A data-driven approach to preprocessing Illumina 450K methylation array data
Open accessAbstract
BACKGROUND: As the most stable and experimentally accessible epigenetic mark, DNA methylation is of great interest to the research community. The landscape of DNA methylation across tissues, through development and in disease pathogenesis is not yet well characterized. Thus there is a need for rapid and cost effective methods for assessing genome-wide levels of DNA methylation. The Illumina Infinium HumanMethylation450 (450K) BeadChip is a very useful addition to the available methods for DNA methylation analysis but its complex design, incorporating two different assay methods, requires careful consideration. Accordingly, several normalization schemes have been published. We have taken advantage of known DNA methylation patterns associated with genomic imprinting and X-chromosome inactivation (XCI), in addition to the performance of SNP genotyping assays present on the array, to derive three independent metrics which we use to test alternative schemes of correction and normalization. These metrics also have potential utility as quality scores for datasets. RESULTS: The standard index of DNA methylation at any specific CpG site is β = M/(M + U + 100) where M and U are methylated and unmethylated signal intensities, respectively. Betas (βs) calculated from raw signal intensities (the default GenomeStudio behavior) perform well, but using 11 methylomic datasets we demonstrate that quantile normalization methods produce marked improvement, even in highly consistent data, by all three metrics. The commonly used procedure of normalizing betas is inferior to the separate normalization of M and U, and it is also advantageous to normalize Type I and Type II assays separately. More elaborate manipulation of quantiles proves to be counterproductive. CONCLUSIONS: Careful selection of preprocessing steps can minimize variance and thus improve statistical power, especially for the detection of the small absolute DNA methylation changes likely associated with complex disease phenotypes. For the convenience of the research community we have created a user-friendly R software package called wateRmelon, downloadable from bioConductor, compatible with the existing methylumi, minfi and IMA packages, that allows others to utilize the same normalization methods and data quality tests on 450K data.
AI snapshot
Written from the paper’s full text.
TL;DR
Using 11 Illumina 450K datasets (total n=696), the authors derive three performance metrics from imprinted DMRs, SNP probes, and X-chromosome inactivation and use them to rank 15 preprocessing/normalization methods. They find that dasen—background adjustment plus separate between-array quantile normalization of methylated and unmethylated Type I and Type II intensities—performs best overall.
Key findings
- Across 11 datasets and 15 methods, dasen had the best average rank (2; Type I 2.5, Type II 1.5), followed by danet (3.5) and daten1 (4); raw averaged 8.75, betaqn 13.5, Fuks 10.75, Tost 13.5, and SWAN 12.5.
- Normalizing M and U separately outperformed normalizing β (naten vs betaqn), and separate Type I/Type II normalization improved metrics; more elaborate segmented QN schemes (Tost, SWAN) were counterproductive.
- In Cohort 1Ai, a sex-difference t-test at Bonferroni-corrected p<0.05 yielded 9,796 significant probes, of which 8,969 were X-chromosome loci, recovering 80% of X-chromosome probes; dasen added 296 (3%) sex differences compared with raw.
- The metrics used 237 iDMR probes expected to have β=0.5, 65 SNP probes, and 11,232 X-chromosome features.
Methods
- 11 450K datasets (total n=696) from post-mortem brain regions and whole blood; 500 ng genomic DNA was bisulfite-treated, run on Illumina HumanMethylation450 BeadChips, and raw signal intensities were extracted with GenomeStudio without background correction or normalization.
- Three performance metrics were calculated: DMRSE from 237 imprinted DMR probes expected at β=0.5; GCOSE from 65 SNP probes partitioned by k-means into three genotype clusters; and Seabird = 1−AUC from ROC analysis predicting X-chromosome location using sex-difference p-values.
- Fifteen preprocessing methods were ranked by mean ranks across metrics and datasets for Type I and Type II probes separately; analyses used R 2.15.0 and Bioconductor 2.12, with functions packaged as wateRmelon, and data deposited in GEO under GSE43414.
Limitations
- The authors state that the three metrics may miss some aspect of performance; the datasets analyzed were relatively homogeneous brain or whole blood, multi-tissue datasets were not extensively addressed, and more heterogeneous datasets may require different handling.
- The authors also state that only QN-based methods were tested, and a different approach such as scaling could possibly work better.
- Evident from the text: less than 1% of individual samples were excluded as technical failures based on atypical raw intensity distribution and poor correlation with other samples, and the ranking relies on array-derived natural controls rather than external spike-in validation.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
15 of 21 cited sources checked · 7 supported · 0 flagged · accuracy not yet measured
- Premisecitation
Basic analysis of β values works well for detecting large DNA methylation differences, such as between tissues or tumor-normal pairs.
“The β values are indeed fairly stable and the basic analysis method works well, particularly for the detection of large differences, for example between tissues or tumor-normal pairs [8].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [8] Carén H, Djos A, Nethander M, Sjöberg RM, Kogner P, Enström C, Nilsson S, Martinsson T: Identification of epigenetically regulated genes that predict patient outcome in neuroblastoma. BMC Cancer. 2011, 11: 66-10.1186/1471-2407-11-66. [http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3045360&tool=pmcentrez&rendertype=abstract] source
The Methods define β values, and the Results (Differential methylation analysis; page unavailable) show that comparing them detected differences between neuroblastoma patient groups. The Discussion describes DNA methylation measurement as robust and stable. The paper does not establish that basic β-value analysis works particularly well for large differences between tissues or tumor–normal pairs. (LLM, conf=0.83)
Experimental — accuracy not yet measured
- Premisecitation
Autosomal sex-specifically methylated loci are very few in number.
“True, autosomal sex-specifically methylated loci are very small in number”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [17] Chen Ya, Choufani S, Ferreira JC, Grafodatskaya D, Butcher DT, Weksberg R: Sequence overlap between autosomal and sex-linked probes on the Illumina HumanMethylation27 microarray. Genomics. 2011, 97 (4): 214-222. 10.1016/j.ygeno.2010.12.004. [http://www.ncbi.nlm.nih.gov/pubmed/21211562] source
Section 2.1 reports that, among 26,486 autosomal CpGs tested in blood lymphocytes, only seven of the 19 apparent sex-specific findings were not considered technical artifacts. That supports the claim that such loci were very few in this study, but not an unqualified claim about all autosomal loci or tissues. (LLM, conf=0.84)
Experimental — accuracy not yet measured
- Interpretationcitation
Tests of distribution similarity between technical replicates, such as the Kolmogorov-Smirnov test, are potentially misleading because samples can be identically distributed but uncorrelated.
“tests of distribution similarity between technical replicates(for example the Kolmogorov-Smirnov test used by Maksimovic et al[6]) are also potentially misleading because samples can be identically distributed but uncorrelated.”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [6] Maksimovic J, Gordon L, Oshlack A: SWAN: Subset quantile within-array normalization for Illumina Infinium HumanMethylation450 BeadChips. Genome Biol. 2012, 13 (6): R44-10.1186/gb-2012-13-6-r44. [http://www.ncbi.nlm.nih.gov/pubmed/22703947] source
- Interpretationcitation
Disappearance of batch effects is a useful but insufficient indicator of normalization performance because it does not show whether true differences can still be detected.
“the disappearance of batch effects is a useful indicator [12], but not sufficient as a performance indicator because it does not indicate whether true differences can”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [12] Sun Z, Chai HS, Wu Y, White WM, Donkena KV, Klein CJ, Garovic VD, Therneau TM, Kocher JPa: Batch effect correction for genome-wide methylation data with Illumina Infinium platform. BMC Med Genomics. 2011, 4: 84-10.1186/1755-8794-4-84. [http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3265417&tool=pmcentrez&rendertype=abstract] source
- Premisecitation
Autosomal sex-specifically methylated loci may represent autosomal probes mapping to sex-chromosome sequences.
“and potentially represent autosomal probes mapping to sequences on the sex-chromosome [17].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [17] Chen Ya, Choufani S, Ferreira JC, Grafodatskaya D, Butcher DT, Weksberg R: Sequence overlap between autosomal and sex-linked probes on the Illumina HumanMethylation27 microarray. Genomics. 2011, 97 (4): 214-222. 10.1016/j.ygeno.2010.12.004. [http://www.ncbi.nlm.nih.gov/pubmed/21211562] source
- Interpretationcitation
Profile correlations are insensitive to normalization problems because they are dominated by the majority of probes that do not show true differences.
“Profile correlations are also insensitive to the potential problems of normalization because they are dominated by the majority of probes which do not show true differences”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [14] Roessler J, Ammerpohl O, Gutwein J, Hasemeier B, Anwar SL, Kreipe H, Lehmann U: Quantitative cross-validation and content analysis of the 450k DNA methylation array from Illumina, Inc. BMC Res Notes. 2012, 5: 210-10.1186/1756-0500-5-210. [http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3420245&tool=pmcentrez&rendertype=abstract] source
In Findings > Cross-validation of 450k methylation array and pyrosequencing (page not provided), the paper says correlations can be misleading because data points cluster near 0% and 100% methylation, masking poorer agreement in the middle range. It does not say this makes profile correlations insensitive to normalization problems, or that the dominant probes are those without true differences. (LLM, conf=0.84)
Experimental — accuracy not yet measured
- Premisecitation
Type I and Type II probes perform differently.
“Type I and Type II probes perform differently.”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [5] Dedeurwaerder S, Defrance M, Calonne E, Sotiriou C, Fuks F: Evaluation of the Infinium Methylation 450K technology. Epigenetics. 2011, 3 (6): 771-784. source
- Interpretationreasoning
Where array features overlap iDMRs, monoallelic methylation at approximately 50% (β=0.5) is expected.
“where array features overlap these, we would expect monoallelic (therefore 50% (β=0.5)) methylation.”
Quote in the paperSubstantial claim1 of 2 sources checked- cites [15] Schulz R, Woodfine K, Menheniott TR, Bourc’his D, Bestor T, Oakey RJ: WAMIDEX: a web atlas of murine genomic imprinting and differential expression. Epigenetics. 2008, 3 (2): 89-96. 10.4161/epi.3.2.5900. [http://www.landesbioscience.com/journals/epi/article/5900/] source
The Introduction describes methylation differences between maternal and paternal alleles at imprinting control regions, and the Methods say most such regions are methylated on the maternal allele. That supports the monoallelic premise, but the paper does not give an expected 50% methylation value or β=0.5 reading for array features overlapping iDMRs; its described microarrays measure gene expression. (LLM, conf=0.82)
Experimental — accuracy not yet measured - cites [16] Schulz R: Wamidex, accessed 26 Jan2012. [https://atlas.genetics.kcl.ac.uk/]
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Interpretationcitation
It may be advantageous to minimize the differences between Type I and Type II probes so that the ranking of potential differentially methylated loci is more accurate.
“it may be advantageous to minimize the differences so that the ranking of potential differentially methylated loci is more accurate [5].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [5] Dedeurwaerder S, Defrance M, Calonne E, Sotiriou C, Fuks F: Evaluation of the Infinium Methylation 450K technology. Epigenetics. 2011, 3 (6): 771-784. source
The Discussion (p. 11) says correcting differences between Type I and Type II data improves accuracy and facilitates detection of differentially methylated cytosines. The paper does not assess whether correction makes the ranking of candidate loci more accurate. (LLM, conf=0.84)
Experimental — accuracy not yet measured
- Premisecitation
Previous DNA methylation analysis methods have been assessed using DNA methylation differences between experimental groups verified by an independent method.
