BOREALIS: an R/Bioconductor package to detect outlier methylation from bisulfite sequencing data

Background: Rare genetic disease studies have benefited from the era of high throughput sequencing. DNA sequencing results in diagnosis 18-40% previously unsolved cases, while incorporation RNA-Seq analysis has more recently been shown to generate significant numbers unattainable diagnoses. While methylation remains less explored, multiple inborn diseases resulting disorders genomic imprinting are well characterized and a growing body literature suggests causative or correlative role aberrant diverse rare inherited conditions. Complex pictures patterning also emerging, including association regional, specific-site even single-site methylation, with disease. The systematic application genomic-wide methylation-based for undiagnosed cases is logical progression current testing paradigms. Similar rationale exploited RNA-based studies, we can assume that disease-associated aberrations an individual will demonstrate differences other individuals unrelated phenotypes. Thus, aberrantly methylated sites be outliers heterogeneous cohort individuals. Methods: Based on this rationale, present BOREALIS: Bisulfite-seq OutlieR MEthylation At SingLeSIte ReSolution. BOREALIS uses beta binomial model identify outlier at single CpG site resolution bisulfite data. Results: Utilizing power analyses, within samples. Furthermore, show tolerant inclusion identical sufficient size depth. Conclusions: The method demonstrates improved performance versus standard statistical suited multi-site downstream analysis.