Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray

DNA methylation, an important type of epigenetic modification in humans, participates in crucial cellular processes, such as embryonic development, X-inactivation, genomic imprinting and chromosome stability. Several platforms have been developed to study genome-wide DNA methylation. Many investigators in the field have chosen the Illumina Infinium HumanMethylation microarray for its ability to reliably assess DNA methylation following sodium bisulfite conversion. Here, we analyzed methylation profiles of 489 adult males and 357 adult females generated by the Infinium HumanMethylation450 microarray. Among the autosomal CpG sites that displayed significant methylation differences between the two sexes, we observed a significant enrichment of cross-reactive probes co-hybridizing to the sex chromosomes with more than 94% sequence identity. This could lead investigators to mistakenly infer the existence of significant autosomal sex-associated methylation. Using sequence identity cutoffs derived from the sex methylation analysis, we concluded that 6% of the array probes can potentially generate spurious signals because of co-hybridization to alternate genomic sequences highly homologous to the intended targets. Additionally, we discovered probes targeting polymorphic CpGs that overlapped SNPs. The methylation levels detected by these probes are simply the reflection of underlying genetic polymorphisms but could be misinterpreted as true signals. The existence of probes that are cross-reactive or of target polymorphic CpGs in the Illumina HumanMethylation microarrays can confound data obtained from such microarrays. Therefore, investigators should exercise caution when significant biological associations are found using these array platforms. A list of all cross-reactive probes and polymorphic CpGs identified by us are annotated in this paper.     

Different measures of “genome-wide” DNA methylation exhibit unique properties in placental and somatic tissues

DNA methylation of CpGs located in two types of repetitive elements—LINE1 (L1) and Alu—is used to assess “global” changes in DNA methylation in studies of human disease and environmental exposure. L1 and Alu contribute close to 30% of all base pairs in the human genome and transposition of repetitive elements is repressed through DNA methylation. Few studies have investigated whether repetitive element DNA methylation is associated with DNA methylation at other genomic regions, or the biological and technical factors that influence potential associations. Here, we assess L1 and Alu DNA methylation by Pyrosequencing of consensus sequences and using subsets of probes included in the Illumina Infinium HumanMethylation27 BeadChip array. We show that evolutionary age and assay method affect the assessment of repetitive element DNA methylation. Additionally, we compare Pyrosequencing results for repetitive elements to average DNA methylation of CpG islands, as assessed by array probes classified into strong, weak and non-islands. We demonstrate that each of these dispersed sequences exhibits different patterns of tissue-specific DNA methylation. Correlation of DNA methylation suggests an association between L1 and weak CpG island DNA methylation in some of the tissues examined. We caution, however, that L1, Alu and CpG island DNA methylation are distinct measures of dispersed DNA methylation and one should not be used in lieu of another. Analysis of DNA methylation data is complex and assays may be influenced by environment and pathology in different or complementary ways.     

Different measures of “genome-wide” DNA methylation exhibit unique properties in placental and somatic tissues

DNA methylation of CpGs located in two types of repetitive elements—LINE1 (L1) and Alu—is used to assess “global” changes in DNA methylation in studies of human disease and environmental exposure. L1 and Alu contribute close to 30% of all base pairs in the human genome and transposition of repetitive elements is repressed through DNA methylation. Few studies have investigated whether repetitive element DNA methylation is associated with DNA methylation at other genomic regions, or the biological and technical factors that influence potential associations. Here, we assess L1 and Alu DNA methylation by Pyrosequencing of consensus sequences and using subsets of probes included in the Illumina Infinium HumanMethylation27 BeadChip array. We show that evolutionary age and assay method affect the assessment of repetitive element DNA methylation. Additionally, we compare Pyrosequencing results for repetitive elements to average DNA methylation of CpG islands, as assessed by array probes classified into strong, weak and non-islands. We demonstrate that each of these dispersed sequences exhibits different patterns of tissue-specific DNA methylation. Correlation of DNA methylation suggests an association between L1 and weak CpG island DNA methylation in some of the tissues examined. We caution, however, that L1, Alu and CpG island DNA methylation are distinct measures of dispersed DNA methylation and one should not be used in lieu of another. Analysis of DNA methylation data is complex and assays may be influenced by environment and pathology in different or complementary ways.     

