Genetic effects on the skin methylome in healthy older twins

Summary Whole-skin DNA methylation variation has been implicated in several diseases, including melanoma, but its genetic basis has not yet been fully characterized. Using bulk skin tissue samples from 414 healthy female UK twins, we performed twin-based heritability and methylation quantitative trait loci (meQTL) analyses for >400,000 DNA methylation sites. We find that the human skin DNA methylome is on average less heritable than previously estimated in blood and other tissues (mean heritability: 10.02%). meQTL analysis identified local genetic effects influencing DNA methylation at 18.8% (76,442) of tested CpG sites, as well as 1,775 CpG sites associated with at least one distal genetic variant. As a functional follow-up, we performed skin expression QTL (eQTL) analyses in a partially overlapping sample of 604 female twins. Colocalization analysis identified over 3,500 shared genetic effects affecting thousands of CpG sites (10,067) and genes (4,475). Mediation analysis of putative colocalized gene-CpG pairs identified 114 genes with evidence for eQTL effects being mediated by DNA methylation in skin, including in genes implicating skin disease such as ALOX12 and CSPG4. We further explored the relevance of skin meQTLs to skin disease and found that skin meQTLs and CpGs under genetic influence were enriched for multiple skin-related genome-wide and epigenome-wide association signals, including for melanoma and psoriasis. Our findings give insights into the regulatory landscape of epigenomic variation in skin.

Figure S16.Proportion of eGenes colocalised in both the original meQTL-eQTL colocalisation analysis using TwinsUK data, and the validation analysis using the suprapubic skin GTEx eQTL data with TwinsUK meQTL data.Figure S17.Proportion of eGenes tested in both the original meQTL-eQTL colocalisation analysis using TwinsUK data, and the validation analysis using the suprapubic skin GTEx eQTL data with TwinsUK meQTL data, which colocalise in each analysis.

Figure
Figure S1.(a) Biplot of PC1 and PC2 from PCA of whole skin DNA methylation data from 414 samples, with dermis and epidermis DNA methylation data from 92 samples from Vandiver et al. 1 (b) Scatter plot of PC1 from S1a ("Tissue Type PC") with PC1 from PCA on just the 414 whole skin DNA methylation data.

Figure S2 .
Figure S2.Scatter plots of PC1 from PCA of whole skin DNA methylation data from 414 samples vs EpiSCORE cell-type estimates for each sample.

Figure S3 .
Figure S3.Scatter plot of significant whole skin cis-meQTL effect betas against blood cis-meQTL effect betas from Min et al.,2 , coloured by the density of points.

Figure S4 .
Figure S4.Number of discovered eGenes in whole skin at a Benjamini-Hochberg FDR of 5% in cis-eQTL analyses with different numbers of PEER factors included as covariates.30 PEER factors were retained for our primary cis-eQTL analysis, to maximise eGene discovery whilst avoiding over-fitting.

Figure S5 .
Figure S5.Distance from lead eQTL SNP from conditional analysis to TSS of its eGene vs -log10(P-value) of the eQTL association.

Figure S6 .
FigureS6.Plots used to find the optimal value of p12 in eQTL-meQTL co-localisation analysis.NSNPs vs observed Posterior Probability of a Common Causal Variant for p12 values corresponding to 10%, 25%, 50%, and 75% probability of a causal cis-eQTL SNP also being a causal cis-meQTL SNP.The purple curve is the loess smoothed curve of these points.The orange dashed line is the loess smoothed curve of NSNPs vs the Prior Probability of a Common Causal Variant (not plotted), calculated using the formula described by Guo et al.,3 .

Figure S7 .
Figure S7.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and the mean GC content per RNA sample.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S8 .
Figure S8.Box plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and the RNAseq batch.A yellow background indicates a nominally significant difference between groups (Anova P<0.05).

Figure S9 .
Figure S9.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and chronological age.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S10 .
Figure S10.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and Endothelial Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S11 .
Figure S11.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and Fibroblast Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S12 .
Figure S12.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and Differentiated Keratinocyte Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S13 .
Figure S13.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and Undifferentiated Keratinocyte Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S14 .
Figure S14.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and Macrophage Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).

Figure S15 .
Figure S15.Scatter plots showing the correlation between the 30 PEER factors used as covariates in the skin eQTL analysis and T-Cell proportion as estimated from DNA methylation data.A yellow background indicates a nominally significant correlation (Pearson P<0.05).