Sample-specific CpG loci are important for accurate long-read methylation analysis

Methylated CpG loci mutate at high rates, but it remains challenging to identify sample-specific non-reference CpG loci for methylation analyses. Here, we present a likelihood-based method that identifies sample-specific CpG loci at high precision in long-read sequencing data from Nanopore and PacBio platforms. Inclusion of non-reference CpG loci increased the number of loci and reduced reference bias, and is important for accurate estimation of global methylation levels across samples.