SWAN: Subset-quantile Within Array Normalization for Illumina Infinium HumanMethylation450 BeadChips

DNA methylation is the most widely studied epigenetic mark and is known to be essential to normal development and frequently disrupted in disease. The Illumina HumanMethylation450 BeadChip assays the methylation status of CpGs at 485,577 sites across the genome. Here we present Subset-quantile Within Array Normalization (SWAN), a new method that substantially improves the results from this platform by reducing technical variation within and between arrays. SWAN is available in the minfi Bioconductor package.     

Functional normalization of 450k methylation array data improves replication in large cancer studies

We propose an extension to quantile normalization that removes unwanted technical variation using control probes. We adapt our algorithm, functional normalization, to the Illumina 450k methylation array and address the open problem of normalizing methylation data with global epigenetic changes, such as human cancers. Using data sets from The Cancer Genome Atlas and a large case–control study, we show that our algorithm outperforms all existing normalization methods with respect to replication of results between experiments, and yields robust results even in the presence of batch effects. Functional normalization can be applied to any microarray platform, provided suitable control probes are available.

Electronic supplementary material

The online version of this article (doi:10.1186/s13059-014-0503-2) contains supplementary material, which is available to authorized users.     

Deep Bisulfite Sequencing of Aberrantly Methylated Loci in a Patient with Multiple Methylation Defects

NLRP7 is a maternal effect gene as maternal mutations in this gene cause recurrent hydatidiform moles, spontaneous abortions and stillbirths, whereas live births are very rare. We have studied a patient with multiple anomalies born to a mother with a heterozygous NLRP7 mutation. By array-based CpG methylation analysis of blood DNA from the patient, his parents and 18 normal controls on Illumina Infinium HumanMethylation27 BeadChips we found that the patient had methylation changes (delta ß ≥ 0.3) at many imprinted loci as well as at 87 CpGs associated with 85 genes of unknown imprinting status. Using a pseudoproband (permutation) approach, we found methylation changes at only 7-24 CpGs (mean 15; standard deviation 4.84) in the controls. Thus, the number of abberantly methylated CpGs in the patient is more than 14 standard deviations higher. In order to identify novel imprinted genes among the 85 conspicuous genes in the patient, we selected 19 (mainly hypomethylated) genes for deep bisulfite amplicon sequencing on the ROCHE/454 Genome Sequencer in the patient and at least two additional controls. These controls had not been included in the array analysis and were heterozygous for a single nucleotide polymorphism at the test locus, so that allele-specific DNA methylation patterns could be determined. Apart from FAM50B, which we proved to be imprinted in blood, we did not observe allele-specific DNA methylation at the other 18 loci. We conclude that the patient does not only have methylation defects at imprinted loci but (at least in blood) also an excess of methylation changes at apparently non-imprinted loci.     

ENmix: a novel background correction method for Illumina HumanMethylation450 BeadChip

The Illumina HumanMethylation450 BeadChip is increasingly utilized in epigenome-wide association studies, however, this array-based measurement of DNA methylation is subject to measurement variation. Appropriate data preprocessing to remove background noise is important for detecting the small changes that may be associated with disease. We developed a novel background correction method, ENmix, that uses a mixture of exponential and truncated normal distributions to flexibly model signal intensity and uses a truncated normal distribution to model background noise. Depending on data availability, we employ three approaches to estimate background normal distribution parameters using (i) internal chip negative controls, (ii) out-of-band Infinium I probe intensities or (iii) combined methylated and unmethylated intensities. We evaluate ENmix against other available methods for both reproducibility among duplicate samples and accuracy of methylation measurement among laboratory control samples. ENmix out-performed other background correction methods for both these measures and substantially reduced the probe-design type bias between Infinium I and II probes. In reanalysis of existing EWAS data we show that ENmix can identify additional CpGs, and results in smaller P-value estimates for previously-validated CpGs. We incorporated the method into R package ENmix, which is freely available from Bioconductor website.     

