epialleleR is an R package for calling and reporting cytosine DNA
methylation. Developed to help identify and quantify epimutations (aberrant DNA
methylation events), it has now acquired multiple additional functions to
dissect DNA methylation in many useful ways. But the main feature of the
package is to report frequencies of epimutations (variant epiallele
frequencies, VEF) at the level of genomic regions or individual cytosines. All
you need in order to use it, is a binary alignment map (BAM) file from
basically any next-generation (methylation or native) sequencing experiment.
- very fast!
- reference-free
- designed with epimutation studies in mind
- probably, the most accurate tool for conventional cytosine methylation reporting
For details, see the related publication and vignette.
- short-read and long-read (native)
- paired-end and single-end
- whole-genome, genome-wide (e.g., hybridization capture or adaptive sampling), and narrowly targeted (e.g., amplicon panels)
- call cytosine methylation and save calls in a new BAM file
(
callMethylation) - create sample BAM files from scratch given mandatory and optional
BAM fields (
simulateBam) - create conventional reports of cytosine methylation
(
generateCytosineReport) - evaluate epimutation frequencies both at the
level of genomic regions (
generate[Bed|Amplicon|Capture]Report) and individual cytosines (generateCytosineReport) - evaluate linearised Methylated Haplotype Load (lMHL,
generateMhlReport) - extract methylation patterns for genomic region of interest
(
extractPatterns) - visualise methylation patterns (
plotPatterns) - test for the association between epiallele methylation
status and sequence variations (
generateVcfReport) - assess the distribution of per-read beta values for genomic regions of
interest (
generateBedEcdf)
- using genomic coordinates of targets, only a subset of BAM reads or only fragments of BAM reads that are overlapping the targets can now be loaded
- disrupting changes in all
generate*Reportfunctions (from version 1.19.1 onwards):cytosine.contextparameter instead ofthreshold.context/haplotype.context- new parameter
filter.readsregulates filtering of reads with too few cytosines or presumable incomplete conversion of cytosines - reads with too few within-the-context cytosines (less than
min.context.sites) or out-of-context cytosine methylation higher thanmax.outofcontext.betaare filtered out (discarded) instead of being counted as hypomethylated reads (as was in v1.19.0 and earlier) - new default value of 0 for
min.context.sites
- creates pretty plots of methylation patterns
- inputs long-read sequencing alignments
- full support for short-read sequencing alignments by Illumina DRAGEN, Bismark, bwa-meth, BSMAP
- RRBS-specific options
- lower memory usage
- inputs both single-end and paired-end sequencing alignments
- makes and stores methylation calls
- creates sample BAM files
- reports linearised MHL
- significant speed-up
- method to extract and visualize methylation patterns
- even faster and more memory-efficient BAM loading (by means of HTSlib)
- min.baseq parameter to reduce the effect of low quality bases on
methylation or SNV calling (in v1.0 the output of
generateVcfReportwas equivalent to the one ofsamtools mpileup -Q 0 ...)
check out NEWS for more!
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("epialleleR")library(devtools)
install_github("BBCG/epialleleR", build_vignettes=FALSE,
repos=BiocManager::repositories(),
dependencies=TRUE, type="source")Please read epialleleR vignette
at GitHub pages
or within the R environment: vignette("epialleleR", package="epialleleR"), or
consult the function's help pages for the extensive information on usage,
parameters and output values.
Comparison of beta, VEF and lMHL values for various use cases is given by the
values
vignette (vignette("values", package="epialleleR"))
Very brief synopsis:
library(epialleleR)
# make methylation calls if necessary
callMethylation(
input.bam.file=system.file("extdata", "test", "dragen-se-unsort-xg.bam", package="epialleleR"),
output.bam.file=tempfile(pattern="output-", fileext=".bam"),
genome=system.file("extdata", "test", "reference.fasta.gz", package="epialleleR")
)
# make a sample BAM file from scratch
simulateBam(output.bam.file=tempfile(pattern="simulated-", fileext=".bam"),
pos=c(1, 2), XM=c("ZZZzzZZZ", "ZZzzzzZZ"), XG=c("CT", "AG"))
# or use external files
amplicon.bam <- system.file("extdata", "amplicon010meth.bam",
package="epialleleR")
amplicon.bed <- system.file("extdata", "amplicon.bed", package="epialleleR")
amplicon.vcf <- system.file("extdata", "amplicon.vcf.gz", package="epialleleR")
# preload the data
bam.data <- preprocessBam(amplicon.bam)
# methylation patterns and their plot
patterns <- extractPatterns(bam=amplicon.bam, bed=amplicon.bed, bed.row=3)
plotPatterns(patterns)
# conventional cytosine report
cx.report <- generateCytosineReport(bam.data, filter.reads=FALSE,
threshold.reads=FALSE, report.context="CX")
# CpG VEF report for individual bases
cg.vef.report <- generateCytosineReport(bam.data)
# BED-guided VEF report for genomic ranges
bed.report <- generateBedReport(bam=amplicon.bam, bed=amplicon.bed,
bed.type="capture")
# VCF report
vcf.report <- generateVcfReport(bam=amplicon.bam, bed=amplicon.bed,
vcf=amplicon.vcf, vcf.style="NCBI")
# lMHL report
mhl.report <- generateMhlReport(bam=amplicon.bam)Oleksii Nikolaienko, Per Eystein Lønning, Stian Knappskog, epialleleR: an R/Bioconductor package for sensitive allele-specific methylation analysis in NGS data. GigaScience, Volume 12, 2023, giad087, https://doi.org/10.1093/gigascience/giad087. Data: GSE201690
Per Eystein Lonning, Oleksii Nikolaienko, Kathy Pan, Allison W. Kurian, Hans Petter Petter Eikesdal, Mary Pettinger, Garnet L Anderson, Ross L Prentice, Rowan T. Chlebowski, and Stian Knappskog. Constitutional BRCA1 methylation and risk of incident triple-negative breast cancer and high-grade serous ovarian cancer. JAMA Oncology 2022. https://doi.org/10.1001/jamaoncol.2022.3846
Oleksii Nikolaienko, Hans P. Eikesdal, Elisabet Ognedal, Bjørnar Gilje, Steinar Lundgren, Egil S. Blix, Helge Espelid, Jürgen Geisler, Stephanie Geisler, Emiel A.M. Janssen, Synnøve Yndestad, Laura Minsaas, Beryl Leirvaag, Reidun Lillestøl, Stian Knappskog, Per E. Lønning. Prenatal BRCA1 epimutations contribute significantly to triple-negative breast cancer development. Genome Medicine 2023. https://doi.org/10.1186/s13073-023-01262-8. Data: GSE243966
Oleksii Nikolaienko, Garnet L Anderson, Rowan T Chlebowski, Su Yon Jung, Holly R Harris, Stian Knappskog, and Per E Lønning. MGMT epimutations and risk of incident cancer of the colon, glioblastoma multiforme, and diffuse large B-cell lymphomas. Clinical Epigenetics 2025. https://doi.org/10.1186/s13148-025-01835-x
Artistic License 2.0