Abstract
Background: A statistical methodology is available to estimate the proportion of cell types (cellular heterogeneity) in adult whole blood specimens used in epigenome-wide association studies (EWAS). However, there is no methodology to estimate the proportion of cell types in umbilical cord blood (also a heterogeneous tissue) used in EWAS.
Objectives: The objectives of this study were to determine whether differences in DNA methylation (DNAm) patterns in umbilical cord blood are the result of blood cell type proportion changes that typically occur across gestational age and to demonstrate the effect of cell type proportion confounding by comparing preterm infants exposed and not exposed to antenatal steroids.
Methods: We obtained DNAm profiles of cord blood using the Illumina HumanMethylation27k BeadChip array for 385 neonates from the Boston Birth Cohort. We estimated cell type proportions for six cell types using the deconvolution method developed by Houseman et al. (2012).
Results: The cell type proportion estimates segregated into two groups that were significantly different by gestational age, indicating that gestational age was associated with cell type proportion. Among infants exposed to antenatal steroids, the number of differentially methylated CpGs dropped from 127 to 1 after controlling for cell type proportion.
Discussion: EWAS utilizing cord blood are confounded by cell type proportion. Careful study design including correction for cell type proportion and interpretation of results of EWAS using cord blood are critical.