“Previous DNA methylation data analysis methods have been assessed using the DNA methylaton differences between experimental groups verified using an independent method [5].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [5] Dedeurwaerder S, Defrance M, Calonne E, Sotiriou C, Fuks F: Evaluation of the Infinium Methylation 450K technology. Epigenetics. 2011, 3 (6): 771-784. source
- Premisecitation
Stable iDMRs have been characterized for 25 human imprinting regions.
“Stable iDMRs have been characterized for 25 human imprinting regions [15,16]”
Quote in the paperSubstantial claim1 of 2 sources checked- cites [15] Schulz R, Woodfine K, Menheniott TR, Bourc’his D, Bestor T, Oakey RJ: WAMIDEX: a web atlas of murine genomic imprinting and differential expression. Epigenetics. 2008, 3 (2): 89-96. 10.4161/epi.3.2.5900. [http://www.landesbioscience.com/journals/epi/article/5900/] source
The Introduction describes imprinting control regions and reports approximately 50 human imprinted genes, but it does not give a count of human regions with stable iDMRs. Searches for “25 human imprinting regions” and “human differentially methylated regions” found no passage in our copy supporting the claimed 25 regions. (LLM, conf=0.72)
Experimental — accuracy not yet measured - cites [16] Schulz R: Wamidex, accessed 26 Jan2012. [https://atlas.genetics.kcl.ac.uk/]
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Findingown data
It is advantageous to normalize Type I and Type II assays separately.
“it is also advantageous to normalize Type I and Type II assays separately”
Quote in the paperSubstantial claim - Findingown data
The commonly used procedure of normalizing betas is inferior to separate normalization of M and U.
“The commonly used procedure of normalizing betas is inferior to the separate normalization of M and U”
Quote in the paperSubstantial claim - Findingown data
Using 11 methylomic datasets, quantile normalization methods produce marked improvement by all three metrics, even in highly consistent data.
“using 11 methylomic datasets we demonstrate that quantile normalization methods produce marked improvement, even in highly consistent data, by all three metrics”
Quote in the paperSubstantial claim - Findingown data
More elaborate manipulation of quantiles is counterproductive.
“More elaborate manipulation of quantiles proves to be counterproductive”
Quote in the paperSubstantial claim - Findingown data
In the density plot of raw β values, the two peaks of Type II probes, representing methylated and unmethylated CpG sites, are compressed toward β=0.5.
“In the density plot of raw β values (Figure 1), the two peaks of the Type II probes (representing methylated and unmethylated CpG sites) are compressed toward β=0.5”
Quote in the paperSubstantial claim - Findingnone given
Betas calculated from raw signal intensities, the default GenomeStudio behavior, perform well.
“Betas (βs) calculated from raw signal intensities (the default GenomeStudio behavior) perform well”
Quote in the paperSubstantial claim - Findingown data
Density plots of raw β values show that Type I and Type II probes have different distributions.
“Density plots of the raw β values confirm that Type I and Type II probes have different distributions (Figure 1)”
Quote in the paperSubstantial claim - Findingown data
For Cohort 1Ai, dasen gives 296 (3%) additional sex differences at the Bonferroni-corrected p <.05 level, compared to raw.
“For Cohort 1Ai for example, dasen gives us 296 (3%) additional sex differences at the Bonferroni-corrected p <.05 level, compared to raw.”
Quote in the paperNot a substantial claim - Findingown data
Dasen is the best performing preprocessing method across both probe types.
“Using this approach, our data indicates that dasen is the best performing method across both probe types (Table 2).”
- Also stated in: § Systematic ranking of methods based on metrics · plain
“Dasen is shown to be the best preprocessing method across both probe types.”
Quote in the paperSubstantial claim- Rests on an assumption: The three performance metrics (DMRSE, GCOSE, Seabird) are valid indicators of preprocessing method performance.
- Also stated in: § Systematic ranking of methods based on metrics · plain
- Findingown data
The data is much improved for the Type II probes.
“The data is much improved for the Type II probes, which represent 72% of the probes on the array.”
Quote in the paperSubstantial claim- Rests on an assumption: Type II probes represent 72% of the probes on the array.
- Findingown data
β values at SNP probes cluster into three groups corresponding to heterozygous and homozygous genotypes.
“As expected the β values for these sites cluster into three groups, depending on whether the samples are heterozygous, or homozygous at each SNP (see Figure 2b).”
Quote in the paperSubstantial claim- Rests on an assumption: 65 probes on the 450K BeadChip assay highly polymorphic SNPs rather than CpG sites.
- Findingown data
More complicated segmented QN schemes seem to be counterproductive.
“but more complicated segmented QN schemes seem to be counterproductive despite embodying valuable insights about the properties of the assay (compare Tost[7] and SWAN[6] with dasen).”
Quote in the paperNot a substantial claim - Findingown data
According to the standard error measures DMRSE and GCOSE, Type II probes perform better than Type I probes across most methods.
“The first observation is that according to the standard error measures DMRSE and GCOSE, Type II probes perform better than Type I probes across most methods”
Quote in the paperNot a substantial claim - Findingown data
The dasen method performs consistently well for both the Type I and Type II probe data.
“It is evident from the quantitative ranking method (Table 2) and figures for each individual dataset (Additional files 3, 4 and 5) that the dasen method performs consistently well for both the Type I and Type II probe data.”
Quote in the paperSubstantial claim- Rests on an assumption: The three performance metrics (DMRSE, GCOSE, Seabird) are valid indicators of preprocessing method performance.
- Findingown data
Type I probes are more stable than Type II probes.
“and Type I probes are more stable than Type II probes.”
Quote in the paperNot a substantial claim - Findingown data
Both methylated and unmethylated raw intensities display a characteristic peak close to zero, which differs slightly between Type I and II probes.
“Both methylated and unmethylated raw intensities display a characteristic peak close to zero, which differs slightly between Type I and II probes (Additional file 2).”
Quote in the paperSubstantial claim - Findingown data
QN produces a slightly narrower peak in the iDMR β density than raw.
“QN produces a slightly narrower peak visible on a density plot (Figure 2a)”
Quote in the paperNot a substantial claim - Findingown data
In Cohort 1Ai, sex-difference t-tests identified 9,796 probes significant at Bonferroni-corrected p=0.05.
“In Cohort 1Ai for example, a t-test of each probe for sex differences gives 9,796 significant at the p=0.05 level (Bonferroni corrected for 485,577 tests)”
Quote in the paperSubstantial claim - Findingown data
Methods that do not use any normalization (raw and Fuks) clearly perform worse for both types than those that do.
“Methods that do not use any normalization (raw and Fuks[5]) clearly perform worse for both types than those that do.”
Quote in the paperSubstantial claim - Findingown data
Of the 9,796 significant probes in Cohort 1Ai, 8,969 correspond to X-chromosome loci.
“of which 8,969 correspond to X-chromosome loci.”
Quote in the paperSubstantial claim - Findingown data
Background adjustment does not introduce noticeable extra variance.
“Background adjustment does not introduce noticeable extra variance (compare raw with danen).”
Quote in the paperNot a substantial claim - Findingown data
Normalizing M and U is clearly better than normalizing β.
“Normalizing M and U is clearly better than normalizing β (compare naten with betaqn).”
Quote in the paperSubstantial claim - Findingreasoning
For the datasets used, the three metrics give a clear picture of the optimal quantile-based normalization procedure.
“For the datasets we use these give a clear picture about the optimal quantile-based normalization procedure.”
Quote in the paperSubstantial claim - Findingown data
The dasen method decreases the standard error across CpG sites within iDMRs.
“Figure 2a shows that the dasen method decreases the standard error across CpG sites within iDMRs, as expected (DMRSE test).”
Quote in the paperSubstantial claim - Findingown data
QN of Type I and Type II intensities separately appear to reduce variance, particularly for the iDMR measures.
“Furthermore, QN of Type I and Type II intensities separately appear to reduce variance, particularly for the iDMR measures (compare daten1 and dasen)”
Quote in the paperNot a substantial claim - Findingown data
We observed a gradient in background in some datasets.
“we observed a gradient in background in some datasets”
Quote in the paperSubstantial claim - Findingown data
In each dataset, β values for iDMR probes show a single peak at approximately 0.5.
“From each dataset we observe a distribution of β for these probes with a single peak at approximately 0.5, as expected.”
Quote in the paperSubstantial claim - Interpretationown data
Preprocessing methods may perform differently for the two assays on the 450K array.
“preprocessing methods may perform differently for the two assays”
Quote in the paperSubstantial claim - Study design and conductreasoning
Three independent metrics based on known DNA methylation patterns associated with genomic imprinting and X-chromosome inactivation, plus SNP genotyping assay performance, were used to test alternative correction and normalization schemes.
“We have taken advantage of known DNA methylation patterns associated with genomic imprinting and X-chromosome inactivation (XCI), in addition to the performance of SNP genotyping assays present on the array, to derive three independent metrics which we use to test alternative schemes of correction and normalization”
Quote in the paperSubstantial claim - Interpretationown data
Differing distributions of Type I and Type II probes may make overall rankings of differentially methylated probes inaccurate.
“differing distributions may make overall rankings of differentially methylated probes inaccurate”
Quote in the paperSubstantial claim - Interpretationown data
Careful selection of preprocessing steps can minimize variance and thus improve statistical power, especially for detecting small absolute DNA methylation changes likely associated with complex disease phenotypes.
“Careful selection of preprocessing steps can minimize variance and thus improve statistical power, especially for the detection of the small absolute DNA methylation changes likely associated with complex disease phenotypes”
Quote in the paperSubstantial claim- Rests on an assumption: Small absolute DNA methylation changes are likely associated with complex disease phenotypes.
- Interpretationown data
A combination of background adjustment and between-array quantile normalization is optimal for data processing to allow detection of differential methylation between samples.
“The results of our tests reveal that a combination of background adjustment and between-array quantile normalization is optimal for data processing, to allow the detection of differential methylation between samples.”
Quote in the paperSubstantial claim- Rests on an assumption: The array features that assay known sites of differential methylation or genotype are valid indicators of normalization performance.
- Interpretationown data
91% of the sex-difference probes recovered at this cutoff are X-chromosome loci.
“91% of the recovered differences are in X-chromosome loci.”
Quote in the paperSubstantial claim - Interpretationreasoning
Technical variation is likely more simply dealt with by adjusting M and U intensities than by adjusting derived β values.
“technical variation is likely to be more simply dealt with by adjusting these, rather than the derived β value”
Quote in the paperSubstantial claim - Premisecitation
237 probes on the 450K array lie within a conservative set of defined iDMRs.
“There are 237 probes on the array that lie within a conservative set of defined iDMRs [16]”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [16] Schulz R: Wamidex, accessed 26 Jan2012. [https://atlas.genetics.kcl.ac.uk/]
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Interpretationreasoning
The difference in performance between Type I and Type II probes is not in itself a normalization issue.
