Tissue-Specific Effects of Genetic and Epigenetic Variation on Gene Regulation and Splicing
Figure 6
Associations between DNA methylation and alternative splicing.
(A) Methylation sites associated to alternative splicing (asQTMs) are significantly enriched for active promoters, elongation marks, CTCF peaks, middle exons, transcription factor binding (TFB) peak motifs, DNase I hypersensitive sites (HSs) and enhancer marks; as shown by the higher percentage of overlap in observed asQTMs compared to null sites. One star indicates P < 0.05, two stars indicate P < 5E-04, Fisher’s exact test (see also S6 Table). For T-cells there was no ChIP-seq data available at the time of analysis so the data of an LCL was used instead (see Materials and Methods). (B) Effect sizes of the union of asQTMs (best methylation site per exon-exon link) for each pair of cell-types are plotted. Effect size is measured as the slope of the linear regression of alternative splicing levels given methylation levels on scaled values. Hence, effect sizes are quantified in terms of number of alternative splicing standard deviations changed by every methylation standard deviation change. Black dots depict associations significant in both cell-types compared; orange, blue and purple dots are associations significant in only fibroblasts, LCLs and T-cells, respectively, within each pair compared. Coefficient of determination R2, reflecting the proportion of effect size variance in one cell-type explained by the other cell-type, is shown in the top left corner of each plot (S1 Table). Given lower sample size in T-cells, there are many less significant associations and hence, it is difficult to assess cell-type specificity between this cell-type and the rest. Fibroblasts and LCLs show a significant fraction of shared correlations between DNA methylation and alternative splicing, and also a high abundance of tissue-specific effects.