Reducing the risk of false discovery enabling identification of biologically significant genome-wide methylation status using the HumanMethylation450 array

Abstract

BackgroundThe Illumina HumanMethylation450 BeadChip (HM450K) measures the DNA methylation of 485,512 CpGs in the human genome. The technology relies on hybridization of genomic fragments to probes on the chip. However, certain genomic factors may compromise the ability to measure methylation using the array such as single nucleotide polymorphisms (SNPs), small insertions and deletions (INDELs), repetitive DNA, and regions with reduced genomic complexity. Currently, there is no clear method or pipeline for determining which of the probes on the HM450K bead array should be retained for subsequent analysis in light of these issues.ResultsWe comprehensively assessed the effects of SNPs, INDELs, repeats and bisulfite induced reduced genomic complexity by comparing HM450K bead array results with whole genome bisulfite sequencing. We determined which CpG probes provided accurate or noisy signals. From this, we derived a set of high-quality probes that provide unadulterated measurements of DNA methylation.ConclusionsOur method significantly reduces the risk of false discoveries when using the HM450K bead array, while maximising the power of the array to detect methylation status genome-wide. Additionally, we demonstrate the utility of our method through extraction of biologically relevant epigenetic changes in prostate cancer.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2164-15-51) contains supplementary material, which is available to authorized users.

Systematic Mendelian randomization framework elucidates hundreds of CpG sites which may mediate the influence of genetic variants on disease

“…The multiple testing threshold was calculated as 0.05 divided by the derived number of independent tests. CpG sites for effects which survived this threshold were annotated based on evaluations of the 450K array (61,62). When only one genetic instrument was available MR effect estimates are based on the Wald ratio test (21): ${\hat{\beta }}_{\mathrm{normalW}\mathrm{normala}\mathrm{normall}\mathrm{normald}\mathrm{normal }\mathrm{normalr}\mathrm{normala}\mathrm{normalt}\mathrm{normali}\mathrm{normalo}}= \frac{{\hat{\beta }}_{Yfalse|Z}\mathrm{normal }}{{\hat{\beta }}_{Xfalse|Z}}$${se(\hat{\beta }}_{\mathrm{normalW}\mathrm{normala}\mathrm{normall}\mathrm{normald}\mathrm{normal }\mathrm{normalr}\mathrm{normala}\mathrm{normalt}\mathrm{normali}\mathrm{normalo}})\mathrm{normal }=\sqrt{\frac{\mathrm{s}\mathrm{e}({{\widehat{\beta }}_{Y|Z})\mathrm{ }}^2}{{{\widehat{\beta }}_{X|Z}}^2}+\mathrm{ }\frac{{{\widehat{\beta }}_{Y|Z}}^2\mathrm{s}\mathrm{e}({{\widehat{\beta }}_{X|Z})\mathrm{ }}^2}{{{\widehat{\beta }}_{X|Z}}^4}-\mathrm{ }\frac{2{\widehat{\beta }}_{Y|Z}\mathrm{c}\mathrm{o}\mathrm{v}({\widehat{\beta }}_{X|Z},{\widehat{\beta }}_{Y|Z})\mathrm{ }}{{{\widehat{\beta }}_{X|Z}}^3}\mathrm{ }}$ where $\widehat{\beta } Yfalse|Z$ is the coefficient of the genetic variant in the regression of the exposure (e.g.…”

Dopamine gene methylation patterns are associated with obesity markers and carbohydrate intake

“…Methylation data were peak‐based corrected for type I and type II bias and subsequently normalized using a categorical Subset Quantile Normalization method (Touleimat & Tost, 2012). Probes containing single nucleotide polymorphisms, those hybridizing to multiple genomic locations, or associated with X and Y chromosomes, were removed (Naeem et al., 2014; Nordlund et al., 2013). The ComBat normalization method was applied to adjusting for nonbiological experimental variation (Johnson, Li, & Rabinovic, 2007).…”

Epigenome-wide association study of asthma and wheeze in childhood and adolescence

“…All models were adjusted for sex, maternal smoking during pregnancy, maternal age, parity, maternal education derived from maternal questionnaires and unknown confounders using surrogate variables (by use of maximum 10 surrogate variables obtained from the sva R package [45]), as shown in Additional file 1: Table S2. Probes that were known to contain SNPs or be of low quality [46] were removed. The genomic inflation factor (lambda λ) and quantile-quantile (Q-Q) plots were used to compare the genome-wide distribution of P values with the expected null distribution.…”