Skip to content

Software Overview

GarrettJenkinson edited this page May 31, 2018 · 2 revisions

This repository contains the MATLAB, C++ and R software that implements informME, as well as the BASH wrappers and example submission scripts used to run informME on a Sun Grid Engine (SGE) cluster or a SLURM computer cluster.

informME is an information-theoretic approach to methylation analysis developed in [1] and [2], using the Ising model of statistical physics. This method is applied on BAM files with reads aligned to a reference genome, which are generated during a whole-genome bisulphite sequencing (WGBS) experiment, and produces genome-wide information about the statistical properties of methylation in a single-sample and in a differential analysis framework. The resulting information is stored in multiple MATLAB MAT files for subsequent processing and is summarized by bedGraph genomic tracks that can be visualized using a genome browser (such as the UCSC genome browser, see https://genome.ucsc.edu).

The current implementation of informME has been tested within the following environments:

  • Red Hat Enterprise Linux Server Release 6.5 (Santiago), CentOS release 6.7 (Final) and Ubuntu 16.04 (xenial)

  • Sun Grid Engine OGS/GE 2011.11 and Slurm 17.02.3

  • MATLAB R2013b 64-bit, MATLAB R2016b 64-bit, MATLAB R2017a 64-bit, and MATLAB R2018a 64-bit with Bioinformatics and Symbolic Math Toolboxes

and by using the following tools:

  • Bismark Bisulphite Mapper - v0.13.1 (to make BAM files)

  • SAMtools - v0.1.19 (to parse BAM files)

  • gcc/g++ compiler - v4.4.7-4, v4.9.2, v6.3.0 (to compile MATLAB MEX code, version must be compatible with corresponding MATLAB release, see Dependencies Section)