Correcting for cell-type composition bias in epigenome-wide association studies

Recent epigenome-wide association studies have indicated a potential role for epigenetic variation in the etiology of complex human diseases. However, one major challenge is to distinguish true epigenetic variation from changes caused by differences in cellular composition between the disease and non-disease state, a problem that is particularly relevant when analyzing whole blood. For studies with large numbers of samples, it can be expensive and very time consuming to perform cell sorting, and it is often not clear which is the correct cell type to profile. Two recently published papers have attempted to address this confounding issue using bioinformatics.