Analyzing whole genome bisulfite sequencing data from highly divergent genotypes

Abstract

In the study of DNA methylation, genetic variation between species, strains or individuals can result in CpG sites that are exclusive to a subset of samples, and insertions and deletions can rearrange the spatial distribution of CpGs. How to account for this variation in an analysis of the interplay between sequence variation and DNA methylation is not well understood, especially when the number of CpG differences between samples is large. Here, we use whole-genome bisulfite sequencing data on two highly divergent mouse strains to study this problem. We show that alignment to personal genomes is necessary for valid methylation quantification. We introduce a method for including strain-specific CpGs in differential analysis, and show that this increases power. We apply our method to a human normal-cancer dataset, and show this improves accuracy and power, illustrating the broad applicability of our approach. Our method uses smoothing to impute methylation levels at strain-specific sites, thereby allowing strain-specific CpGs to contribute to the analysis, while accounting for differences in the spatial occurrences of CpGs. Our results have implications for joint analysis of genetic variation and DNA methylation using bisulfite-converted DNA, and unlocks the use of personal genomes for addressing this question.

Divergent DNA methylation contributes to duplicated gene evolution and chilling response in tea plants

“…We here generated the single‐base‐resolution DNA methylation map of tea plants under chilling stress through whole‐genome bisulfite sequencing. The average bisulfite non‐conversion rate is approximately 0.3% in our samples, which is comparable with previous studies (Li et al, 2020; Wulfridge et al, 2019). Approximately 21, 23 and 56% of the cytosines in the tea plant genome are methylated at CG, CHG and CHH sites, respectively.…”

Genome-wide detection of cytosine methylation by single molecule real-time sequencing

“…It was previously reported that the genomic variations, including single nucleotide variants, insertions, and deletions, would introduce “quantification bias” of methylation levels in the step of alignment for BS-seq reads ( 31 ). For example, if a sample genome has a “CG-to-TG” variant relative to the reference genome sequence, standard alignment approaches would consider a “TG” dinucleotide in a read to be derived from an unmethylated CpG dinucleotide, resulting in an underestimation of methylation level ( 31 ). Such a quantification bias would lead to inaccurate data interpretation when comparing methylation patterns between species and human normal-cancer datasets with divergent genotypes ( 31 ).…”

“…For example, if a sample genome has a “CG-to-TG” variant relative to the reference genome sequence, standard alignment approaches would consider a “TG” dinucleotide in a read to be derived from an unmethylated CpG dinucleotide, resulting in an underestimation of methylation level ( 31 ). Such a quantification bias would lead to inaccurate data interpretation when comparing methylation patterns between species and human normal-cancer datasets with divergent genotypes ( 31 ). In addition to a dedicated bioinformatics approach for mitigating the quantification bias ( 31 ), the HK model presented in this study would provide an opportunity to address the quantification bias due to the issue of mappability in the standard alignment approaches for BS-seq.…”

“…Such a quantification bias would lead to inaccurate data interpretation when comparing methylation patterns between species and human normal-cancer datasets with divergent genotypes ( 31 ). In addition to a dedicated bioinformatics approach for mitigating the quantification bias ( 31 ), the HK model presented in this study would provide an opportunity to address the quantification bias due to the issue of mappability in the standard alignment approaches for BS-seq. In this regard, we wish to highlight two attributes of the approach described in our study.…”

Introduction to Single-Cell DNA Methylation Profiling Methods

“…This acts as a confounding factor in downstream analysis, including finding differentially methylated regions. The conversion efficiency can be directly derived from unmethylated lambda DNA spiked in by calculating the fraction of converted bases at cytosine sites or using available software programs [37,77,78].…”