“This is not in itself a normalization issue”
Quote in the paperSubstantial claim - Study design and conductown data
The study excluded less than 1% of individual samples as clear technical failures based on atypical raw intensity distribution and poor correlation of β with other samples.
“we identified and excluded a small number (less than 1%) of individual samples which were clearly technical failures based on criteria including atypical raw intensity distribution and poor correlation of β with other samples.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Quantile normalizing M and U against each other would remove dye bias variation.
“quantile normalizing M and U against each other would have the advantage of removing dye bias variation”
Quote in the paperSubstantial claim- Rests on an assumption: Dye bias variation is a potential problem with Type II assays.
- Interpretationown data
At the Bonferroni-corrected p=0.05 cutoff, the sex-difference test recovers 80% of X-chromosome probes.
“The sex difference test with this cut-off thus recovers 80% of the X-chromosome probes.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Adjusting the background difference between Type I and II probes for both M and U intensities is the simplest way to minimize the differences between Type I and Type II probes.
“This is achieved most simply by adjusting the background difference between Type I and II probes for both M and U intensities (danes, danet, danen, daten1, daten2, dasen).”
Quote in the paperSubstantial claim - Interpretationnone given
Three metrics of data homogeneity evaluate different and substantial parts of the array.
“We propose three metrics of data homogeneity that evaluate different and substantial parts of the array.”
Quote in the paperNot a substantial claim - Premisecitation
The iDMR probes have an expected β value of 0.5 because they are uniparentally methylated in most tissues.
“and have an expected β value of 0.5 because they are uniparentally methylated in most tissues.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [16] Schulz R: Wamidex, accessed 26 Jan2012. [https://atlas.genetics.kcl.ac.uk/]
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Interpretationreasoning
Assessing DNA methylation methods by group differences verified with an independent method can be misleading because an unknown portion of those differences are artefacts.
“This has the potential to be misleading as an unknown portion of the differences are artefacts.”
Quote in the paperSubstantial claim - Interpretationnone given
Standard spiking control datasets, although influential in gene expression, would not be as suitable for DNA methylation analysis.
“Standard, specially constructed control datasets produced by spiking samples have been influential in the gene expression field [13], but would not be as suitable for analysis of DNA methylation.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [13] Irizarry Ra, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP: Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics (Oxford, England). 2003, 4 (2): 249-264. 10.1093/biostatistics/4.2.249. [http://www.ncbi.nlm.nih.gov/pubmed/12925520] source
- Interpretationown data
The study analyzed only relatively homogeneous datasets derived from brain or whole blood.
“In particular, we have only analyzed relatively homogeneous datasets derived from brain or whole blood.”
Quote in the paperSubstantial claim - Interpretationown data
The narrower iDMR β peak after QN indicates a reduction in inter-sample variance.
“indicating that we can detect a reduction in inter-sample variance.”
Quote in the paperSubstantial claim - Interpretationreasoning
Male-female differences in DNA methylation are expected at X-chromosome CpG sites influenced by XCI, with females showing at least 50% methylation and males substantially less.
“We do, however, expect male-female differences, with females showing at least 50% methylation at CpG sites on the X-chromosome that are influenced by XCI, and males substantially less.”
Quote in the paperSubstantial claim- Rests on an assumption: In females, one copy of the X chromosome is predominantly inactivated and largely methylated.
- Classified as probably not a substantial claim. 14
- May combine two claims. 6
Classified as probably not a substantial claim. May combine two claims.
PremisecitationDedeurwaerder et al. found that the difference between Type I and Type II distributions reflected a difference in performance between the two probe types and devised a custom transformation of Type II β values to accommodate it.
“Dedeurwaerder et al[5] found that this reflected a difference in performance between the two probe types and devised a custom transformation of the Type II β values to accommodate it”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [5] Dedeurwaerder S, Defrance M, Calonne E, Sotiriou C, Fuks F: Evaluation of the Infinium Methylation 450K technology. Epigenetics. 2011, 3 (6): 771-784. source
May combine two claims.
PremisecitationThe majority of CpG sites assayed are fully methylated or fully unmethylated and would change little after normalization.
“the majority of CpG sites assayed are in a fully methylated or fully unmethylated state and would change little after normalization”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [14] Roessler J, Ammerpohl O, Gutwein J, Hasemeier B, Anwar SL, Kreipe H, Lehmann U: Quantitative cross-validation and content analysis of the 450k DNA methylation array from Illumina, Inc. BMC Res Notes. 2012, 5: 210-10.1186/1756-0500-5-210. [http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3420245&tool=pmcentrez&rendertype=abstract] source
The Conclusion says that β-values are a good measure of methylation status for most CpG sites, but it does not say most sites are fully methylated or fully unmethylated. The Findings discuss agreement with pyrosequencing, not changes after normalization. Searches for extreme methylation states and normalization found no passage addressing the claim. (LLM, conf=0.72)
Experimental — accuracy not yet measured
Classified as probably not a substantial claim. May combine two claims.
PremisecitationThe DNA methylation field is technology-limited and characterized by small sample sizes, diverse techniques and variable statistical standards.
“Currently the DNA methylation field resembles gene expression studies of a decade ago: it is technology-limited and characterized by small sample sizes, diverse techniques and variable statistical standards [22].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [22] Heijmans BT, Mill J: Commentary: The seven plagues of epigenetic epidemiology. Int J Epidemiol. 2012, 41: 74-78. 10.1093/ije/dyr225. [http://www.ncbi.nlm.nih.gov/pubmed/22269254] source
May combine two claims.
PremisecitationThe minority of intermediate methylation values are susceptible to greater changes after normalization but are overlooked by a correlation test.
“the minority of intermediate methylation values would be susceptible to far greater changes after normalization but would be overlooked by a correlation test [14]”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [14] Roessler J, Ammerpohl O, Gutwein J, Hasemeier B, Anwar SL, Kreipe H, Lehmann U: Quantitative cross-validation and content analysis of the 450k DNA methylation array from Illumina, Inc. BMC Res Notes. 2012, 5: 210-10.1186/1756-0500-5-210. [http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3420245&tool=pmcentrez&rendertype=abstract] source
In Findings > Cross-validation of 450k methylation array and pyrosequencing (page not specified), the paper says correlation coefficients can be misleading because values near 0% and 100% dominate, masking poorer agreement at intermediate methylation levels. It compares array measurements with pyrosequencing; it does not say that intermediate values undergo greater changes after normalization. (LLM, conf=0.87)
Experimental — accuracy not yet measured
Probably not a substantial claim: may not pass S3 (load-bearing).
Findingown dataThe effect of background adjustment on Type I probes is clearly visible.
“While the difference in peak width is not readily visible at this scale, the effect of background adjustment on Type I probes is clearly visible.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningA higher background in Type II assays would explain the compressed distribution of β values because it inflates both M and U.
“Our insight is that a higher background in the Type II assays would explain such a difference in distribution of β values, because it inflates both M and U”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
InterpretationreasoningQuantile normalization may introduce considerable changes in parts of the distribution with few values.
“A potential weakness of QN is that in parts of the distribution with few values (and therefore relatively large interquantile differences), it may introduce considerable changes.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationQuantile normalization is a well-established technique in gene expression analysis that has been shown to perform well.
“Quantile normalization (QN) is a well established technique in gene expression analysis, where it has been shown to perform well [13].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [13] Irizarry Ra, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP: Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics (Oxford, England). 2003, 4 (2): 249-264. 10.1093/biostatistics/4.2.249. [http://www.ncbi.nlm.nih.gov/pubmed/12925520] source
Probably not a substantial claim: may not pass S3 (load-bearing).
InterpretationreasoningThe SNP control probes can provide an indication of the degree of technical variance between samples.
“Although these are not DNA methylation signals, they could be used to provide an indication of the degree of technical variance between samples.”
Quote in the paperNot a substantial claim- Rests on an assumption: The 450K BeadChip contains 65 control probes that assay highly polymorphic SNPs rather than DNA methylation.
Probably not a substantial claim: may not pass S3 (load-bearing).
Study design and conductown dataInitial quality control checks assessed concordance between reported and genotyped gender.
“Initial quality control checks were performed using functions in the methylumi package to assess concordance between reported and genotyped gender.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
Study design and conductnone givenNon-CpG SNP probes were used to confirm that multiple tissues were sourced from the same individual where expected.
“Non-CpG SNP probes on the array were also used to confirm that multiple tissues were sourced from the same individual where expected.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
InterpretationreasoningThose large changes could increase between-sample variance for individual features rather than reducing it.
“The danger is that these large changes could increase the variance across samples for individual features, rather than reducing it as desired.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenDMRSE should directly predict sensitivity to detect true differences between groups.
“and it should directly predict sensitivity to detect true differences between groups.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenDMRSE is reasonably independent of the size of the dataset.
“This measure was chosen because it is reasonably independent of the size of the dataset”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
InterpretationreasoningEffects on β values may be complicated by interaction between M and U.
“where effects may be complicated by their interaction.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenThe signal from the 65 SNP control probes is expected to cluster into three distinct groups representing the heterozygous and two homozygous genotypes.
“The signal from these probes is expected to cluster into three distinct groups (representing the heterozygous and two homozygous groups).”
Quote in the paperNot a substantial claim- Rests on an assumption: The 450K BeadChip contains 65 control probes that assay highly polymorphic SNPs rather than DNA methylation.
Probably not a substantial claim: may not pass S3 (load-bearing).
Interpretationnone givenThe study only tested normalization methods based on quantile normalization.
“Finally we have only tested methods based on QN.”
Quote in the paperNot a substantial claim
The paper relies on this without support.
Small absolute DNA methylation changes are likely associated with complex disease phenotypes.
“especially for the detection of the small absolute DNA methylation changes likely associated with complex disease phenotypes”
The paper relies on this without support.
Attempts to correct background by subtracting estimates derived from control probes introduce another source of variance.
“attempts to correct them, for example by subtraction of estimates derived from control probes, are undesirable because they introduce another source of variance”
The paper relies on this without support.
The suitability of quantile normalization for DNA methylation data has been assumed from gene expression experience, but there has been little systematic testing.
“The suitability of QN for DNA methylation data has been assumed based on experience in the analysis of gene expression data, but there has been little systematic testing.”
The paper relies on this without support.
The 450K BeadChip contains 65 control probes that assay highly polymorphic SNPs rather than DNA methylation.
“The 450K BeadChip also features 65 control probes which assay highly-polymorphic single nucleotide polymorphisms (SNPs) rather than DNA methylation.”
The paper relies on this without support.
GenomeStudio default settings perform no normalization.
“the manufacturer’s GenomeStudio software calculates βs from raw intensities and under default settings performs no normalization”
The paper relies on this without support.
In females, one copy of the X chromosome is predominantly inactivated and largely methylated.
“In females, one copy of the X-chromosome is predominantly inactivated and largely methylated.”
The paper relies on this without support.
65 probes on the 450K BeadChip assay highly polymorphic SNPs rather than CpG sites.
“There are 65 probes on the 450K BeadChip which assay highly polymorphic SNPs rather than CpG sites.”
The paper relies on this without support.
QN has the potential to impose large changes in parts of a distribution with few members.
