How do you use Ruby on Rails for science (if applicable)?

We do research in systems biology. We prefer to use existing data sets, because collecting new biological data is expensive. Thus, a lot of the scripts we write are little more than transformations of one data set into another.

Eventually, we put our results online -- and more and more journals are requiring this sort of thing.

So it was no great leap for me to try using Rails for my projects. I can set up easily reproducible experiments, transform data step by step through database tables (e.g. using rake), and display results using gems like flotomatic and the gnuplot. If I need something to run very quickly, I can even write a custom gem in C++ using Rice, or parallelize using starling and workling.

Eventually, I started to wonder if anyone else was using Rails to do bioinformatics or science in general.

I thought, "If I were a sci开发者_运维问答ence research Rails gem, what would I do?"

What extra features would such a gem have? Perhaps a Migration adaptation into a rake-able pipeline? Maybe more advanced graphing features? Built-in background jobs?

For Bioinformatics, see http://bioruby.org

I agree with Pierre. I think bioruby is the right place. Many (most?) bioruby users/developers use rails, making rails a natural extension to the bioruby project.

Here's an incomplete list of bioruby code for rails:

• Run bioruby console in a rail app - http://bioruby.open-bio.org/wiki/BioRubyOnRails

• ActiveRecord (Rails default ORM) classes for Ensembl - bioruby-annex.rubyforge.org/

• Plugin for Uniprot db - bioruby.g.hatena.ne.jp/nakao_mitsuteru/20070410/uniprot_on_active_record_plugin

• CHADO/Bioruby integration effort - github.com/robsyme/RubyCHADO

sorry about the messed up links, but as a new user I can't post more than a single link :(

I've used Michael Barton's organized_experiments before. It works quite nicely once you replace DataMapper with ActiveRecord.