-

SolutionChallenge

From Open Bioinformatics Foundation
Revision as of 01:54, 22 February 2010 by Kdahlquist (talk) (added back link to BOSC 2010)
Jump to: navigation, search

Back to BOSC 2010

OpenBio solution challenge: Project updates at BOSC 2010

One of the traditional items we've had at the conference is project updates from open source bioinformatics libraries, like BioPerl, BioJava, Biopython, BioRuby and EMBOSS. This year, we're planning to organize these talks around a central theme: the OpenBio solution challenge. We start with a biological question of general interest, and project talks focus around how you would solve that problem using your toolkit and programming language.

This is meant to provide a challenge for OpenBio contributors, a nice tutorial style overview of various projects and approaches for other programmers, and a fun opportunity to compete and learn from other projects. Conference attendees will vote on their favorite solution, with the winner receiving fame and fortune (warning: fortune not guaranteed).

The challenge is open to any interested participants. Please submit an abstract for BOSC with information about your project and that you are interested in participating in the solution challenge session. Introduce yourself and your ideas on the Open Bio mailing list as well.

Guidelines

  • The goal is to pick a challenging problem that demonstrates the utility of a bioinformatics approach to answering the question, not necessarily an unsolved problem like the Netflix Prize.
  • Feel free to use any publicly available programs and libraries in solving the problem. If you are solving a problem in Python, we certainly want to see how the language and toolkits help you do that most efficiently, but re-writing BLAST in Python makes no sense. How you integrate the external programs into your automated analysis pipeline is of extreme interest.
  • We can select from one overall challenge or a panel of challenges. We'd like to provide some overlap to help compare and contrast various approaches, but some challenges may fit better with a focus of a toolkit then others.
  • The winner will be selected by conference participants; everyone who sees your talk gets a vote. The criteria conference attendees should vote on are
    • Execution of the analysis. How easy is it to understand and replicate?
    • Quality of the answer. What interesting biological was discovered?
    • Uniqueness of the approach. What elements of the toolkit make it useful for the problem?

Project ideas

These are collected project ideas from various members. Please create a wiki account for yourself and add your own ideas. As a group we will decide on one or more challenges to use:

  • Fly RNA-seq gender analysis. Start with short read alignments in SAM format from the modENCODE project. Alignments of total RNA from both male and female flies are available. Your goal is to use these alignments to highlight major expression differences between the genders. This will involve organizing alignment locations relative to genomic features, identifying differences, and providing summary charts and tables that explain the most interesting results.

Discussion

Your thoughts and project ideas are very welcome. Please use the open bio mailing list as a central point of discussion: open-bio-l@lists.open-bio.org