cvlr: Finding heterogeneously methylated genomicregions using ONT reads

SummaryNanopore reads encode information on the methylation status of cytosines in CpG dinucleotides. The length of the reads makes it comparatively easy to look at patterns consisting of multiple loci; here we exploit this property to look for regions where one can define subpopulations of cells based on methylation patterns. As a benchmark we run our clustering algorithm on known imprinted genes and show that the clustering based on methylation is consistent with the phasing of the genome; we then scan chromosome 15 looking for windows corresponding to heterogeneous methylation. We can also compute the covariance of methylation across these regions while keeping into account the mixture of different types of reads. Availabilityhttps://github.com/EmanueleRaineri/releases Contactemanuele.raineri@cnag.crg.eu, simon.heath@cnag.crg.eu Supplementary informationTables, figures, and some further explanations of the algorithms are available as online supplementary information.