“the known potential of QN to impose large changes in those parts of a distribution with few members”
The paper relies on this without support.
M and U distributions differ.
“M and U distributions differ”
The paper relies on this without support.
Dye bias variation is a potential problem with Type II assays.
“which is a potential problem with Type II assays”
The paper relies on this without support.
The inactive X-chromosome in females is hypermethylated.
“The inactive X-chromosome in females is hypermethylated”
The paper relies on this without support.
The 11,232 X-chromosome features annotated for the array show a distinctive β distribution centred on approximately 0.5, compared with the biphasic distribution of autosomal probes.
“the 11,232 X-chromosome features annotated for the array show a distinctive β value distribution, once again centred on approximately 0.5, compared to the biphasic distribution of autosomal probes.”
The paper relies on this without support.
The three performance metrics (DMRSE, GCOSE, Seabird) are valid indicators of preprocessing method performance.
“The test metrics were subsequently used to quantitatively determine the best performing preprocessing method”
The paper relies on this without support.
Type II probes represent 72% of the probes on the array.
“which represent 72% of the probes on the array”
The paper relies on this without support.
The array features that assay known sites of differential methylation or genotype are valid indicators of normalization performance.
“We then evaluated each method using features on the array that assay known sites of differential methylation or genotype.”
The paper relies on this without support.
Quantile normalization is strongly favoured by consensus in the field.
“Although there is a strong consensus in favour of this method, it is possible that a different approach based, for example, on scaling, could possibly work better.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
SWAN normalisation removed technical differences between MINFI package Infinium I and Infinium II probes.
“SWAN normalisation removed technical differences between MINFI package Infinium I and Infinium II probes (Maksimovic, Gordon & Oshlack, 2012;Pidsley et al., 2013).”
From DNA methylation of NR3C1 and FKBP5 is associated with posttraumatic stress disorder, posttraumatic growth, and resilience. (2020) · cites this paper as Pidsley et al., 2013
In Results and discussion, “Qualitative observations on preprocessing and normalization” (page unavailable), the paper describes SWAN’s segmented quantile-normalization scheme as counterproductive relative to dasen. Its “Preprocessing and normalization method design” section attributes reduction of Type I/II background differences to other methods, not SWAN. The paper therefore does not report that SWAN removed those technical differences. (LLM, conf=0.76)
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
One cohort is said to supply seven datasets but eight are listed
MinorThe Methods section labels the cohort as the source of seven datasets, then names eight: 1Ai, 1Aii, 1B, 1C, 1D, 1E, 1BC and 1AD.
“Seven of the datasets were obtained from one cohort of post-mortem brain samples”
“cerebellum (1Ai) (n=91), rescan of cerebellum data (1Aii) (n=36), frontal cortex (1B) (n=89), entorhinal cortex (1C) (n=93), superior temporal gyrus (1D) (n=94), whole blood (1E) (n=95), frontal cortex and entorhinal cortex (1BC) (n=46), cerebellum and superior temporal gyrus (1AD) (n=47).”
Cites
Bioconductor: open software development for computational biology and bioinformatics
Gentleman, R., Carey, V., Bates, D., Bolstad, B., Dettling, M., Dudoit, S., Ellis, B., Gautier, L., Ge, Y., Gentry, J., et al. (2004) · Genome Biology · doi:10.1186/gb-2004-5-10-r80 · Cited by 13,133 · cited 1× · see citing works
“12 [24].”
Exploration, normalization, and summaries of high density oligonucleotide array probe level data
Irizarry, R. A. (2003) · Biostatistics · doi:10.1093/biostatistics/4.2.249 · Cited by 10,603 · cited 2× · see citing works
“Standard, specially constructed control datasets produced by spiking samples have been influential in the gene expression field [13], but would not be as suitable for analysis of DNA methylation.”
BackgroundMethodlimma: Linear Models for Microarray Data
Smyth, G. K. (2005) · Statistics for Biology and Health · doi:10.1007/0-387-29362-0_23 · Cited by 5,796 · see citing works
Bioinformatics and Computational Biology Solutions Using R and Bioconductor
Gentleman, R. C., Carey, V. J., Huber, W. K., Irizarry, R. A., & Dudoit, S. (2005) · Statistics in the health sciences · doi:10.1007/0-387-29362-0 · Cited by 3,470 · see citing works
ROCR: visualizing classifier performance in R
Sing, T., Sander, O., Beerenwinkel, N., & Lengauer, T. (2005) · Computer applications in the biosciences · doi:10.1093/bioinformatics/bti623 · Cited by 3,205 · see citing works
Du, P., Zhang, X., Huang, C.-C., Jafari, N., Kibbe, W. A., Hou, L., & Lin, S. M. (2010) · BMC Bioinformatics · doi:10.1186/1471-2105-11-587 · Cited by 2,136 · cited 1× · see citing works
“An alternative index not bounded by 0 and 1 is M = log2((M + α)/(U + α)), which is essentially equivalent to a logit transformation of β [4].”
BackgroundHigh density DNA methylation array with single CpG site resolution
Bibikova, M., Barnes, B., Tsan, C., Ho, V., Klotzle, B., Le, J. M., Delano, D., Zhang, L., Schroth, G. P., Gunderson, K. L., et al. (2011) · Genomics · doi:10.1016/j.ygeno.2011.07.007 · Cited by 1,727 · cited 9× · see citing works
“The 450K BeadChip [2,3] achieves increased coverage by utilising two different probe types on each array: Infinium I (n=135,501) and Infinium II (n=350,076) probes.”
BackgroundMethodInfluentialTeschendorff, A. E., Marabita, F., Lechner, M., Bartlett, T. E., Tegner, J. N., Gómez-Cabrero, D., & Beck, S. (2012) · Bioinformatics · doi:10.1093/bioinformatics/bts680 · Cited by 1,652 · see citing works
Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome
Sandoval, J., Heyn, H., Moran, S., Serra-Musach, J., Pujana, M. A., Bibikova, M., & Esteller, M. (2011) · Epigenetics · doi:10.4161/epi.6.6.16196 · Cited by 990 · cited 9× · see citing works
“The 450K BeadChip [2,3] achieves increased coverage by utilising two different probe types on each array: Infinium I (n=135,501) and Infinium II (n=350,076) probes.”
BackgroundMethodInfluentialSWAN: Subset-quantile Within Array Normalization for Illumina Infinium HumanMethylation450 BeadChips
Maksimovic, J., Gordon, L., & Oshlack, A. (2012) · Genome Biology · doi:10.1186/gb-2012-13-6-r44 · Cited by 826 · see citing works
Genome-Wide Dna Methylation Profiling Using Infinium ® Assay
Bibikova, M. V., Le, J. M., Barnes, B., Saedinia-Melnyk, S., Zhou, L., Shen, R., & Gunderson, K. L. (2009) · Epigenomics · doi:10.2217/epi.09.14 · Cited by 592 · cited 1× · see citing works
“The initial HumanMethylation27 (27K) BeadChip interrogated 27,578 CpG sites associated with 14,495 protein-coding gene promoters [1].”
BackgroundEvaluation of the Infinium Methylation 450K technology
Dedeurwaerder, S., Defrance, M., Calonne, E., Denis, H., Sotiriou, C., & Fuks, F. (2011) · Epigenomics · doi:10.2217/epi.11.105 · Cited by 569 · cited 7× · see citing works
“This is not in itself a normalization issue, but it may be advantageous to minimize the differences so that the ranking of potential differentially methylated loci is more accurate [5].”
BackgroundMethodInfluentialTouleimat, N., & Tost, J. (2012) · Epigenomics · doi:10.2217/epi.12.21 · Cited by 509 · cited 2× · see citing works
“A number of research groups have developed preprocessing methods to accommodate the differences in signal between the two probe types [5-7].”
BackgroundDempster, E. L., Pidsley, R., Schalkwyk, L. C., Owens, S., Georgiades, A., Kane, F., Kalidindi, S., Picchioni, M., Kravariti, E., Toulopoulou, T., et al. (2011) · Human Molecular Genetics · doi:10.1093/hmg/ddr416 · Cited by 460 · cited 1× · see citing works
“However, for the investigation of subtle differences, such as those seen in common complex disorders such as schizophrenia [9,10] and diabetes [11],”
BackgroundSelf-citationRakyan, V. K., Beyan, H., Down, T. A., Hawa, M. I., Maslau, S., Aden, D., Daunay, A., Busato, F., Mein, C. A., Manfras, B. J., et al. (2011) · PLoS Genetics · doi:10.1371/journal.pgen.1002300 · Cited by 343 · cited 1× · see citing works
“However, for the investigation of subtle differences, such as those seen in common complex disorders such as schizophrenia [9,10] and diabetes [11],”
BackgroundIMA: an R package for high-throughput analysis of Illumina's 450K Infinium methylation data
Wang, D., Yan, L., Hu, Q., Sucheston, L. E., Higgins, M. J., Ambrosone, C. B., Johnson, C. S., Smiraglia, D. J., & Liu, S. (2012) · Bioinformatics · doi:10.1093/bioinformatics/bts013 · Cited by 309 · cited 2× · see citing works
“The functions in the package are all generic and methods are provided for the objects produced by the packages methylumi [18], minfi [19], and IMA [20], as well as wrappers for the methods of Dedeurwaerder et al [5] A number of packages from CRAN and bioConductor were used: quantile normalization using limma [25]; dat…”
MethodCommentary: The seven plagues of epigenetic epidemiology
Heijmans, B. T., & Mill, J. (2012) · International Journal of Epidemiology · doi:10.1093/ije/dyr225 · Cited by 208 · cited 1× · see citing works
“Currently the DNA methylation field resembles gene expression studies of a decade ago: it is technologylimited and characterized by small sample sizes, diverse techniques and variable statistical standards [22].”
BackgroundSelf-citationRoessler, J., Ammerpohl, O., Gutwein, J., Hasemeier, B., Anwar, S. L., Kreipe, H., & Lehmann, U. (2012) · BMC Research Notes · doi:10.1186/1756-0500-5-210 · Cited by 112 · cited 1× · see citing works
“the majority of CpG sites assayed are in a fully methylated or fully unmethylated state and would change little after normalization, whilst the minority of intermediate methylation values would be susceptible to far greater changes after normalization but would be overlooked by a correlation test [14].”
BackgroundBatch effect correction for genome-wide methylation data with Illumina Infinium platform
Sun, Z., Chai, H. S., Wu, Y., White, W. M., Donkena, K. V., Klein, C. J., Garovic, V. D., Therneau, T. M., & Kocher, J.-P. A. (2011) · BMC Medical Genomics · doi:10.1186/1755-8794-4-84 · Cited by 108 · cited 3× · see citing works
“A pragmatic approach to the limitations of simple ratiobased methods to calculate DNA methylation values, common in the literature [12], has been to quantile normalize β scores.”
BackgroundMethodInfluentialChen, Y.-A., Choufani, S., Ferreira, J. C., Grafodatskaya, D., Butcher, D. T., & Weksberg, R. (2011) · Genomics · doi:10.1016/j.ygeno.2010.12.004 · Cited by 90 · cited 1× · see citing works
“True, autosomal sex-specifically methylated loci are very small in number and potentially represent autosomal probes mapping to sequences on the sex-chromosome [17].”