An evaluation of analysis pipelines for DNA methylation profiling using the Illumina HumanMethylation450 BeadChip platform

The proper identification of differentially methylated CpGs is central in most epigenetic studies. The Illumina HumanMethylation450 BeadChip is widely used to quantify DNA methylation; nevertheless, the design of an appropriate analysis pipeline faces severe challenges due to the convolution of biological and technical variability and the presence of a signal bias between Infinium I and II probe design types. Despite recent attempts to investigate how to analyze DNA methylation data with such an array design, it has not been possible to perform a comprehensive comparison between different bioinformatics pipelines due to the lack of appropriate data sets having both large sample size and sufficient number of technical replicates. Here we perform such a comparative analysis, targeting the problems of reducing the technical variability, eliminating the probe design bias and reducing the batch effect by exploiting two unpublished data sets, which included technical replicates and were profiled for DNA methylation either on peripheral blood, monocytes or muscle biopsies. We evaluated the performance of different analysis pipelines and demonstrated that: (1) it is critical to correct for the probe design type, since the amplitude of the measured methylation change depends on the underlying chemistry; (2) the effect of different normalization schemes is mixed, and the most effective method in our hands were quantile normalization and Beta Mixture Quantile dilation (BMIQ); (3) it is beneficial to correct for batch effects. In conclusion, our comparative analysis using a comprehensive data set suggests an efficient pipeline for proper identification of differentially methylated CpGs using the Illumina 450K arrays.     

Collaboration between CpG sites is needed for stable somatic inheritance of DNA methylation states

Inheritance of 5-methyl cytosine modification of CpG (CG/CG) DNA sequences is needed to maintain early developmental decisions in vertebrates. The standard inheritance model treats CpGs as independent, with methylated CpGs maintained by efficient methylation of hemimethylated CpGs produced after DNA replication, and unmethylated CpGs maintained by an absence of de novo methylation. By stochastic simulations of CpG islands over multiple cell cycles and systematic sampling of reaction parameters, we show that the standard model is inconsistent with many experimental observations. In contrast, dynamic collaboration between CpGs can provide strong error-tolerant somatic inheritance of both hypermethylated and hypomethylated states of a cluster of CpGs, reproducing observed stable bimodal methylation patterns. Known recruitment of methylating enzymes by methylated CpGs could provide the necessary collaboration, but we predict that recruitment of demethylating enzymes by unmethylated CpGs strengthens inheritance and allows CpG islands to remain hypomethylated within a sea of hypermethylation.     

Genome-wide DNA methylation profiling in rheumatoid arthritis identifies disease-associated methylation changes that are distinct to individual T- and B-lymphocyte populations

Changes to the DNA methylome have been described in patients with rheumatoid arthritis (RA). In previous work, we reported genome-wide methylation differences in T-lymphocyte and B-lymphocyte populations from healthy individuals. Now, using HumanMethylation450 BeadChips to interrogate genome-wide DNA methylation, we have determined disease-associated methylation changes in blood-derived T- and B-lymphocyte populations from 12 female patients with seropositive established RA, relative to 12 matched healthy individuals. Array data were analyzed using NIMBL software and bisulfite pyrosequencing was used to validate array candidates. Genome-wide DNA methylation, determined by analysis of LINE-1 sequences, revealed higher methylation in B-lymphocytes compared with T-lymphocytes (P ≤ 0.01), which is consistent with our findings in healthy individuals. Moreover, loci-specific methylation differences that distinguished T-lymphocytes from B-lymphocytes in healthy individuals were also apparent in RA patients. However, disease-associated methylation differences were also identified in RA. In these cases, we identified 509 and 252 CpGs in RA-derived T- and B-lymphocytes, respectively, that showed significant changes in methylation compared with their cognate healthy counterparts. Moreover, this included a restricted set of 32 CpGs in T-lymphocytes and 20 CpGs in B-lymphocytes (representing 15 and 10 genes, respectively, and including two, MGMT and CCS, that were common to both cell types) that displayed more substantial changes in methylation. These changes, apparent as hyper- or hypo-methylation, were independently confirmed by pyrosequencing analysis. Validation by pyrosequencing also revealed additional sites in some candidate genes that also displayed altered methylation in RA. In this first study of genome-wide DNA methylation in individual T- and B-lymphocyte populations in RA patients, we report disease-associated methylation changes that are distinct to each cell type and which support a role for discrete epigenetic regulation in this disease.     

