MethylKit

Locating differentially methylated regions with methylKit

methylKit is an R package for analysis and annotation of DNA methylation information obtained by high-throughput bisulfite sequencing. The package is designed to deal with sequencing data from RRBS and its variants. But, it can potentially handle whole-genome bisulfite sequencing data if proper input format is provided.

Input Requirements

BED File

This workflow currently accepts 3 kinds of input files:

Bismark Coverage Files
Bismark Cytosine Reports
bedMethyl Files

Bismark Coverage Files

This file type is an output of the bismark2bedGraph function of Bismark. It contains coverage in the file format below (using 1-based genomic genomic coordinates):

<chromosome> <start position> <end position> <methylation percentage> <count methylated> <count unmethylated>

Typically, these files end in the suffix .bismark.cov. They are not stranded and more information on how they are used by Bismark can be found on page 41 of this guide.

Bismark Cytosine Reports

This file is a genome-wide cytosine report output from Bismark. Starting from the coverage output, the Bismark methylation extractor can optionally also output a genome-wide cytosine methylation report. The module coverage2cytosine (part of the Bismark package) may also be run individually. It is also sorted by chromosomal coordinates but also contains the sequence context and is in the following format:

<chromosome> <position> <strand> <count methylated> <count unmethylated> <C-context> <trinucleotide context>

bedMethyl Format

These are BED files in the bedMethyl format from ENCODE. The bedMethyl file is a bed9+2 file containing the number of reads and the percent methylation. Each column represents the following:

Reference chromosome or scaffold
Start position in chromosome
End position in chromosome
Name of item
Score from 0-1000. Capped number of reads
Strandedness, plus (+), minus (-), or unknown (.)
Start of where display should be thick (start codon)
End of where display should be thick (stop codon)
Color value (RGB)
Coverage, or number of reads
Percentage of reads that show methylation at this position in the genome

Treatment Information

For each BED file, there is a boolean flag used to indicate which samples should be compared against other samples. Group samples by this toggle.

Sample Name

For each sample, give it a name (this will be used for graphing purposes).

Track Name

The Track Name will be used in IGV to label the track of DMR Regions.

Parameters

The following parameters are used in the workflow to process the BED files and assign DMRs:

Minimum Base Coverage (default = 3): Filters a methylRawList and discards bases that have coverage below this threshold (e.g. if set to 10, CpGs with coverage less than 10X will be discarded).
Tiling Parameter (default = 200): Used in TileMethylCounts(). This value tiles the genome with windows of this specified length and step-size and summarizes the methylation information on those tiles.
Tiling Coverage Cutoff (default = 10): Filter based on the number of bases (cytosines) per tiled region.
Percent Methylation Difference Cutoff (default = 25): Used in getMethylDiff(). This value defines what the percent methylation difference must be greater than between samples for a region in order to be counted as a DMR.
q-Value Cutoff (default = 0.05): Used in getMethylDiff(). This value defines what the q-value cutoff is in order for a region to be counted as a DMR.

Output Files

This workflow outputs three folders:

processed_beds: Contains BED files that are processed from methylBED format with a new column representing frequency of methylation as a float percent. These are used as input to MethylKit.
methylkit_results: Contains a list of DMRs and multiple plots on coverage.
IGV_tracks: Contains the IGV track of DMRs between samples.

Test Data

The test data includes Chromosome 1 from the following WGBS samples from ENCODE:

WGBS from skeletal muscle myoblast (ENCLB587BLQ) - Hg38

GEO Link, ENCODE Link

WGBS from body of pancreas (ENCLB777VZI) - Hg38