BackgroundDNA Methylation Signatures of Peripheral Leukocytes in Schizophrenia
Kinoshita, M., Numata, S., Tajima, A., Shimodera, S., Ono, S., Imamura, A., Iga, J., Watanabe, S., Kikuchi, K., Kubo, H., et al. (2012) · NeuroMolecular Medicine · doi:10.1007/s12017-012-8198-6 · Cited by 78 · cited 1× · see citing works
“However, for the investigation of subtle differences, such as those seen in common complex disorders such as schizophrenia [9,10] and diabetes [11],”
BackgroundIdentification of epigenetically regulated genes that predict patient outcome in neuroblastoma
Carén, H., Djos, A., Nethander, M., Sjöberg, R.-M., Kogner, P., Enström, C., Nilsson, S., & Martinsson, T. (2011) · BMC Cancer · doi:10.1186/1471-2407-11-66 · Cited by 74 · cited 1× · see citing works
“The β values are indeed fairly stable and the basic analysis method works well, particularly for the detection of large differences, for example between tissues or tumornormal pairs [8].”
BackgroundWAMIDEX: A web atlas of murine genomic imprinting and differential expression
Schulz, R., Woodfine, K., Menheniott, T. R., Bourc'his, D., Bestor, T., & Oakey, R. J. (2008) · Epigenetics · doi:10.4161/epi.3.2.5900 · Cited by 56 · cited 1× · see citing works
“Stable iDMRs have been characterized for 25 human imprinting regions [15,16] and where array features overlap these, we would expect monoallelic (therefore 50% (β=0.”
Background
Flagged first, then by citations
1 of the 24 references couldn’t be matched to a known paper.
Research works printed but not linked: 1
- Davis S, Du P, Bilke S, Triche JrT, Bootwalla M: Methylumi: Handle Illumina Methylation Data. 2012, R package version 2.2.0, .
Cited by
Sorted by citations
Aryee, M. J., Jaffe, A. E., Bravo, H. C., Ladd‐Acosta, C., Feinberg, A. P., Hansen, K. D., & Irizarry, R. A. (2014) · Bioinformatics · doi:10.1093/bioinformatics/btu049 · Cited by 4,423 · see citing works
Pidsley, R., Zotenko, E., Peters, T. J., Lawrence, M. G., Risbridger, G. P., Molloy, P., Van Djik, S., Muhlhausler, B., Stirzaker, C., & Clark, S. J. (2016) · Genome biology · doi:10.1186/s13059-016-1066-1 · Cited by 1,480 · see citing works
Self-citationComprehensive and Integrated Genomic Characterization of Adult Soft Tissue Sarcomas
Abeshouse, A., Adebamowo, C., Adebamowo, S. N., Akbani, R., Akeredolu, T., Ally, A., Anderson, M. L., Anur, P., Appelbaum, E. L., Armenia, J., et al. (2017) · Cell · doi:10.1016/j.cell.2017.10.014 · Cited by 1,069 · see citing works
ChAMP: 450k Chip Analysis Methylation Pipeline
Morris, T. J., Butcher, L. M., Feber, A., Teschendorff, A. E., Chakravarthy, A., Wojdacz, T. K., & Beck, S. (2013) · Bioinformatics · doi:10.1093/bioinformatics/btt684 · Cited by 1,004 · see citing works
De novo identification of differentially methylated regions in the human genome
Peters, T. J., Buckley, M. J., Statham, A. L., Pidsley, R., Samaras, K., Lord, R. V., Clark, S. J., & Molloy, P. L. (2015) · Epigenetics & Chromatin · doi:10.1186/1756-8935-8-6 · Cited by 956 · see citing works
“For each comparison, samples were normalised separately using the dasen method from the R package wateRmelon [23].”
MethodSelf-citationWhole-Genome and Epigenomic Landscapes of Etiologically Distinct Subtypes of Cholangiocarcinoma
Jusakul, A., Cutcutache, I., Yong, C. H., Lim, J. Q., Huang, M. N., Padmanabhan, N., Nellore, V., Kongpetch, S., Ng, A. W. T., Ng, L. M., et al. (2017) · Cancer Discovery · doi:10.1158/2159-8290.cd-17-0368 · Cited by 972 · see citing works
“Data was preprocessed using the “minfi” (73) and “wateRmelon” (74) R packages.”
DNA methylation in newborns and maternal smoking in pregnancy: genome-wide consortium meta-analysis
Joubert, B. R., Felix, J. F., Yousefi, P., Bakulski, K. M., Just, A. C., Breton, C., Reese, S. E., Markunas, C. A., Richmond, R. C., Xu, C.-J., et al. (2016) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2016.02.019 · Cited by 886 · see citing works
Epigenetic Signatures of Cigarette Smoking
Joehanes, R., Just, A. C., Marioni, R. E., Pilling, L. C., Reynolds, L. M., Mandaviya, P. R., Guan, W., Xu, T., Elks, C. E., Aslibekyan, S. W., et al. (2016) · Circulation Cardiovascular Genetics · doi:10.1161/circgenetics.116.001506 · Cited by 915 · see citing works
“Despite the decline in the prevalence of smoking in many countries, it remains the leading preventable cause of death in the world, accounting for ≈6 million deaths each year.(2)”
BackgroundFunctional normalization of 450k methylation array data improves replication in large cancer studies
Fortin, J.-P., Labbe, A., Lemire, M., Zanke, B. W., Hudson, T. J., Fertig, E. J., Greenwood, C. M. T., & Hansen, K. D. (2014) · Genome biology · doi:10.1186/s13059-014-0503-2 · Cited by 893 · see citing works
“, 2012] and dasen [Pidsley et al., 2013].”
BackgroundInfluentialAn epigenetic clock analysis of race/ethnicity, sex, and coronary heart disease
Horvath, S., Gurven, M., Levine, M. E., Trumble, B. C., Kaplan, H. S., Allayee, H., Ritz, B. R., Chen, B. H., Lu, A. T.-H., Rickabaugh, T. M., et al. (2016) · Genome biology · doi:10.1186/s13059-016-1030-0 · Cited by 729 · see citing works
“The R package wateRmelon [70] was used for β-value normalization and quality control.”
MethodGenomic hallmarks of localized, non-indolent prostate cancer
Fraser, M. E., Sabelnykova, V. Y., Yamaguchi, T. N., Heisler, L. E., Livingstone, J. M., Huang, V., Shiah, Y.-J., Yousif, F., Lin, X., Masella, A. P., et al. (2017) · Nature · doi:10.1038/nature20788 · Cited by 589 · see citing works
Comprehensive analysis of DNA methylation data with RnBeads
Assenov, Y., Müller, F., Lutsik, P., Walter, J., Lengauer, T., & Bock, C. (2014) · Nature Methods · doi:10.1038/nmeth.3115 · Cited by 587 · see citing works
Systematic identification of genetic influences on methylation across the human life course
Gaunt, T. R., Shihab, H. A., Hemani, G., Min, J. L., Woodward, G., Lyttleton, O., Zheng, J., Duggirala, A., McArdle, W. L., Ho, K., et al. (2016) · Genome Biology · doi:10.1186/s13059-016-0926-z · Cited by 598 · see citing works
“wateRmelon package [36] using the Touleimat and Tost [37] algorithm to reduce the non-biological differences be-”
MethodMethylomic profiling implicates cortical deregulation of ANK1 in Alzheimer's disease
Lunnon, K., Smith, R., Hannon, E., De Jager, P. L., Srivastava, G., Volta, M., Troakes, C., Al-Sarraj, S., Burrage, J., Macdonald, R., et al. (2014) · Nature Neuroscience · doi:10.1038/nn.3782 · Cited by 574 · see citing works
Self-citationHiggins‐Chen, A., Thrush, K., Wang, Y., Minteer, C. J., Kuo, P., Wang, M., Niimi, P., Sturm, G., Lin, J., Moore, A. Z., et al. (2022) · Nature Aging · doi:10.1038/s43587-022-00248-2 · Cited by 498 · see citing works
“We examined 34 individuals with two scans each in cerebellum 36 .”
BackgroundMethodDisease variants alter transcription factor levels and methylation of their binding sites
Bonder, M. J., Luijk, R., Zhernakova, D. V., Moed, M., Deelen, P., Vermaat, M., van Iterson, M., van Dijk, F., van Galen, M., Bot, J. J., et al. (2016) · Nature Genetics · doi:10.1038/ng.3721 · Cited by 462 · see citing works
“Following quality control, we performed background correction and probe type normalization as implemented in DASEN 45 .”
Hannon, E., Lunnon, K., Schalkwyk, L., & Mill, J. (2015) · Epigenetics · doi:10.1080/15592294.2015.1100786 · Cited by 455 · see citing works
“Data was subsequently normalized in the R package wateRmelon using the dasen function as previously described.(35) Prior to data analysis, we removed the 65 nonCpG SNP probes and probes characterized by either non-specific binding (n D 43,233) or containing common (minor allele frequency > 5%) SNPs within 10 bp of the…”
MethodSelf-citationENmix: a novel background correction method for Illumina HumanMethylation450 BeadChip
Xu, Z., Niu, L., Li, L., & Taylor, J. A. (2015) · Nucleic Acids Research · doi:10.1093/nar/gkv907 · Cited by 442 · see citing works
“proposed background adjustment instead of background correction: adding the offset between Infinium I and II probe intensity values back to Infinium I probe intensities (5).”
BackgroundMethodInfluentialDNA methylation-based estimator of telomere length
Lu, A. T., Seeboth, A., Tsai, P.-C., Sun, D., Quach, A., Reiner, A. P., Kooperberg, C., Ferrucci, L., Hou, L., Baccarelli, A. A., et al. (2019) · Aging · doi:10.18632/aging.102173 · Cited by 440 · see citing works
Demerath, E. W., Guan, W., Grove, M. L., Aslibekyan, S., Mendelson, M., Zhou, Y.-H., Hedman, Å. K., Sandling, J. K., Li, L.-A., Irvin, M. R., et al. (2015) · Human Molecular Genetics · doi:10.1093/hmg/ddv161 · Cited by 355 · see citing works
“Further, the results show that despite differences in blood cell type (GOLDN used CD4+ T cells, ARIC and FHS used whole blood), normalization approach [GOLDN used COMBAT followed by polynomial regression normalization (43), FHS used wateRmelon and Beta MIxture Quantile dilation (BMIQ) (44) and ARIC used BMIQ] and some…”
MethodInfluentialLehne, B., Drong, A. W., Loh, M., Zhang, W., Scott, W. R., Tan, S.-T., Afzal, U., Scott, J., Jarvelin, M.-R., Elliott, P., et al. (2015) · Genome biology · doi:10.1186/s13059-015-0600-x · Cited by 380 · see citing works
“On the basis of these results, which are in agreement with previous findings [23,24], we performed quantile normalisation of intensity values for all samples in this study.”
ResultRnBeads 2.0: comprehensive analysis of DNA methylation data
Müller, F., Scherer, M., Assenov, Y., Lutsik, P., Walter, J., Lengauer, T., & Bock, C. (2019) · Genome biology · doi:10.1186/s13059-019-1664-9 · Cited by 369 · see citing works
“html), watermelon [38], and ChAMP [39], and with the methyKit [40] package for analyzing bisulfite sequencing data.”
