Comprehensive Evaluation of Differential Methylation Analysis Methods for Bisulfite Sequencing Data

Abstract

Background: With advances in next-generation sequencing technologies, the bisulfite conversion of genomic DNA followed by sequencing has become the predominant technique for quantifying genome-wide DNA methylation at single-base resolution. A large number of computational approaches are available in literature for identifying differentially methylated regions in bisulfite sequencing data, and more are being developed continuously. Results: Here, we focused on a comprehensive evaluation of commonly used differential methylation analysis methods and describe the potential strengths and limitations of each method. We found that there are large differences among methods, and no single method consistently ranked first in all benchmarking. Moreover, smoothing seemed not to improve the performance greatly, and a small number of replicates created more difficulties in the computational analysis of BS-seq data than low sequencing depth. Conclusions: Data analysis and interpretation should be performed with great care, especially when the number of replicates or sequencing depth is limited.

Cold exposure impacts DNA methylation patterns in cattle sperm

“…Using Methylkit, with a threshold of methylation difference ≥10% and p < 0.05, used by another study for evaluating cattle sperm methylation, we could not detect DMCs, but with DSS, we got a striking difference (Stiavnicka et al, 2022). Consistently, DSS identified more DMCs (Piao et al, 2021). DSS quantifies methylation by the Bayesian hierarchical model based on direct methylation count, in contrast Methylkit uses logistic regression by converting the methylation count to percentage (Akalin et al, 2012;Feng et al, 2014).…”

DNA methylation in T cell leukaemia

“…The downstream analysis of whole genome methylation sequencing is most commonly the identification of differential DNA methylation between samples. Various tools have been developed for the analysis of differentially methylated cytosines utilizing several statistical tests, such as logistic regression, logistic ratio test, modified t-test, and fisher's exact test [367]. Due to high sequencing costs or a lack of input material whole genome methylation sequencing is often limited by the omission of biological replicates.…”

Assessing the Differential Methylation Analysis Quality for Microarray and NGS Platforms

“…Benchmarks based on experimental data are limited to imprinted DMRs [15], goldstandard DMRs inferred from the results of RNA-seq and DNase-seq experiments [16] and methylation titration data [17]. While using real data, the assumptions and biases that are inevitable in the simulation approach can be avoided.…”

“…(1) Signature quality Recall [12][13][14]16,18] Simulated [17] Real True negative rate [13,18] Simulated False discovery rate [12] Simulated Precision [14] Simulated [17] Real ROC AUC [13,14,16,18,19] Simulated False positive rate [13,16,18] Simulated [14] Real Empirical distribution of p-values under the null [13,16] Simulated DMR overlapping fraction [16] Real…”