2012-01-23

Ion Torrent Suite on GitHub

Good news - the Ion Torrent Suite is now freely available open source software on GitHub under the GPL v2 licence, as promised late last year. There is now something more substantial behind talk of Ion Torrent "democratising sequencing", and a clear advantage over the closed source tools of rival companies. I commend them!

The Ion Torrent GitHub account was registered on Friday 13th January 2012, and the Torrent Mapping Alignment Program (TMAP) added (GPL v2 licenced).

A few days later they added TorrentScout, an interactive Java tool to access to raw data generated by the Ion Torrent Personal Genome Machine (PGM). Curiously this doesn't have a clear license statement, nor even a readme file. Since at the time issues (GitHub's bug tracker) wasn't enabled for this repository, I filled these as a bug on the parent repository, Chantal Roth's TorrentScout repository (see missing readme, missing license).

Then on Monday 23rd Jan, we got the Ion Torrent Variant Hunter (GPL v2), and what we've been waiting for most of all - the main Ion Torrent Suite (GPL v2, some third party code also under LGPL or BSD). Hooray!

3 comments:

  1. You were too quick :-). All files are there now (including all documentation). Also, all class files have the license file included on top ;-).
    Cheers,
    Chantal

    ReplyDelete
  2. Nice.

    I still have a few questions (and "you" does not refer to the owner of the blog but Ion Torrent / Life Technologies / Innocentive, who dont allow me to post anything without agreeing to T&C):

    *where can I download the Ion Torrent example data set without signing T&C? not exactly open

    I am interested in the accuracy challenge, however since I dont feel like signing T&C, I would prefer not to enter their challenge officially, but still interested in beating the challenge with the open source coli data on github (from the outbreak in europe).

    *suppose I changed part of the signal processing, what sequence of commands from which tools should I execute to check my if my results reach the benchmark? one of my (and I think I am not alone) conditions before participating in such a contest would be to have a failproof way to check if my changes pass the required performance. (of course cheating by changing the code of phred result calculation etc would always be possible, however an honest contestant would not touch those parts of the code, and could get a good view of his results if the sequence of steps to verify the changed code's performance) so what piece of software or what sequence of commands can I use to churn out the result to compare with Ion Torrents performance requirement for the accuracy challenge?

    Now that the code is open-source, if an open-source community can solve this problem using the open-source E.coli dataset from the outbreak and beats
    *the closed-source Ion Torrent develloper community using multiple closed-source datasets
    *the open-source community who submit themselves to the T&C to view the single closed-source E.coli data-set
    then this can send a strong signal to Life Technologies:
    If you intend to hold an open challenge, open up every part necessary for the challenge, it is only as open as the most closed part. Most companies exclude their employees from challenges specifically for this reason.
    Once employees are excluded you can make the challenge fair for all participants (who do or do not submit to the IP/trade secret clausules).

    It may seem like I insist on opensourcing the whole company, this is not true. for example if you wish to keep specifics hidden, keep them hidden. It is not in the providing or not providing of source code, or data sets that you promote democracy. For whoever (from wherever in this world) beats the limit, it is in the financial (prize) gradient that you insurrect over the signing or not signing of the T&C...

    ReplyDelete
    Replies
    1. Perhaps a thread on SEQanswers.com would be a good place to discuss this openly?

      Delete