Skip to main content

Table 2 Blood-based genome-wide meQTL studies (sample size > 100) in whole blood or blood-derived cell samples

From: Genetic impacts on DNA methylation: research findings and future perspectives

Ref. Sample sizea Methylation assay Genotyping assay Significant CpGsb Significance thresholdc Remarks
[50] 27,750 Illumina 450K 1000Gd cis (≤ 1 Mbp), 43.5% trans (> 1 Mbp), 4.5% pcis<1×10−8,ptrans<1×10−14 Study design: two-phase meta-analysis in a total of 36 cohorts. Additional analysis: replication of 188,017 meQTLs in external sample (76% for cis and 79% for trans).
[35] 4170 Illumina 450K Affymetrix 500K and 50K MIP, 1000Gd cis (≤ 1 Mbp), 29.3% trans (> 1 Mbp), 3.3% FWER1 < 5% (pcis<2×10−11,ptrans<1.5×10−14) Additional analysis: replication of effect direction in two external samples (81–99%).
[136] 429 Illumina 450K GOLDN studye cis (≤ 1 Mbp), 0.3% FWER1 < 5% (p<5.6×10−10) Study design: response meQTL, with log transformed post-/pre-treatment methylation values.
[38] 1111 Illumina EPIC Illumina CoreExome, 1000Gd trans, 12.2% FWER1 < 5% (p<6.5×10−14) Effect size: mean methylation change per additional reference allele of 3.46% (s=3.01%).
[60] 156 [monocytes] (two samples) Illumina EPIC Illumina Omni, WGS, 1000Gd cis (≤ 100 kbp), 12.6% FDR3 < 5% (p<1×10−5) Study design: separate meQTLs discovery for the two samples, joint results reported. Additional analysis: trans association restricted to SNPs in TFs genes vicinity.
[53] 1980 (two samples) Illumina 450K Illumina 610-Quad cis (≤ 2 Mbp), 15.4–15.7% trans (> 2 Mbp), 0.5–0.6% FWER1 < 5% (pcis<1×10−11,ptrans<1×10−13) Study design: separate meQTLs discovery for the two samples. Additional analysis: replication of CpGs in both samples (13.3% for cis and 0.5% for trans).
[93] 337 Illumina 450K Illumina CytoSNP, 1000Gd cis (≤ 500 kbp), 18.3% FDR3 < 1%  
[157] 729 Illumina 450K Illumina Exome, Hap300 and Omni, 1000Gd cis (≤ 500 kbp), 12.1% FWER1 < 5% (p<9.4×10−11) Additional analysis: variance meQTLs.
[85] 460 [cord blood and whole blood] (two samples) Illumina 450K Illumina Omni, 1000Gd cis (≤ 0.5–1 Mbp), 8.1–19.2% FDR3 < 5% Study design: separate meQTLs discovery for two samples with different parameters. Additional analysis: meQTLs co-localization with results from two published studies.
[51] 3,841 Illumina 450K CODAM34, LLD9, LLS38, NTR12, and RS studiese, GoNLd cis (≤ 250 kbp), 34.4% FDR3 < 5% (p<1.4×10−4) Additional analysis: trans association restricted to 6111 informative SNPs.
[65] 744 [T cells and whole blood] (two samples) MCC-seq, WGBS WGS, Illumina Omni, inferred from WGBS, 1000Gd cis (≤ 250 kbp), 16.0–18.1% FDR4 < 10% Study design: separate meQTLs discovery for the two samples. Additional analysis: meQTL in visceral adipose tissue samples, ASM analyses and genotype-independent tests, and validation on Illumina 450K.
[78] 525 [neutrophils, monocytes and T cells] (three samples) Illumina 450K WGS cis (≤ 500 kbp), 9.9% FWER1 < 5% Study design: separate meQTLs discovery for the three samples, mean results reported.
[37] 3948 [cord blood and whole blood] (five samples) Illumina 450K Illumina Hap550 and 660W, 1000Gd cis (≤ 1 Mbp), 6.1–8.0% trans(>1 Mbp), 0.5–0.7% FWER1 < 0.2% (p<1×10−14) Study design: separate meQTLs discovery for the five samples. Additional analysis: replication of CpGs in pairwise comparisons (83–98% for cis and 88–98% for trans).
[54] 850 Illumina450K Illumina Hap550, Exon510, 1M and 1M-Duo cis (≤ 50 kbp), 13.8% FWER1 < 5% (p<2.6×10−9)  
[55] 1748 [lymphocytes] (cancer-case and control samples) Illumina 450K OFCCRe, Affymetrix 500K cis (≤ 1 Mbp), 13.9% trans(> 1 Mbp), 0.5% FDR4 < 5% (pcis<4.8×10−10,ptrans<3.2×10−13) Study design: joint meQTLs discovery for the two samples; trans analysis excluded CpG sites with cis meQTLs. Additional analysis: replication of meQTL-CpG pairs in two external samples (83.6% for trans).
[52] 697 MBD-seq Affymetrix SNP 5.0 and 6.0, Illumina Omni, 1000Gd cis (≤ 1 Mbp), 15% trans(> 1 Mbp), 0.1% FDR4 < 1% Additional analysis: replication of findings in one sample of schizophrenia cases (π1= 95% for cis, 98.7% for trans same chromosome, and 99.3% for trans different chromosome).
[86] 264 [cord blood and whole blood] (three samples) Illumina 27K Illumina Omni, Affymetrix SNP 5.0 and 6.0, HapMapd cis (≤50 kbp), 0.7–1.5% FWER2 <5% Study design: separate meQTLs discovery for the three samples. Additional analysis: replication of meQTL-CpG pairs in pairwise comparisons (17.8–69.5%); meQTL in four brain regions samples.
[59] 177 [T cells and LCL] Illumina 450K Illumina Omni, 1000Gd cis (≤ 5 kbp), 5.4–7.8% FDR3 < 10% Study design: separate meQTLs discovery for the two samples. Additional analysis: meQTL in fibroblasts sample.
[32] 171 Illumina 27K Illumina Hap300, 610-Quad, 1M-Duo and 1.2M-Duo, HapMapd cis (≤ 50), 6.3% FDR3 < 5% (p<1×10−5)  
  1. aIf not specified, the sample type is whole blood. If more than one sample per analysis, the pooled size and number of samples is reported
  2. bIn parenthesis, maximum or minimum distances are indicated for cis and trans analysis, respectively. The range of results is presented if more than one analysis was done (unless otherwise stated)
  3. cMultiple-testing criteria, with the corresponding p-value threshold for cis and trans meQTLs (where it differs). Different approaches to estimate FWER and FDR are as follows:
  4. 1FWER based on Bonferroni correction
  5. 2FWER based on Holm-Bonferroni correction
  6. 3FDR based on permutations
  7. 4FDR based on Benjamini-Hochberg correction
  8. dReference panel for imputations
  9. eDatabase or biobank
  10. FWER family-wise error rate, FDR false discovery rate, LCL lymphoblastoid cell lines, WGS whole genome sequencing, MCC-seq methylC-capture sequencing, WGBS whole genome bisulfite sequencing, MBD-seq methyl-CpG-binding domain sequencing, 1000G 1000 genotypes, GoNL Genome of the Netherlands, TF transcription factor, ASM allele-specific methylation