MethodSmith, A. K., Kilaru, V., Klengel, T., Mercer, K. B., Bradley, B., Conneely, K. N., Ressler, K. J., & Binder, E. B. (2015) · American Journal of Medical Genetics Part B: Neuropsychiatric Genetics · doi:10.1002/ajmg.b.32278 · Cited by 355 · see citing works
“Next, we sought to compare DNA methylation in each tissue to that of four brain regions (cerebellum, frontal cortex, entorhinal cortex, and superior temporal gyrus) from a publically available dataset (GEO GSE43414) that has been previously described [Pidsley et al., 2013].”
MethodAir pollution and DNA methylation: effects of exposure in humans
Rider, C. F., & Carlsten, C. (2019) · Clinical Epigenetics · doi:10.1186/s13148-019-0713-2 · Cited by 368 · see citing works
Hannon, E., Dempster, E., Viana, J., Burrage, J., Smith, A. R., Macdonald, R., St Clair, D., Mustard, C., Breen, G., Therman, S., et al. (2016) · Genome Biology · doi:10.1186/s13059-016-1041-x · Cited by 349 · see citing works
“Normalization of the DNA methylation data was performed used the dasen() function in the wateRmelon package [52].”
Self-citationMethylation QTLs in the developing brain and their enrichment in schizophrenia risk loci
Hannon, E. J., Spiers, H., Viana, J. F., Pidsley, R., Burrage, J., Murphy, T. M., Troakes, C., Turecki, G., O’Donovan, M. C., Schalkwyk, L. C., et al. (2015) · Nature Neuroscience · doi:10.1038/nn.4182 · Cited by 336 · see citing works
Self-citationGuidance for DNA methylation studies: statistical insights from the Illumina EPIC array
Mansell, G., Gorrie-Stone, T. J., Bao, Y., Kumari, M., Schalkwyk, L. S., Mill, J., & Hannon, E. (2019) · BMC Genomics · doi:10.1186/s12864-019-5761-7 · Cited by 340 · see citing works
“Raw signal intensity data were processed from idat files through a standard pipeline using the bigmelon [48] and wateRmelon [49] packages in R.”
MethodSelf-citationLehne, B. C., Drong, A., Loh, M. C. S., Zhang, W., Scott, W. R., Tan, S.-T., Afzal, U., Schulz, R., Scott, J., Järvelin, M., et al. (2016) · Genome biology · doi:10.1186/s13059-016-0934-z · Cited by 284 · see citing works
Methylomic trajectories across human fetal brain development
Spiers, H., Hannon, E., Schalkwyk, L. C., Smith, R., Wong, C. C. Y., O’Donovan, M. C., Bray, N. J., & Mill, J. (2015) · Genome Research · doi:10.1101/gr.180273.114 · Cited by 287 · see citing works
“DNA methylation was quantified using the Illumina Infinium HumanMethylation450 BeadChip, with preprocessing, normalization, and stringentqualitycontrol undertakenaspreviouslydescribed (Pidsley et al. 2013).”
BackgroundMethodSelf-citationData Resource Profile: Accessible Resource for Integrated Epigenomic Studies (ARIES)
Relton, C. L., Gaunt, T. R., McArdle, W. L., Ho, K. M., Duggirala, A., Shihab, H. A., Woodward, G., Lyttleton, O., Evans, D. M., Reik, W., et al. (2015) · International Journal of Epidemiology · doi:10.1093/ije/dyv072 · Cited by 287 · see citing works
Walton, E., Hass, J., Liu, J., Roffman, J. L., Bernardoni, F., Roessner, V., Kirsch, M., Schackert, G., Calhoun, V., & Ehrlich, S. (2016) · Schizophrenia Bulletin · doi:10.1093/schbul/sbv074 · Cited by 286 · see citing works
Breton, C. V., Marsit, C. J., Faustman, E. M., Nadeau, K. C., Goodrich, J. M., Dolinoy, D. C., Herbstman, J. B., Holland, N., LaSalle, J. M., Schmidt, R. J., et al. (2017) · Environmental Health Perspectives · doi:10.1289/ehp595 · Cited by 261 · see citing works
“…have been proposed for removal of batch effects (Fortin et al. 2014; Heiss and Brenner 2015; Leek and Storey 2007, 2008; Maksimovic et al. 2015; Pidsley et al. 2013; Teschendorff et al. 2011), and ComBat 516 volume 125 | number 4 | April 2017 • Environmental Health Perspectives (Johnson et al.…”
BackgroundTraumatic stress and accelerated DNA methylation age: a meta-analysis
Wolf, E. J., Maniates, H., Nugent, N., Maihofer, A. X., Armstrong, D., Ratanatharathorn, A., Ashley-Koch, A. E., Garrett, M., Kimbrel, N. A., Lori, A., et al. (2018) · Psychoneuroendocrinology · doi:10.1016/j.psyneuen.2017.12.007 · Cited by 245 · see citing works
The Proteogenomic Landscape of Curable Prostate Cancer
Sinha, A., Huang, V., Livingstone, J. M., Wang, J., Fox, N. S., Kurganovs, N. J., Ignatchenko, V., Fritsch, K., Donmez, N., Heisler, L. E., et al. (2019) · Cancer Cell · doi:10.1016/j.ccell.2019.02.005 · Cited by 250 · see citing works
Huang, T., Kim, C. K., Alvarez, A. A., Pangeni, R. P., Wan, X., Song, X., Shi, T., Yang, Y., Sastry, N. G., Horbinski, C., et al. (2017) · Cancer Cell · doi:10.1016/j.ccell.2017.11.005 · Cited by 251 · see citing works
“In order to preserve the biological variations on methylation profiles among different subtypes, no further normalization was performed (Laird, 2010; Pidsley et al., 2013).”
Genetic impacts on DNA methylation: research findings and future perspectives
Villicaña, S., & Bell, J. T. (2021) · Genome biology · doi:10.1186/s13059-021-02347-6 · Cited by 247 · see citing works
An Organoid Biobank of Neuroendocrine Neoplasms Enables Genotype-Phenotype Mapping
Kawasaki, K., Toshimitsu, K., Matano, M., Fujita, M., Fujii, M., Togasaki, K., Ebisudani, T., Shimokawa, M., Takano, A., Takahashi, S., et al. (2020) · Cell · doi:10.1016/j.cell.2020.10.023 · Cited by 238 · see citing works
Characterization of whole-genome autosomal differences of DNA methylation between men and women
Singmann, P., Shem‐Tov, D., Wahl, S., Grallert, H., Fiorito, G., Shin, S.-Y., Schramm, K., Wolf, P., Kunze, S., Baran, Y., et al. (2015) · Epigenetics & Chromatin · doi:10.1186/s13072-015-0035-3 · Cited by 224 · see citing works
“3 [66]).”
Germline BRCA2 mutations drive prostate cancers with distinct evolutionary trajectories
Taylor, R. A., Fraser, M. E., Livingstone, J. M., Espiritu, S. M. G., Thorne, H., Huang, V., Lo, W. W., Shiah, Y.-J., Yamaguchi, T. N., Sliwinski, A., et al. (2017) · Nature Communications · doi:10.1038/ncomms13671 · Cited by 234 · see citing works
Marzi, S. J., Leung, S. K., Ribarska, T., Hannon, E., Smith, A. R., Pishva, E., Poschmann, J., Moore, K., Troakes, C., Al-Sarraj, S., et al. (2018) · Nature Neuroscience · doi:10.1038/s41593-018-0253-7 · Cited by 232 · see citing works
Self-citationPlasma proteomic biomarker signature of age predicts health and life span
Tanaka, T., Basisty, N., Fantoni, G., Candia, J., Moore, A. Z., Biancotto, A., Schilling, B., Bandinelli, S., & Ferrucci, L. (2020) · eLife · doi:10.7554/elife.61073 · Cited by 213 · see citing works
“Quality filtering (bead count, detection rate) and normalization was performed using the DASEN method in the watermelon R package (Pidsley et al., 2013).”
MethodGruzieva, O., Xu, C.-J., Breton, C. V., Annesi-Maesano, I., Antó, J. M., Auffray, C., Ballereau, S., Bellander, T., Bousquet, J., Bustamante, M., et al. (2017) · Environmental Health Perspectives · doi:10.1289/ehp36 · Cited by 227 · see citing works
A comprehensive overview of Infinium HumanMethylation450 data processing
Dedeurwaerder, S., Defrance, M., Bizet, M., Calonne, E., Bontempi, G., & Fuks, F. (2013) · Briefings in Bioinformatics · doi:10.1093/bib/bbt054 · Cited by 226 · see citing works
“For example, in the wateRmelon package, the ‘nasen’ method consists in four quantile normalizations between samples, as the data are separated according to probe type (Infinium I or Infinium II) and color channel [16].”
MethodDNA Methylation Analysis Identifies Loci for Blood Pressure Regulation
Richard, M. A., Huan, T., Ligthart, S., Gondalia, R., Jhun, M. A., Brody, J. A., Irvin, M. R., Marioni, R. E., Shen, J., Tsai, P.-C., et al. (2017) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2017.09.028 · Cited by 225 · see citing works
The Role of DNA Methylation in Cardiovascular Risk and Disease
Zhong, J., Agha, G., & Baccarelli, A. (2016) · Circulation Research · doi:10.1161/circresaha.115.305206 · Cited by 223 · see citing works
“Additional technical issues that are unique to the current technology for EWAS analyses include background fluorescence signal (background noise) that can distort measurement of methylation intensities,(113) probe-design-type bias because of mixing 2 different technological assays (type I and II) in the Infinium 450K,…”
MethodHuan, T., Joehanes, R., Song, C., Peng, F., Guo, Y., Mendelson, M., Yao, C., Liu, C., Ma, J., Richard, M. A., et al. (2019) · Nature Communications · doi:10.1038/s41467-019-12228-z · Cited by 216 · see citing works
Irvin, M. R., Zhi, D., Joehanes, R., Mendelson, M., Aslibekyan, S. W., Claas, S. A., Thibeault, K. S., Patel, N. M., Day, K. R., Jones, L. W., et al. (2014) · Circulation · doi:10.1161/circulationaha.114.009158 · Cited by 217 · see citing works
“Methylation data were normalized within laboratory batches by using a DASEN method from the wateRmelon package.(27) Robust multichip average method was used to normalize the gene expression values with quality-control measures as previously reported.”
MethodDNA methylation methods: Global DNA methylation and methylomic analyses
Li, S., & Tollefsbol, T. O. (2020) · Methods · doi:10.1016/j.ymeth.2020.10.002 · Cited by 219 · see citing works
Reynolds, L. M., Taylor, J. R., Ding, J., Lohman, K., Johnson, C. W., Siscovick, D. S., Burke, G. L., Post, W. S., Shea, S., Jacobs, D. R., et al. (2014) · Nature Communications · doi:10.1038/ncomms6366 · Cited by 201 · see citing works
“…passed the following filters: “detected” methylation levels 90% of MESA samples using a detection p-value cut-off of 0.05, 65 control probes which assay highly-polymorphic single nucleotide polymorphisms (SNPs) rather than DNA methylation 41 , or overlap with a repetitive element or region.”