Genome-Wide DNA Methylation Analysis of Systemic Lupus Erythematosus Reveals Persistent Hypomethylation of Interferon Genes and Compositional Changes to CD4+ T-cell Populations

Systemic lupus erythematosus (SLE) is an autoimmune disease with known genetic, epigenetic, and environmental risk factors. To assess the role of DNA methylation in SLE, we collected CD4+ T-cells, CD19+ B-cells, and CD14+ monocytes from 49 SLE patients and 58 controls, and performed genome-wide DNA methylation analysis with Illumina Methylation450 microarrays. We identified 166 CpGs in B-cells, 97 CpGs in monocytes, and 1,033 CpGs in T-cells with highly significant changes in DNA methylation levels (p<1×10⁻⁸) among SLE patients. Common to all three cell-types were widespread and severe hypomethylation events near genes involved in interferon signaling (type I). These interferon-related changes were apparent in patients collected during active and quiescent stages of the disease, suggesting that epigenetically-mediated hypersensitivity to interferon persists beyond acute stages of the disease and is independent of circulating interferon levels. This interferon hypersensitivity was apparent in memory, naïve and regulatory T-cells, suggesting that this epigenetic state in lupus patients is established in progenitor cell populations. We also identified a widespread, but lower amplitude shift in methylation in CD4+ T-cells (>16,000 CpGs at FDR<1%) near genes involved in cell division and MAPK signaling. These cell type-specific effects are consistent with disease-specific changes in the composition of the CD4+ population and suggest that shifts in the proportion of CD4+ subtypes can be monitored at CpGs with subtype-specific DNA methylation patterns.     

Genome-wide DNA methylation profiling in rheumatoid arthritis identifies disease-associated methylation changes that are distinct to individual T- and B-lymphocyte populations

Changes to the DNA methylome have been described in patients with rheumatoid arthritis (RA). In previous work, we reported genome-wide methylation differences in T-lymphocyte and B-lymphocyte populations from healthy individuals. Now, using HumanMethylation450 BeadChips to interrogate genome-wide DNA methylation, we have determined disease-associated methylation changes in blood-derived T- and B-lymphocyte populations from 12 female patients with seropositive established RA, relative to 12 matched healthy individuals. Array data were analyzed using NIMBL software and bisulfite pyrosequencing was used to validate array candidates. Genome-wide DNA methylation, determined by analysis of LINE-1 sequences, revealed higher methylation in B-lymphocytes compared with T-lymphocytes (P ≤ 0.01), which is consistent with our findings in healthy individuals. Moreover, loci-specific methylation differences that distinguished T-lymphocytes from B-lymphocytes in healthy individuals were also apparent in RA patients. However, disease-associated methylation differences were also identified in RA. In these cases, we identified 509 and 252 CpGs in RA-derived T- and B-lymphocytes, respectively, that showed significant changes in methylation compared with their cognate healthy counterparts. Moreover, this included a restricted set of 32 CpGs in T-lymphocytes and 20 CpGs in B-lymphocytes (representing 15 and 10 genes, respectively, and including two, MGMT and CCS, that were common to both cell types) that displayed more substantial changes in methylation. These changes, apparent as hyper- or hypo-methylation, were independently confirmed by pyrosequencing analysis. Validation by pyrosequencing also revealed additional sites in some candidate genes that also displayed altered methylation in RA. In this first study of genome-wide DNA methylation in individual T- and B-lymphocyte populations in RA patients, we report disease-associated methylation changes that are distinct to each cell type and which support a role for discrete epigenetic regulation in this disease.     

