Based on an extensive simulation study, McGregor and colleagues recently recommended the use of surrogate variable analysis (SVA) to control for the confounding effects of cell-type heterogeneity in DNA methylation association studies in scenarios where no cell-type proportions are available. |