/parallel-aracne

Primary LanguageR

mk-aracne

The pipeline run by CSB-IG to make aracne adjs with p=1.

This pipeline computes the superior diagonal of the adj only. You should transform your files to the format required for further analyses.

Running a job

Add to the data directory your files containing the expression matrices in the format:

genes <tab> sample1 <tab> ... <tab> sample_n

  • The first line must be the header.

  • The expression matrix filename(s) must end with .txt

Then execute:

bin/targets | xargs mk

The expression value of each probe will be colapsed by mean, alphabetically sorted and placed in results/001/.

The adj for each probe will be placed at results/002/ but may be joined into a single sif or adj.

You can edit the parameters to run aracne with by editing config.mk.

ADJ output {#adj}

To join every probe for an experiment into one file:

bin/target-adjs | xargs mk

Output will be added in the results/ directory.

SIF output {#sif}

To transform every probe for an experiment into one sif file:

bin/target-sifs | xargs mk

Output will be added in the results/ directory.

Cleaning intermediate results:

Requirements

  • mk: For running the jobs.

  • R 3.3: For colapsing the expression matrices by mean. All required libraries are provided in this repository.

  • awk: For knowing which jobs to run.

  • ARACNe: The whole point of this pipeline.

    The binary file should be on your PATH and be named aracne (not aracne2). As the binary needs modifications to be installed correctly on your PATH, you can use a script containing:

    #!/bin/sh
/path/to/aracne2 -H "@"

    We use aracne 2 as published as of march 2017: 87ab986280492fab1058b9f8bfb411d5 ARACNE.src.tar.gz