Covariate adjusted differential variability analysis of DNA methylation with propensity score method

Abstract

It has been proposed recently that differentially variable CpG methylation (DVC) may contribute to transcriptional aberrations in human diseases. In large scale epigenetic studies, potential confounders could affect the observed methylation variabilities and need to be accounted for. In this paper, we develop a robust statistical model for differential variability DVC analysis that accounts for potential confounding covariates by utilizing the propensity score method. Our method is based on a weighted score test on strata generated propensity score stratification. To the best of our knowledge, this is the first proposed statistical method for detecting DVCs that adjusts for confounding covariates. We show that this method is robust against model misspecification and achieves good operating characteristics based on extensive simulations and a case study.

methylDMV: SIMULTANEOUS DETECTION OF DIFFERENTIAL DNA METHYLATION AND VARIABILITY WITH CONFOUNDER ADJUSTMENT

“…32 In such cases, confounding factors need to be properly accounted for to avoid biases in DNA methylation biomarker detection. There are several approaches for DM analysis which allow for confounders adjustment, 33 however to the best of our knowledge existing DV analysis approaches are not tailored for confounders adjustments, except for our earlier work 16 which proposed a DV only analysis in the presence of confounders within large scale hypothesis testings framework. This paper extends our earlier work which allows for simultaneous detection of DM and DV in large scale hypothesis testings framework, and at the same time provides a candidate feature selection mechanism for the prediction algorithm.…”

“…On the other hand, several algorithms have been proposed in recent years to identify CpGs which exhibit differential variability in large scale hypothesis testing. 15 introduced a generalized additive models for location, scale and shape (GAMLSS) framework and Kuan (2014) 16 proposed a general linear model with propensity score method for detecting CpGs with differential variability.…”