Assessment of methylation level prediction accuracy in methyl-DNA immunoprecipitation and sodium bisulfite based microarray platforms

In this study, we verified the accuracy of two array methods—methylated DNA immunoprecipitation coupled with CpG island microarrays (MeDIP-CGI-arrays) and sodium bisulfite conversion based microarrays (BC-arrays)—in predicting regional methylation levels as measured by pyrosequencing of bisulfite converted DNA (BC-pyrosequencing). To test the accuracy of these methods we used the Agilent Human CpG island and the Illumina HumanMethylation27 microarrays respectively, and compared microarray outputs to the data from targeted BC-pyrosequencing assays from several genomic regions of corresponding samples. We observed relatively high correlation with BC-pyrosequencing data for both array platforms, R = 0.87 for BC-Array and R = 0.79 for MeDIP-CGI array. However, MeDIP-CGI array were less reliable in predicting intermediate levels of DNA methylation. Several bioinformatics strategies, to ameliorate the performance of the MeDIP-CGI-Arrays did not improve the correlation with BC-pyrosequencing data. The high scalability, low cost and simpler analysis of BC-arrays, together with the recent extended coverage may make them a more versatile methylation analysis tool.     

Constitutively methylated CpG dinucleotides as mutation hot spots in the retinoblastoma gene (RB1).

A wide spectrum of mutations, ranging from point mutations to large deletions, have been described in the retinoblastoma gene (RB1). Mutations have been found throughout the gene; however, these genetic alterations do not appear to be homogeneously distributed. In particular, a significant proportion of disease-causing mutations results in the premature termination of protein synthesis, and the majority of these mutations occur as C-->T transitions at CpG dinucleotides (CpGs). Such recurrent CpG mutations, including those found in RB1, are likely the result of the deamination of 5-methylcytosine within these CpGs. In the present study, we used the sodiumbisulfite conversion method to detect cytosine methylation in representative exons of RB1. We analyzed DNA from a variety of tissues and specifically targeted CGA codons in RB1, where recurrent premature termination mutations have been reported. We found that DNA methylation within RB1 exons 8, 14, 25, and 27 appeared to be restricted to CpGs, including six CGA codons. Other codons containing methylated cytosines have not been reported to be mutated. Therefore, disease-causing mutations at CpGs in RB1 appear to be determined by several factors, including the constitutive presence of DNA methylation at cytosines within CpGs, the specific codon within which the methylated cytosine is located, and the particular region of the gene within which that codon resides.     

Global Analysis of DNA Methylation Variation in Adipose Tissue from Twins Reveals Links to Disease-Associated Variants in Distal Regulatory Elements

Epigenetic modifications such as DNA methylation play a key role in gene regulation and disease susceptibility. However, little is known about the genome-wide frequency, localization, and function of methylation variation and how it is regulated by genetic and environmental factors. We utilized the Multiple Tissue Human Expression Resource (MuTHER) and generated Illumina 450K adipose methylome data from 648 twins. We found that individual CpGs had low variance and that variability was suppressed in promoters. We noted that DNA methylation variation was highly heritable (h²_median = 0.34) and that shared environmental effects correlated with metabolic phenotype-associated CpGs. Analysis of methylation quantitative-trait loci (metQTL) revealed that 28% of CpGs were associated with nearby SNPs, and when overlapping them with adipose expression quantitative-trait loci (eQTL) from the same individuals, we found that 6% of the loci played a role in regulating both gene expression and DNA methylation. These associations were bidirectional, but there were pronounced negative associations for promoter CpGs. Integration of metQTL with adipose reference epigenomes and disease associations revealed significant enrichment of metQTL overlapping metabolic-trait or disease loci in enhancers (the strongest effects were for high-density lipoprotein cholesterol and body mass index [BMI]). We followed up with the BMI SNP rs713586, a cg01884057 metQTL that overlaps an enhancer upstream of ADCY3, and used bisulphite sequencing to refine this region. Our results showed widespread population invariability yet sequence dependence on adipose DNA methylation but that incorporating maps of regulatory elements aid in linking CpG variation to gene regulation and disease risk in a tissue-dependent manner.