Welcome to Eoulsan

We developed Eoulsan, a versatile framework based on the Hadoop implementation of the MapReduce algorithm, dedicated to high throughput sequencing data analysis on distributed computers. With Eoulsan, users can easily set up a cloud computing cluster and automate the analysis of several samples at once using various software solutions available. Working either on standalone workstations or cloud computing clusters, Eoulsan provides an integrated and flexible solution for RNA-Seq data analysis of differential expression.

What's new in Eoulsan 2.3?

Eoulsan 2.3 has successfully passed our integration tests and will generate the same outputs obtained with Eoulsan 1.2.2.

Warning: this documentation may not be up to date for the latest features of Eoulsan 2.x.

The main new features in this version of Eoulsan are:

  • New workflow engine
  • New steps:
    • DESeq2
    • FastQC
    • MultiQC
    • sam2bam
    • splice2bed
    • bam2sam
    • sam2fastq
    • shell
    • expression additional annotation
    • diffana additional annotation
    • import
    • technicalreplicatemerger
    • Bioconductor SingleCellExperiment object creator
  • Separate log file for each step
  • Integration tests engine
  • Support for multiple versions of a mapper
  • Rewritten mapper launcher
  • New JSON format for step result file
  • Docker support
  • New mappers added: STAR, Minimap2
  • Support of flavors of mappers (e.g. standard and large index binary versions of Bowtie 1, Bowtie 2 and STAR)
  • Update of the bundled versions of the mappers: Bowtie (1.1.1), Bowtie2 (2.2.4), BWA (0.7.12 and 0.7.15), GSNAP (2014-12-21, 2017-02-25 and 2017-04-24) and Minimap2 (2.5 and 2.10)
  • Hadoop 2.x support
  • Support of many job schedulers: HT-Condor, SLURM, TORQUE, PBS Pro and TGCC
  • Galaxy tool XML file can be interpreted as an Eoulsan step
  • New enhanced design file format
  • New step for normalization and differential analysis using DESeq2
  • R code can be executed on local system, on a Rserve server or in a Docker container
  • New Eoulsan actions: info, modules and formats to get the Eoulsan configuration and the list of available modules and formats.
  • A "eoulsan-data" directory is now created when no repository for mapper indexes or genome description are defined. This new directory avoid to relaunch the index or genome description computation.
  • Support for Java 8, 9 and 10.
  • The Eoulsan HTSeq-count implementation has now the same features as the original HTSeq-count.
  • And many other enhancements, see the full changelog for more information

You can now easily run Eoulsan using Docker with a dedicated eoulsan-docker.sh script that automatically configure your Docker environment like Eoulsan resources mounts. For more information about Docker and Eoulsan see the the related Eoulsan documentation. To install this script just copy and paste the following command line in your terminal:

$ curl http://outils.genomique.biologie.ens.fr/eoulsan/eoulsan-docker-installer.sh | bash

Only few Eoulsan modes are currently working with the current version. The following table summarize the status of each mode:

Eoulsan Mode Action name Status
Local mode exec Working
Hadoop mode hadoopexec Working, Hadoop 2.x is required
Cluster mode clusterexec Working
Grid mode gridexec Planned, prototype works, waiting the Hadoop mode achievement
Amazon EMR mode emrexec Not working, dedicated code must be rewritten for the new workflow engine
Other cloud mode not defined Planned, implementation not started

Note: Eoulsan 2.x is not binary compatible with Eoulsan 1.2.x, so step plug-ins written for previous versions will not work this release. The Eoulsan 2.x API is currently almost stable and small breaking compatibility may occurs in future.

Screencast demo

The following screencast show how to install Eoulsan, create a design file and launch an analysis in local mode. For more information see the quickstart guide.

RNA-Seq analysis

Eoulsan contains out the box a fully customizable RNA-Seq pipeline that generate many reports about the performed analysis. The following images are samples of the graphics generated by the normalization and differential analysis steps:

This version currently use DESeq 1 and DESeq 2 for normalization and differential analysis but support for complex designs is only available with DESeq 2.

Chip-Seq analysis

Chip-Seq workflow support is currently under development, it will available in a future version of Eoulsan 2.x.

Object model

Eoulsan contains also a java object model that can be used outside the Eoulsan workflow engine to handle and manipulate NGS data. See the Eoulsan developer wiki for more informations about this part of the project.


Eoulsan is distributed under the Lesser General Public Licence and CeCill-C.


This work was supported by the Infrastructures en Biologie Santé et Agronomie (IBiSA), France Genomique and Amazon Web Services research grant.


Subscribe to Eoulsan RSS feed