Background

DNA methylation plays an important role in many cellular processes and is currently being widely studied to gain a better understanding of human development and disease (Robertson, 2005). Most epigenome-wide association studies (EWAS) search for associations between DNA methylation and disease (Flanagan, 2015). For this aim, Illumina’s BeadChip arrays are widely used to measure DNA methylation in humans. Methylation in promoters is associated with gene expression repression (Boyes and Bird, 1992). The mechanisms of expression repression include impeding the binding of transcription factors and recruiting transcription repressors (Cedar and Bergman, 2012). Aberrant DNA methylation in these regions has been linked to several diseases, including cancer (Ehrlich and Lacey, 2013) and autoimmune disorders (Dozmorov et al., 2014). There are several R packages designed to detect differentially methylated regions that apply de novo and predefined strategies, as previously reviewed (Martorell-Marugán et al., 2019). Most of the predefined methods can be applied directly to detect differentially methylated promoters. On the contrary, de novo methods search for differentially methylated regions along the entire genome that should be annotated in order to detect which regions are located at promoters.

In this protocol, we present the complete data and statistical analysis pipeline to detect differentially methylated promoters in disease phenotypes from Illumina BeadChip data based on the mCSEA R package (Martorell-Marugán et al., 2019), which applies a predefined regions strategy. We used previously published data from patients with two rare neurodevelopmental diseases: Williams syndrome (WS) and 7q11.23 duplication syndrome (Dup7), as well as typically developing (TD) patients. Methylation in blood cells was measured in all the samples by the authors of the original study (Strong et al., 2015). This pipeline can be easily adapted to study other genomic regions such as gene body specific methylation patterns or CpG islands (CGIs). The complete code for this protocol is available as Supplemental_script file.