InfluentialNeedham, B. L., Smith, J. A., Zhao, W., Wang, X., Mukherjee, B., Kardia, S. L. R., Shively, C. A., Seeman, T. E., Liu, Y., & Roux, A. V. D. (2015) · Epigenetics · doi:10.1080/15592294.2015.1085139 · Cited by 210 · see citing works
“The candidate-gene approach limited the number of tests performed but may have underestimated the extent to which life course SES is associated with DNAm.”
MethodInfluentialPrediction of gestational age based on genome-wide differentially methylated regions
Bohlin, J., Håberg, S. E., Magnus, P. M., Reese, S. E., Gjessing, H. K., Magnus, M. C., Parr, C. L., Page, C. M., London, S. J., & Nystad, W. (2016) · Genome biology · doi:10.1186/s13059-016-1063-4 · Cited by 209 · see citing works
Finer, S., Mathews, C. S., Lowe, R., Smart, M., Hillman, S. L., Foo, L. F., Sinha, A., Williams, D. J., Rakyan, V. K., & Hitman, G. A. (2015) · Human Molecular Genetics · doi:10.1093/hmg/ddv013 · Cited by 208 · see citing works
“Our sampling and analysis pipeline are consistent with consensus approaches (28,29) and we have established that there is no significant difference in cellular composition of cord blood between cases and controls.”
MethodHannon, E., Knox, O., Sugden, K., Burrage, J., Wong, C., Belsky, D. W., Corcoran, D., Arseneault, L., Moffitt, T., Caspi, A., et al. (2018) · PLoS Genetics · doi:10.1371/journal.pgen.1007544 · Cited by 187 · see citing works
Self-citationRecalibrating the epigenetic clock: implications for assessing biological age in the human cortex
Shireby, G. L., Davies, J. P., Francis, P. T., Burrage, J., Walker, E., Neilson, G., Dahir, A., Thomas, A. J., Love, S., Smith, R. G., et al. (2020) · Brain · doi:10.1093/brain/awaa334 · Cited by 193 · see citing works
“The subsequent normalization of the DNA methylation data was performed using the dasen() function in either wateRmelon or bigmelon (Pidsley et al. , 2013; Gorrie-Stone et al. , 2019).”
InfluentialSelf-citationReview of processing and analysis methods for DNA methylation array data
Wilhelm-Benartzi, C. S., Koestler, D. C., Karagas, M. R., Flanagan, J. M., Christensen, B. C., Kelsey, K. T., Marsit, C. J., Houseman, E. A., & Brown, R. E. (2013) · British Journal of Cancer · doi:10.1038/bjc.2013.496 · Cited by 202 · see citing works
Epigenome-wide association studies identify DNA methylation associated with kidney function
Chu, A. Y., Tin, A., Schlosser, P., Ko, Y., Qiu, C., Yao, C., Joehanes, R., Grams, M. E., Liang, L., Gluck, C. A., et al. (2017) · Nature Communications · doi:10.1038/s41467-017-01297-7 · Cited by 198 · see citing works
DNA Methylation of Lipid-Related Genes Affects Blood Lipid Levels
Pfeiffer, L., Wahl, S., Pilling, L. C., Reischl, E., Sandling, J. K., Kunze, S., Holdt, L. M., Kretschmer, A., Schramm, K., Adamski, J., et al. (2015) · Circulation Cardiovascular Genetics · doi:10.1161/circgenetics.114.000804 · Cited by 194 · see citing works
“3.(8) Table I in the Data Supplement provides a summary of normalized β values of the identified lipid-related CpGs in KORA F4.”
Mining cancer methylomes: prospects and challenges
Stirzaker, C., Taberlay, P. C., Statham, A. L., & Clark, S. J. (2013) · Trends in Genetics · doi:10.1016/j.tig.2013.11.004 · Cited by 192 · see citing works
Soriano‐Tárraga, C., Jiménez-Conde, J., Steinhauer, E. G., Mola-Caminal, M., Vivanco‐Hidalgo, R. M., Ois, Á., Rodríguez-Campello, A., Cuadrado‐Godia, E., Sayols-Baixeras, S., Elosúa, R., et al. (2015) · Human Molecular Genetics · doi:10.1093/hmg/ddv493 · Cited by 191 · see citing works
“05 were removed and cross-reactive probes were excluded (47) using the wateRmelon R package (44).”
MethodHannon, E. J., Gorrie-Stone, T. J., Smart, M., Burrage, J., Hughes, A., Bao, Y., Meena Kumari, B., Schalkwyk, L. C., & Mill, J. S. (2018) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2018.09.007 · Cited by 191 · see citing works
“Data were normalized using the dasen function from the wateRmelon package(Pidsley et al. 2013).”
MethodSelf-citationCecil, C. A. M., Lysenko, L., Jaffee, S. R., Pingault, J., Smith, R. G., Relton, C. L., Woodward, G., McArdle, W. L., Mill, J. S., & Barker, E. D. (2014) · Molecular Psychiatry · doi:10.1038/mp.2014.95 · Cited by 187 · see citing works
Self-citationRojas, D., Rager, J. E., Smeester, L., Bailey, K. A., Drobná, Z., Rubio‐Andrade, M., Stýblo, M., García-Vargas, G. G., & Fry, R. C. (2014) · Toxicological Sciences · doi:10.1093/toxsci/kfu210 · Cited by 187 · see citing works
“Probes that represent known single nucleotide polymorphisms (SNPs) were removed (Pidsley et al., 2013; n¼ 59 732), leaving a total of 424 935 probes for further analyses.”
BackgroundSmith, R. G., Hannon, E., De Jager, P. L., Chibnik, L., Lott, S. J., Condliffe, D., Smith, A. R., Haroutunian, V., Troakes, C., Al‐Sarraj, S., et al. (2018) · Alzheimer's & Dementia · doi:10.1016/j.jalz.2018.01.017 · Cited by 184 · see citing works
Self-citationA cross-package Bioconductor workflow for analysing methylation array data
Maksimovic, J., Phipson, B., & Oshlack, A. (2017) · F1000Research · doi:10.12688/f1000research.8839.3 · Cited by 179 · see citing works
“…packages available that have been developed for analysing methylation array data, including minfi (Aryee et al. 2014), missMethyl (B. Phipson, Maksimovic, and Oshlack 2016), wateRmelon (Pidsley et al. 2013), methylumi (S. Davis et al. 2015), ChAMP (Morris et al. 2014) and charm (Aryee et al. 2011).”
BackgroundMethodMarzi, S. J., Sugden, K., Arseneault, L., Belsky, D. W., Burrage, J., Corcoran, D. L., Danese, A., Fisher, H. L., Hannon, E., Moffitt, T. E., et al. (2018) · American Journal of Psychiatry · doi:10.1176/appi.ajp.2017.17060693 · Cited by 163 · see citing works
“Data were processed with the pfilter function from the wateRmelon package (22), excluding zero samples with .”
MethodSelf-citationGene set enrichment analysis for genome-wide DNA methylation data
Maksimovic, J., Oshlack, A., & Phipson, B. (2021) · Genome biology · doi:10.1186/s13059-021-02388-x · Cited by 165 · see citing works
Smith, J. A., Zhao, W., Wang, X., Ratliff, S. M., Mukherjee, B., Kardia, S. L. R., Liu, Y., Roux, A. V. D., & Needham, B. L. (2017) · Epigenetics · doi:10.1080/15592294.2017.1341026 · Cited by 168 · see citing works
“65 probes that assay highlypolymorphic single nucleotide polymorphisms (SNPs) rather than methylation were also excluded.(53) The methylation level for each site was computed as the M-value, the log ratio of the methylated to the unmethylated signal intensity.”
MethodSmith, R. G., Pishva, E., Shireby, G. L., Smith, A. R., Roubroeks, J. A. Y., Hannon, E. J., Wheildon, G., Mastroeni, D., Gasparoni, G., Riemenschneider, M. J., et al. (2021) · Nature Communications · doi:10.1038/s41467-021-23243-4 · Cited by 163 · see citing works
Self-citationCharacterization of the Cartilage DNA Methylome in Knee and Hip Osteoarthritis
Rushton, M. D., Reynard, L. N., Barter, M. J., Refaie, R., Rankin, K. S., Young, D. A., & Loughlin, J. A. (2014) · Arthritis & Rheumatology · doi:10.1002/art.38713 · Cited by 169 · see citing works
Methylomic profiling of human brain tissue supports a neurodevelopmental origin for schizophrenia
Pidsley, R., Viana, J., Hannon, E., Spiers, H., Troakes, C., Al-Saraj, S., Mechawar, N., Turecki, G., Schalkwyk, L. C., Bray, N. J., et al. (2014) · Genome biology · doi:10.1186/s13059-014-0483-2 · Cited by 157 · see citing works
“, San Diego, CA, USA), performing pre-processing, normalization and stringent quality control as previously described [16] (see Materials and methods; Table S3 in Additional file 1).”
MethodInfluentialNaeem, H., Wong, N. C., Chatterton, Z., Hong, M. K. H., Pedersen, J. S., Corcoran, N. M., Hovens, C. M., & Macintyre, G. (2014) · BMC Genomics · doi:10.1186/1471-2164-15-51 · Cited by 154 · see citing works
Zhong, J., Karlsson, O., Wang, G., Li, J., Guo, Y., Lin, X., Zemplenyi, M., Sánchez-Guerra, M., Trevisi, L., Urch, B., et al. (2017) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1618545114 · Cited by 156 · see citing works
Heiss, J. A., & Just, A. C. (2018) · Clinical Epigenetics · doi:10.1186/s13148-018-0504-1 · Cited by 154 · see citing works
Placental epigenetic clocks: estimating gestational age using placental DNA methylation levels
Lee, Y., Choufani, S., Weksberg, R., Wilson, S. L., Yuan, V., Burt, A., Marsit, C. J., Lu, A. T.-H., Ritz, B. R., Bohlin, J., et al. (2019) · Aging · doi:10.18632/aging.102049 · Cited by 154 · see citing works
“In support of this, we found that the RPC validated in test data that were normalized using various methods: beta-mixture quantile dilation (BMIQ, [30]), quantile normalization (quanNorm, [31, 32]), data-driven separate normalization (dasen, [33]) as detailed in Table 1.”
MethodMarttila, S., Kananen, L., Häyrynen, S., Jylhävä, J., Nevalainen, T., Hervonen, A., Jylhä, M. K., Nykter, M., & Hurme, M. (2015) · BMC Genomics · doi:10.1186/s12864-015-1381-z · Cited by 151 · see citing works
“The methylation data were preprocessed as a methylumiset object using R software with the wateRmelon arrayspecific package from Bioconductor [57].”
