README.md

In elqumsan/RNAseqMVA: Application of Multivariate Analysis methods to RNA-seq Data

RNAseqMVA

R package developed to perform an evaluation of multivariate analysis methods for the supervised classification of RNA-seq data.

Authors

• Mustafa AbuElQumsan
• Jacques van Helden ([e-mail])

Downloading

git clone https://github.com/elqumsan/RNAseqMVA.git

Conda environment

The file conda-rnaseqmva.yml specifies the parameters of a conda environment enabling to install all required dependencies (R, RCRAN and Bioconductor libraries).

Installing miniconda

If conda is not yet installed on your system, follow the conda installation instructions.

Loading conda module

This step is specific to the IFB core cluster. If you are working on another infrastructure, you can skip it.

On the IFB core cluster, conda is loaded via a module, so this command must be run before starting conda

module load conda ## Load the conda module (for the IFB-core-cluster)

Installing the conda rnaseqmva environment

The file conda-rnaseqmva.yml specifies all the requirements to run RNAseqMVA. The simplest way to use the RNAseqMVA package is to create a conda environment that will contain all the dependencies. This can be done automatically with the following commands.

cd RNAseqMVA
conda env create -f conda-rnaseqmva.yml

Note: this command needs to be run only once. The next section explains how to update the environment after it has been created.

Updating the conda rnaseqmva environment

In case of changes to the RNAseqMVA environment, an update can be achieved with the following command.

cd RNAseqMVA
conda env update -f conda-rnaseqmva.yml

Loading the environemnt

The next command needs to be adapted depending on your conda version.

conda --version

If your version is < 5, use the command source below.

source activate rnaseqmva

If you have conda verison >=5, you can run conda activateas below.

conda activate rnaseqmva

Compiling the RNAseqMVA package

We assume here that the RNAseqMVA package has been installed in the home directory. If not, you just need to adapt the path below.

cd ~/RNAseqMVA
make build_and_install

Configuring the analysis

All the parameters of an analysis can be specified in a YAML file misc/00_project_parameters.yml. Parameters can be changed easily by editing this file with any text editor (nano, gedit, emacs, vi, ...).

Running all analyses

Rscript --vanilla misc/main_processes.R

Running selected analyses

R --vanilla

Then open the file misc/main_processes.R and decide whether you need to run each command.

Specific settings for the IFB cluster

The cluster of the Institut Français de Bioinformatique (IFB-core-cluster) was used to run comparative assessment of supervised classification methods for RNA-seq.

On this cluster, conda is loaded via a module.

module load conda ## Load the conda module (for the IFB-core-cluster)

after that, the RNAseqMVA environment can be loaded as described above.

Commands are sent to cluster nodes via srun.

srun --mem=32GB Rscript --vanilla misc/main_processes.R

elqumsan/RNAseqMVA documentation built on March 10, 2021, 8:10 a.m.