Presentation is loading. Please wait.

Presentation is loading. Please wait.

Supplemental Figure 1. False trans association due to probe cross-hybridization and genetic polymorphism at single base extension site. (A) The Infinium.

Similar presentations


Presentation on theme: "Supplemental Figure 1. False trans association due to probe cross-hybridization and genetic polymorphism at single base extension site. (A) The Infinium."— Presentation transcript:

1 Supplemental Figure 1. False trans association due to probe cross-hybridization and genetic polymorphism at single base extension site. (A) The Infinium II probe designed to interrogate CpG site cg15830792 within NRBF2 gene was cross-hybridized to a homologous sequence within pseudogene NRBF2P3. Displayed was the output of alignment among probe sequence, target sequence within NRBF2 (input as bisulfite converted forward strand), and target sequence within NRBF2P3 (input as bisulfite converted reverse strand). Three mismatches, two at the 5’ end of probe and one within probe, were boxed in red. One within-probe cytosine in CpG was boxed in orange. Our probe filtering procedure removed probes potentially cross-hybridize to multiple genomic regions with ≤2 mismatches; therefore the probe was not removed. (B) The position for single base extension within NRBF2P3 contains SNP rs12731526 (T/C for reverse strand, boxed in blue), which is in perfect LD with the trans mQTL rs6690971. If the C allele of rs12731526 is heavily methylated within the pseudogene, then ratio of “unmethylated to methylated” estimated by single base extension will be similar to allele ratio of T/C (i.e., genotypes at rs12731526 or rs6690971 correlate with the assay readouts). Displayed was the alignment between target sequence within NRBF2 (forward strand, without bisulfite conversion) and that within NRBF2P3 (reverse strand, without bisulfite conversion). The positions of mismatches and within-probe CpG site were boxed red and orange respectively. The positions for single base extension were bracketed. Output from alignment with blastn: Probe 1 TCACCAAATTAAAAAATCCTTCCATTACTTCCATAAACCCCRAAATAAAC 50 NRBF2_F 258 TCACCAAATTAAAAAATCCTTCCATTACTTCCATAAACCCCRAAATAAAC 209 NRBF2P3_R 164 ACCAAATTAAAAAATCCTTCCATTACTTTCATAAACCCCRAAATAAAC 117 Target sequence without bisulfite conversion: NRBF2_F 208 [C]GCTTACCCCGGGGTCTATGGAAGTAATGGAAGGACCCCTCAACCTGGTGA 258 NRBF2P3_R 116 [T/C]GCTTACCCCGGGGTCTATGAAAGTAATGGAAGGACCCCTCAACCTGGC 164 v rs12731526 A B

2 Supplemental Figure 3. The effect sizes of local genetic epistasis. For CpG sites mapped to both additive and interacting mQTL, the adjusted association r 2 of the full model was plotted against that of the reduced model without the interaction terms, for CEU (blue) and YRI (orange).

3 Supplemental Figure 4. Population specificity of mQTL effects. Both common and population-specific mQTL were detected across the CEU and YRI samples. Test of genotype main effects across-population detected 61950 associations for 7615 CpG sites at 5% FDR. Test of population-specific genotype effects detected 5551 associations for 1216 CpG sites, among which 889 were also CpG sites with genotype main effect. (A) Even for common associations (genotype main effects), the within-population association r 2 was rather different between the two populations. (B-C) Examples of common mQTL (B) and population-specific mQTL (C). Orange points: YRI samples; black points: CEU samples.

4 Supplemental Figure 5. Local genetic epistasis contributes to population specificity of mQTL effects. (A-C)For CpG site cg11960243, the CEU-specific genotype effect of mQTL rs910320 (A) could be partially explained by genetic interaction with rs2071026 in CEU (B) but not in YRI (C). (D-F) For CpG site cg14373797, the YRI- specific genotype effect of mQTL rs3130211 (D) could be partially explained by genetic interaction with rs440841, which has greater genetic heterogeneity in CEU (E) than in YRI (F). In (A) and (D) orange points: YRI samples; black points: CEU samples.

5 Supplemental Figure 6. Genetic and epigenetic dissection of gene expression variation in the YRI samples. (A) mQTL were enriched for eQTL. For genes containing mQTL within the ± 100Kb regions, P-value distribution of associations between expression levels and mQTL (black) were compared with that between expression levels and all SNPs within the regions (grey). (B) Distribution of relative positions between CpG sites and the correlated genes for genetically dependent CpG-gene pairs. (C) Epigenetic regulation of gene expression. For cytosine modification-gene expression correlations detected at 5% FDR, the association r 2 between gene expression levels and residual cytosine modification levels were compared with the association r 2 between gene expression levels and unadjusted cytosine modification levels. The majority of correlations, after adjusted for genetic variations, remained to be significant. (D) Distribution of relative positions between CpG sites and the correlated genes for genetically independent CpG-gene pairs. In (B) and (D), grey bars represent counts and step-wise lines represent proportions.

6 Supplemental Figure 7. Genetic and non-genetic gene regulation mediated through cytosine modifications in the CEU samples. Co-localization of genetically dependent (A, B) and genetically independent (C, D) CpG sites with transcription factor binding sites (A, C) and histone markers (B, D). Facilitative sites were plotted in upper panels and repressive sites were plotted in lower panels. Counts of co-localization were represented by bars, for each type of peaks that exceeded 50% percentile of the corresponding null distributions. The aggregated counts of co-localization with transcription factor binding sites were colored by orange, while counts of co-localization with H3K4me3, H3K36me3, H3K9me3 and H3K27me3 peaks were colored by black, red, green and blue, respectively. Total counts of repressive or facilitative sites were displayed as step-wise lines.

7 Supplemental Figure 8. Genetic and non-genetic gene regulation mediated through cytosine modifications in the YRI samples. Co- localization of genetically dependent (A, B) and genetically independent (C, D) CpG sites with transcription factor binding sites (A, C) and histone markers (B, D). Facilitative sites were plotted in upper panels and repressive sites were plotted in lower panels. Counts of co-localization were represented by bars, for each type of peaks that exceeded 50% percentile of the corresponding null distributions. The aggregated counts of co-localization with transcription factor binding sites were colored by orange, while counts of co-localization with H3K4me3, H3K36me3, H3K9me3 and H3K27me3 peaks were colored by black, red, green and blue, respectively. Total counts of repressive or facilitative sites were displayed as step-wise lines.


Download ppt "Supplemental Figure 1. False trans association due to probe cross-hybridization and genetic polymorphism at single base extension site. (A) The Infinium."

Similar presentations


Ads by Google