Fig. 5

Distinguishing between genetic artifacts and genuine genetic influence on DNA methylation by co-methylation: an extensive follow-up on the meQTL of the first intron of the NINJ2 gene. A Squared correlation between the DNA methylation ratio of monozygotic co-twins and squared co-methylation across samples as a function of the genomic coordinate of a representative genetic artifact, and B the NINJ2 meQTL. M-values were used instead of methylation ratios in the computation of correlations. Plots are centered around the CpG highlighted in red. CpGs highlighted in purple correspond to those assessed to be potentially artifactual probes. Distances between CpGs targeted by contiguous probes (Δδ) are also highlighted in the co-methylation plot. C Methylation ratio heatmap of the meQTL at the first intron of NINJ2 in datasets of EUR and AFR ancestry. Counts for each of the three clusters are included, obtained from the shown dendrogram. For EUR ancestry, twin-to-twin agreement is represented as a matrix. For AFR, this is not shown as the data consists of unrelated individuals. D Employing frequencies in EUR and AFR estimated from U/M plot to map NINJ2 meQTL’s most likely cis-causal SNP (highlighted with an arrow). Only SNPs in the region with MAF > 0.2 are shown. CpGs and SNPs are represented as circles and rhombi, respectively. Red and blue CpGs correspond to those within or outside the expected co-methylation window (average genomic position across co-methylation CpGs ± 1 kb, colored as a brown rectangle). E Confirmation of meQTL co-methylation across the NINJ2 first intron region, inter-individual variation, and co-twin similarity in DNA methylation using whole-genome bisulfite sequencing data of MZ twins in whole blood. Twin pair numbering was kept as in the original dataset (E-MTAB-3549)