2013/12/11 by Andrea Riebler, Riebler, Andrea, Mirco Menigatti +15
Biochemistry, Genetics and Molecular Biology · #Epigenetics and DNA Methylation #Genomic variations and chromosomal abnormalities #Genomics and Phylogenetic Studies
paper · pdf · doi:10.48550/arxiv.1312.3115
DNA methylation (DNAme) is a critical component of the epigenetic regulatory\nmachinery and aberrations in DNAme patterns occur in many diseases, such as\ncancer. Mapping and understanding DNAme profiles offers considerable promise\nfor reversing the aberrant states. There are several approaches to analyze\nDNAme, which vary widely in cost, resolution and coverage. Affinity capture and\nhigh-throughput sequencing of methylated DNA strike a good balance between the\nhigh cost of whole genome bisulphite sequencing (WGBS) and the low coverage of\nmethylation arrays. However, existing methods cannot adequately differentiate\nbetween hypomethylation patterns and low capture efficiency, and do not offer\nflexibility to integrate copy number variation (CNV). Furthermore, no\nuncertainty estimates are provided, which may prove useful for combining data\nfrom multiple protocols or propagating into downstream analysis. We propose an\nempirical Bayes framework that uses a fully methylated (i.e. SssI treated)\ncontrol sample to transform observed read densities into regional methylation\nestimates. In our model, inefficient capture can be distinguished from low\nmethylation levels by means of larger posterior variances. Furthermore, we can\nintegrate CNV by introducing a multiplicative offset into our Poisson model\nframework. Notably, our model offers analytic expressions for the mean and\nvariance of the methylation level and thus is fast to compute. Our algorithm\noutperforms existing approaches in terms of bias, mean-squared error and\ncoverage probabilities as illustrated on multiple reference datasets. Although\nour method provides advantages even without the SssI-control, considerable\nimprovement is achieved by its incorporation. Our method can be applied to\nmethylated DNA affinity enrichment assays (e.g MBD-seq, MeDIP-seq) and a\nsoftware implementation is available in the Bioconductor Repitools package.\n