ComBat-met: adjusting batch effects in DNA methylation data
Junmin Wang
Abstract
Integration of genomics data is routinely hindered by unwanted technical variations known as batch effects. Despite wide availability, existing batch correction methods often fall short in capturing the unique characteristics of DNA methylation data. We present ComBat-met, a beta regression framework to adjust batch effects in DNA methylation studies. Our method fits beta regression models to the data, calculates batch-free distributions, and maps the quantiles of the estimated distributions to their batch-free counterparts. Compared to traditional methods, ComBat-met followed by differential methylation analysis shows improved statistical power without compromising false positive rates based on simulated data. Additionally, we demonstrate the ability of ComBat-met to remove cross-batch variations and recover biological signals using data from The Cancer Genome Atlas.