Detection of outlier methylation from bisulfite sequencing data with novel Bioconductor package BOREALIS

Abstract

DNA sequencing results in genetic diagnosis of 18-40% of previously unsolved cases, while the incorporation of RNA-Seq analysis has more recently been shown to generate significant numbers of previously unattainable diagnoses. Multiple inborn diseases resulting from disorders of genomic imprinting are well characterized and a growing body of literature suggest the causative or correlative role of aberrant DNA methylation in diverse rare inherited conditions. Therefore, the systematic application of genomic-wide methylation-based sequencing for undiagnosed cases of rare disease is a logical progression from current testing paradigms. Following the rationale previously exploited in RNA-based studies of rare disease, we can assume that disease-associated methylation aberrations in an individual will demonstrate significant differences from individuals with unrelated phenotypes. Thus, aberrantly methylated sites will be outliers from a heterogeneous cohort of individuals.Based on this rationale, we present BOREALIS: Bisulfite-seq OutlieR MEthylation At SingLe-SIte ReSolution. BOREALIS uses a beta binomial model to identify outlier methylation at single CpG site resolution from bisulfite sequencing data. This method addresses a need unmet by standard differential methylation analyses based on case-control groups. Utilizing a heterogeneous cohort of 94 rare disease patients undiagnosed following DNA-based testing we show that BOREALIS can successfully identify outlier methylation linked to phenotypically relevant genes, providing a new avenue of exploration in the quest for increased diagnostic rates in rare disease patients. We highlight the case of a patient with previously undetected hypermethylation patterns that are informing clinical decision-making. BOREALIS is implemented in R and is freely available as a Bioconductor package.

BiocMAP: a Bioconductor-friendly, GPU-accelerated pipeline for bisulfite-sequencing data

“…In the second extraction module, DNA methylation proportion extraction is performed within each sample using MethylDackel [ 17 ] or optionally Bismark [ 7 ], and the results are aggregated across samples into a pair of bsseq [ 11 ] R [ 10 ] objects for easy integration with a number of Bioconductor [ 13 ] packages to facilitate downstream statistical analyses. Some examples include limma [ 18 ] for linear modeling, Borealis [ 19 ] for outlier detection, and MethCP [ 20 ] and DMRcate [ 21 ] for finding differentially methylated regions. The first bsseq object contains counts of methylated and unmethylated cytosines in CpG context across the entire reference genome, while the second object contains any additional cytosines in CpH context, when relevant (Fig.…”