Analyzing epigenetic modifications at sample-specific CpG loci in long-read data using a likelihood-based framework in bsseq

SUMMARY: DNA in CpG context is often epigenetically modified in vertebrate genomes which increases their mutation rate. However, it remains challenging to identify sample-specific CpG loci for analysis of epigenetic modifications. Here, we present a read-based likelihood framework for identifying reference and non-reference CpG loci with high precision and sensitivity in single-molecule long-read sequencing data from Nanopore and PacBio platforms. Including all sample-specific CpG loci increases the number of analyzable loci and reduces bias in modification level estimates, both critical factors for accurate analyses of epigenetic modifications. AVAILABILITY AND IMPLEMENTATION: The likelihood framework is implemented in the R package bsseq with the source code available on github https://github.com/hansenlab/bsseq. Installation instructions and a user guide for long-read analysis and likelihood-based filtering of CpG loci are available on Bioconductor https://www.bioconductor.org/packages/release/bioc/html/bsseq.html. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.