This page hosts open-source analysis tools for DNA methylation data generated on the CHARM microarray platform. The main functions are provided by the R packages oligo and charmR.

Software features:

  • Data quality assessment
  • Identification of differentially methylated regions (DMRs)
  • Probe-level percentage methylation estimates

The software on this page is under active development and will be updated frequently. Check back often to learn about important updates.

Requirements

  • R
  • Bioconductor
  • The Bioconductor packages oligo and charmR and their dependencies.
  • Raw Nimblegen data (.xys) files. These can be requested from Nimblegen.
  • 3Gb RAM

Installation

  • Install a recent version of R (>=2.10) if not already installed.
  • Open R and install the packages
> source("http://www.bioconductor.org/biocLite.R")
> repos <- c("http://r.martinaryee.com", biocinstallRepos())
> pkgs <- c("charmR", "pd.feinberg.hg18.me.hx1")
> install.packages(pkgs=pkgs, repos=repos, dependencies=c("Depends", "Imports"))

This will also download and install the additional necessary dependencies.

Analysis quick start example:

Download the example data zip file that contains xys files for 4 samples and a corresponding sample description text file.

Start R, load the charmR package and change to the directory containing the xys files. 

 > library(charmR)
 > setwd("path/to/the/xysfiles")

Load the tab-delimited sample description file. Only two columns are necessary: a filename column and a group label column (assuming that your file names have the suffixes _532.xys and _635.xys corresponding to the untreated and methyl-depleted channels respectively). In this example the two columns are called 'Filename' and 'Tissue_Type'. 

 > pd <- read.delim("sample_description_file.txt")

 > pd
         Filename            DNA Individual Tissue_Type
 1 136413_532.xys      untreated        441       brain
 2 136421_532.xys      untreated        441       liver 
 3 136593_532.xys      untreated        449       brain
 4 186974_532.xys      untreated        432       liver
 5 136413_635.xys methyldepleted        441       brain
 6 136421_635.xys methyldepleted        441       liver
 7 136593_635.xys methyldepleted        449       brain
 8 186974_635.xys methyldepleted        432       liver

Read in the data 

 > rawData <- readCharm(files=pd$Filename, type=pd$Tissue_Type)

Run a quality report. The qcReport function returns a quality score between 0 and 100 for each array. If passed a filename it will also produce a pdf report with the given name. 

 > qual <- qcReport(rawData, file="qcReport.pdf")

 > qual
   136413   136421   136593   186974 
 77.62826 78.59368 77.64333 83.97405

Find DMRs 

 > grp <- pData(rawData)$type

 > grp
 [1] "brain" "liver" "brain" "liver"

 > dmr <- dmrFinder(rawData, groups=grp)

The output of dmrFinder is a list with components that include:

  • tabs: a list of DMR tables, one for each pair-wise comparison
  • p: the matrix of percentage methylation estimates
  • chr, pos: chromosomal coordinates for each probe (row) in p 
 > names(dmr)
 [1] "tabs"         "p"            "chr"          "pos"          "pns"         
 [6] "controlIndex" "gm"           "gs" 

 > names(dmr$tabs)
 [1] "brain-liver"

 > head(dmr$tabs'brain-liver'?)
         chr     start       end        p1        p2               regionName
 6594  chr15  91163203  91164042 0.1289797 0.7342032  chr15:91150286-91166158
 11216 chr20  55267777  55268617 0.2036257 0.7734801  chr20:55266497-55276390
 13436  chr2  54538781  54539636 0.1489372 0.6450484   chr2:54535410-54540793
 14102  chr3 149898845 149899442 0.6974871 0.1410209 chr3:149897739-149899949
 17864  chr6  52637957  52638681 0.7222176 0.1365534   chr6:52635302-52638967
 18262  chr7 130439092 130439758 0.1643708 0.6602480 chr7:130437932-130444273
       indexStart indexEnd      area
 6594      530285   530309 15.130587
 11216    1053004  1053028 14.246361
 13436     947841   947864 11.906668
 14102    1220947  1220964 10.016393
 17864    1422899  1422915  9.956292
 18262    1538393  1538412  9.917545

Update the installed software

To update your installation to the latest versions of the CHARM analysis tools: 

 > update.packages(repos="http://r.martinaryee.com")

References

RA Irizarry et al., Comprehensive high-throughput arrays for relative methylation (CHARM), Genome Res. 2008 May;18(5):780-90. PMID 18316654.