MethodHannon, E., Dempster, E. L., Mansell, G., Burrage, J., Bass, N., Bohlken, M. M., Corvin, A., Curtis, C. J., Dempster, D., Di Forti, M., et al. (2021) · eLife · doi:10.7554/elife.58430 · Cited by 132 · see citing works
Self-citationImproving Phenotypic Prediction by Combining Genetic and Epigenetic Associations
Shah, S. H., Bonder, M. J., Marioni, R. E., Zhu, Z., McRae, A. F., Zhernakova, A., Harris, S. E., Liewald, D., Henders, A. K., Mendelson, M., et al. (2015) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2015.05.014 · Cited by 133 · see citing works
RELIC: a novel dye-bias correction method for Illumina Methylation BeadChip
Xu, Z., Langie, S. A. S., Boever, P. D., Taylor, J. A., & Niu, L. (2017) · BMC Genomics · doi:10.1186/s12864-016-3426-3 · Cited by 143 · see citing works
“We did not include the nanes method (implemented in wateRmelon) in the comparisons as the nanet method is better than the nanes method, according to the Table two in [10].”
MethodHedman, Å. K., Mendelson, M., Marioni, R. E., Gustafsson, S., Joehanes, R., Irvin, M. R., Zhi, D., Sandling, J. K., Yao, C., Liu, C., et al. (2017) · Circulation Cardiovascular Genetics · doi:10.1161/circgenetics.116.001487 · Cited by 142 · see citing works
Maros, M., Capper, D., Jones, D. T. W., Hovestadt, V., von Deimling, A., Pfister, S., Benner, A., Zucknick, M., & Sill, M. (2020) · Nature Protocols · doi:10.1038/s41596-019-0251-6 · Cited by 114 · see citing works
“Alternative software packages that provide comparable functionalities are the Bioconductor packages RnBeads 100 , ChAMP 101 and wateRmelon 102 .”
MethodGenome-wide methylation data mirror ancestry information
Rahmani, E., Shenhav, L., Schweiger, R., Yousefi, P., Huen, K., Eskenazi, B., Eng, C. S., Huntsman, S., Hu, D., Galanter, J., et al. (2017) · Epigenetics & Chromatin · doi:10.1186/s13072-016-0108-y · Cited by 141 · see citing works
“Beta Mixture Quantile (BMIQ) [40] normalization was applied to the methylation levels using the R package wateR-melon, version 1.0.3 [41].”
MethodInfluentialLogue, M. W., Smith, A. K., Wolf, E. J., Maniates, H., Stone, A., Schichman, S. A., McGlinchey, R. E., Milberg, W., & Miller, M. W. (2017) · Epigenomics · doi:10.2217/epi-2017-0078 · Cited by 139 · see citing works
McCartney, D. L., Stevenson, A. J., Walker, R. M., Gibson, J., Morris, S. W., Campbell, A., Murray, A. D., Whalley, H. C., Porteous, D. J., McIntosh, A. M., et al. (2018) · Alzheimer s & Dementia Diagnosis Assessment & Disease Monitoring · doi:10.1016/j.dadm.2018.05.006 · Cited by 134 · see citing works
Analysis pipelines and packages for Infinium HumanMethylation450 BeadChip (450k) data
Morris, T. J., & Beck, S. (2014) · Methods · doi:10.1016/j.ymeth.2014.08.011 · Cited by 136 · see citing works
Alzheimer’s disease-associated (hydroxy)methylomic changes in the brain and blood
Lardenoije, R., Roubroeks, J. A. Y., Pishva, E., Leber, M., Wagner, H., Iatrou, A., Smith, A. R., Smith, R. G., Eijssen, L., Kleineidam, L., et al. (2019) · Clinical Epigenetics · doi:10.1186/s13148-019-0755-5 · Cited by 136 · see citing works
Self-citationSystematic underestimation of the epigenetic clock and age acceleration in older subjects
Khoury, L. Y. E., Gorrie-Stone, T. J., Smart, M., Hughes, A., Bao, Y., Andrayas, A., Burrage, J., Hannon, E. J., Meena Kumari, B., Mill, J. S., et al. (2019) · Genome biology · doi:10.1186/s13059-019-1810-4 · Cited by 133 · see citing works
“These were later normalized using dasen function from the wateRmelon package [48].”
MethodSelf-citationvan Dijk, S. J., Peters, T. J., Buckley, M. J., Zhou, J., Jones, P. A., Gibson, R. A., Makrides, M., Mühlhäusler, B. S., & Molloy, P. L. (2017) · International Journal of Obesity · doi:10.1038/ijo.2017.228 · Cited by 135 · see citing works
Sánchez‐Cabo, F., Fuster, V., Silla-Castro, J. C., González, G., Lorenzo-Vivas, E., Álvarez, R., Callejas, S., Benguría, A., Gil, E., Núñez, E., et al. (2023) · European Heart Journal · doi:10.1093/eurheartj/ehad361 · Cited by 137 · see citing works
Senut, M.-C., Sen, A., Cingolani, P., Shaik, A., Land, S. J., & Ruden, D. M. (2014) · Toxicological Sciences · doi:10.1093/toxsci/kfu028 · Cited by 135 · see citing works
“Several studies have indicated that the Infinium 1 probes have a more stable signal and extended dynamic range than the Infinium 2 probes, which differ in that Infinium 1 probes have separate beads for methylated and unmethylated CpG sites whereas Infinium 2 probes consist of single beads that can detect both types of…”
BackgroundA cross-package Bioconductor workflow for analysing methylation array data
Maksimovic, J., Phipson, B., & Oshlack, A. (2016) · F1000Research · doi:10.12688/f1000research.8839.2 · Cited by 132 · see citing works
Genome-wide Methylomic Analysis of Monozygotic Twins Discordant for Adolescent Depression
Dempster, E. L., Wong, C. C. Y., Lester, K. J., Burrage, J., Gregory, A. M., Mill, J. S., & Eley, T. C. (2014) · Biological Psychiatry · doi:10.1016/j.biopsych.2014.04.013 · Cited by 132 · see citing works
“4), and DNA was isolated using a standard protocol (30).”
Self-citationCardona, A., Day, F. R., Perry, J. R. B., Loh, M. C. S., Chu, A. Y., Lehne, B. C., Paul, D. S., Lotta, L. A., Stewart, I. D., Kerrison, N. D., et al. (2019) · Diabetes · doi:10.2337/db18-0290 · Cited by 130 · see citing works
McDermott, E., Ryan, E. J., Tosetto, M., Gibson, D. J., Burrage, J., Keegan, D., Byrne, K., Crowe, E., Sexton, G., Malone, K. M., et al. (2015) · Journal of Crohn s and Colitis · doi:10.1093/ecco-jcc/jjv176 · Cited by 128 · see citing works
Self-citationShireby, G. L., Dempster, E. L., Policicchio, S. S. S., Smith, R. G., Pishva, E., Chioza, B. A., Davies, J. P., Burrage, J., Lunnon, K., Vellame, D. S., et al. (2022) · Nature Communications · doi:10.1038/s41467-022-33394-7 · Cited by 125 · see citing works
Self-citationAllard, C., Desgagné, V., Patenaude, J., Lacroix, M., Guillemette, L., Battista, M., Doyon, M., Ménard, J., Jl, A., Perron, P., et al. (2015) · Epigenetics · doi:10.1080/15592294.2015.1029700 · Cited by 127 · see citing works
Statistical methods for detecting differentially methylated loci and regions
Robinson, M. D., Kahraman, A., Law, C. W., Lindsay, H. A., Nowicka, M., Weber, L. M., & Zhou, X. (2014) · Frontiers in Genetics · doi:10.3389/fgene.2014.00324 · Cited by 124 · see citing works
“The platform can be thought of as genotyping BS-treated DNA to reveal the relative proportion of methylated and unmethylated alleles (Pidsley et al., 2013).”
BackgroundSchmitz, L., Zhao, W., Ratliff, S. M., Goodwin, J., Miao, J., Lu, Q., Guo, X., Taylor, K. D., Ding, J., Liu, Y., et al. (2021) · Epigenetics · doi:10.1080/15592294.2021.1939479 · Cited by 114 · see citing works
“65 probes that assay highly polymorphic single nucleotide polymorphisms (SNPs) rather than methylation were also excluded.(85) The methylation level for each site was computed as the M-value, the log ratio of the methylated to the unmethylated signal intensity.”
MethodEpigenetic signatures of childhood abuse and neglect: Implications for psychiatric vulnerability
Cecil, C. A. M., Smith, R. G., Walton, E., Mill, J. S., McCrory, E. J. P., & Viding, E. M. (2016) · Journal of Psychiatric Research · doi:10.1016/j.jpsychires.2016.09.010 · Cited by 122 · see citing works
Self-citationSmith, A. R., Smith, R. G., Pishva, E., Hannon, E. J., Roubroeks, J. A. Y., Burrage, J., Troakes, C., Al‐Sarraj, S., Sloan, C., Mill, J. S., et al. (2019) · Clinical Epigenetics · doi:10.1186/s13148-019-0636-y · Cited by 118 · see citing works
Self-citationElevated polygenic burden for autism is associated with differential DNA methylation at birth
Hannon, E. J., Schendel, D., Ladd‐Acosta, C., Grove, J., iPSYCH-Broad ASD Group, Hansen, C. S., Andrews, S. V., HOUGAARD, D. M., Bresnahan, M. A., Mors, O. N., et al. (2018) · Genome Medicine · doi:10.1186/s13073-018-0527-4 · Cited by 117 · see citing works
Self-citation
- Authors
- Pidsley, R., Y Wong, C. C., Volta, M., Lunnon, K., Mill, J., & Schalkwyk, L. C.
- Affiliations
- King's College London (GB)
- Corresponding author
- Schalkwyk, Leonard C
- Venue
- BMC Genomics
- Published
- 2013-05-01
- Volume
- 14
- Issue
- 1
- Pages
- 293-293
- ISSN
- 1471-2164 (ISSN-L)
- PMC
- PMC3769145
- OpenAlex
- W2119883345
- Semantic Scholar
- d6780662d62ce843992ea64f68c71fc50c1a2c37
- Cited by
- 1,143
- References
- 24
- Citations by year
- 2026: 74 · 2025: 97 · 2024: 102 · 2023: 109 · 2022: 112 · 2021: 141 · 2020: 94 · 2019: 70 · 2018: 76 · 2017: 94 · 2016: 71 · 2015: 60 · 2014: 33 · 2013: 7
- FWCI
- 22.99
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Locations
- BMC Genomics (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Open Access at Essex (University of Essex) (acceptedVersion, cc-by)
- Europe PMC (PubMed Central) (submittedVersion, cc-by)
- PubMed Central (submittedVersion, cc-by)
- Figshare (submittedVersion)
- Research Portal (King's College London) (submittedVersion, other-oa)
- Open Research Exeter (University of Exeter) (submittedVersion, cc-by)
- Funders
- American Asthma Foundation
- National Institutes of Health
- Medical Research Council
- National Institute on Aging (R01AG036039)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Epigenetics and DNA Methylation
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Syndromes and Imprinting
- Keywords
- DNA methylation, quantile normalization, genomic imprinting, X chromosome inactivation, data quality metrics, watermelon, R package, statistical power
- MeSH terms
- Humans, Statistics as Topic, Genomic Imprinting, DNA Methylation, Computational Biology, Oligonucleotide Array Sequence Analysis, Polymorphism, Single Nucleotide, Chromosomes, Human, X
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-18.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-18.
Cited sources are checked by gpt-6-sol.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
