction that GMOD Components will take in the future. I have seen that there is a GMOD RPC API (http://gmod.org/wiki/GMOD_RPC_API) too, but it is not clear whether this API is still supported. The FlyBase URIs that I tried led to a "Page not Found" error. Chado support could be achieved too, but it is not clear to me whether the Chado schema are still actively being developed.

  - Commitment of Support
    Yoctogi and opacmo are privately funded as citizen science projects. I am therefore highly committed to continue these projects since they reflect the output of my spare time research.
 
  - Users and Support Mailing List(s)
    Yoctogi offers a mailman mailing list through the usual mailman interfaces. I can also set-up a mailing list for opacmo if it is a requirement for GMOD Components.

> I'll also add that the process is administered by the GMOD Community Support, which is currently vacant (although we're in the process of hiring a candidate). We may not be able to start the process for the projects you propose until someone is in that position again, and so you're unlikely to delay that if you take some time for reviewing the above and your responses.
  I actually expected to get a response from Scott, but I assume his duties are elsewhere. I am in no rush with getting Yoctogi and opacmo approved as GMOD Components and I am happy to wait until the post of a GMOD Community Support officer is filled again.

Best,
Joachim