Galaxy Community Conference 2016 (GCC16)

→ Slides    doi:10.7490/f1000research.1112573.1
→ Video

Authors
Virginie Rossard, INRA-LBE
Fabien Gogé, IRSTEA Montpellier
Eric Latrille, INRA-LBE
Jean-Michel Roger, IRSTEA Montpellier
Jean-Claude Boulet, INRA-SPO

Abstract
Infrared spectroscopy is widely used in academic research and industry as simple, fast, cheap and safe measurement tool. The infrared data are displayed as spectra, and chemometric is a science which aims at extracting informations from spectra.

We are developing a comprehensive package which contains (1) a MOOC broadcasted in september 2016; (2) a chemometric tool, named ChemFlow, which is an application of Galaxy; and (3) a spectral database. We will focus on ChemFlow.

The required specifications were:

• a free tool;
• a tool which recycles code from Matlab, Scilab, R, Python and C;
• a tool accessible via internet with new devices such as smartphones.

That's why we chose Galaxy. ChemFlow is being implemented with our own functions. By now it includes most of the processing tools : import and convert our data; run chemometrics methods such as calibrations and classifications.

We are very satisfied of the performances of Chemflow running on a server. Nevertheless, some issues were fixed, others are still pending:

• Speed performance was improved by switching the Galaxy server to Apache and PostgreSQL.
• Hundreds of users are expected. We plan to deploy 2 servers of 48-cores each, without knowing how ChemFlow will behave with many users submitting little tasks.
• The graphical toolbox in Galaxy is our main work in progress, and we are currently implementing several original visualisation tools such as R-Shiny.    
• The development of a specific toolshed is discussed.    

As a summary, Galaxy is used in a new domain, chemometrics, adressed to a new user community, and will be a central platform for a new e-learning module, as a MOOC.