Limits...
Meta-analysis of human methylation data for evidence of sex-specific autosomal patterns.

McCarthy NS, Melton PE, Cadby G, Yazar S, Franchina M, Moses EK, Mackey DA, Hewitt AW - BMC Genomics (2014)

Bottom Line: After excluding non-specific probes, we found 184 autosomal CpG sites differentially methylated by sex after correction for multiple testing.Many of the most significantly associated CpG probes were new.This study represents a comprehensive analysis of sex-specific methylation patterns.

View Article: PubMed Central - PubMed

Affiliation: Centre for the Genetic Origins of Health and Disease (GOHaD), University of Western Australia, Perth, Australia. nina.mccarthy@uwa.edu.au.

ABSTRACT

Background: Several individual studies have suggested that autosomal CpG methylation differs by sex both in terms of individual CpG sites and global autosomal CpG methylation. However, these findings have been inconsistent and plagued by spurious associations due to the cross reactivity of CpG probes on commercial microarrays. We collectively analysed 76 published studies (n = 6,795) for sex-associated differences in both autosomal and sex chromosome CpG sites.

Results: Overall autosomal methylation profiles varied substantially by study, and we encountered substantial batch effects. We accounted for these by conducting random effects meta-analysis for individual autosomal CpG methylation associations. After excluding non-specific probes, we found 184 autosomal CpG sites differentially methylated by sex after correction for multiple testing. In line with previous studies, average beta differences were small. Many of the most significantly associated CpG probes were new. Of note was differential CpG methylation in the promoters of genes thought to be involved in spermatogenesis and male fertility, such as SLC9A2, SPESP1, CRISP2, and NUPL1. Pathway analysis revealed overrepresentation of genes differentially methylated by sex in several broad Gene Ontology biological processes, including RNA splicing and DNA repair.

Conclusions: This study represents a comprehensive analysis of sex-specific methylation patterns. We demonstrate the existence of sex-specific methylation profiles and report a large number of novel DNA methylation differences in autosomal CpG sites between sexes.

Show MeSH

Related in: MedlinePlus

Plot of the first two principal components (PC1 and PC2) based on the 999 X chromosome CpG sites for known males (n = 2,277), females (n = 2,870)and other (n = 60) (plot A) and with those of unknown sex (n = 2,126) superimposed (plot B).
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4255932&req=5

Fig1: Plot of the first two principal components (PC1 and PC2) based on the 999 X chromosome CpG sites for known males (n = 2,277), females (n = 2,870)and other (n = 60) (plot A) and with those of unknown sex (n = 2,126) superimposed (plot B).

Mentions: We initially sought to investigate whether sex could be inferred from X chromosome methylation data using principal component analysis (PCA). The first two principal components (PCs) were plotted against each other for all samples of known sex (Figure 1(A)). Colouring by the recorded sex from the EBI phenotype files indicates that sex can be determined by classifying samples based on their first PC, with samples recorded as ‘other/trisomy’ (n = 60) clustering in the middle. The second PC, in contrast, contributes little to the separation of males and females. Figure 1(B) shows that the samples of unknown sex cluster well with those of known sex. Logistic regression of recorded sex on PC1 and recorded sex on PC2 showed that both relationships were significant and, as expected, PC1 was a much better predictor than PC2 (PC1: P < 2e-16, AIC 2447.4; PC2: P = 1e-08, AIC 7037.2).Figure 1


Meta-analysis of human methylation data for evidence of sex-specific autosomal patterns.

McCarthy NS, Melton PE, Cadby G, Yazar S, Franchina M, Moses EK, Mackey DA, Hewitt AW - BMC Genomics (2014)

Plot of the first two principal components (PC1 and PC2) based on the 999 X chromosome CpG sites for known males (n = 2,277), females (n = 2,870)and other (n = 60) (plot A) and with those of unknown sex (n = 2,126) superimposed (plot B).
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4255932&req=5

Fig1: Plot of the first two principal components (PC1 and PC2) based on the 999 X chromosome CpG sites for known males (n = 2,277), females (n = 2,870)and other (n = 60) (plot A) and with those of unknown sex (n = 2,126) superimposed (plot B).
Mentions: We initially sought to investigate whether sex could be inferred from X chromosome methylation data using principal component analysis (PCA). The first two principal components (PCs) were plotted against each other for all samples of known sex (Figure 1(A)). Colouring by the recorded sex from the EBI phenotype files indicates that sex can be determined by classifying samples based on their first PC, with samples recorded as ‘other/trisomy’ (n = 60) clustering in the middle. The second PC, in contrast, contributes little to the separation of males and females. Figure 1(B) shows that the samples of unknown sex cluster well with those of known sex. Logistic regression of recorded sex on PC1 and recorded sex on PC2 showed that both relationships were significant and, as expected, PC1 was a much better predictor than PC2 (PC1: P < 2e-16, AIC 2447.4; PC2: P = 1e-08, AIC 7037.2).Figure 1

Bottom Line: After excluding non-specific probes, we found 184 autosomal CpG sites differentially methylated by sex after correction for multiple testing.Many of the most significantly associated CpG probes were new.This study represents a comprehensive analysis of sex-specific methylation patterns.

View Article: PubMed Central - PubMed

Affiliation: Centre for the Genetic Origins of Health and Disease (GOHaD), University of Western Australia, Perth, Australia. nina.mccarthy@uwa.edu.au.

ABSTRACT

Background: Several individual studies have suggested that autosomal CpG methylation differs by sex both in terms of individual CpG sites and global autosomal CpG methylation. However, these findings have been inconsistent and plagued by spurious associations due to the cross reactivity of CpG probes on commercial microarrays. We collectively analysed 76 published studies (n = 6,795) for sex-associated differences in both autosomal and sex chromosome CpG sites.

Results: Overall autosomal methylation profiles varied substantially by study, and we encountered substantial batch effects. We accounted for these by conducting random effects meta-analysis for individual autosomal CpG methylation associations. After excluding non-specific probes, we found 184 autosomal CpG sites differentially methylated by sex after correction for multiple testing. In line with previous studies, average beta differences were small. Many of the most significantly associated CpG probes were new. Of note was differential CpG methylation in the promoters of genes thought to be involved in spermatogenesis and male fertility, such as SLC9A2, SPESP1, CRISP2, and NUPL1. Pathway analysis revealed overrepresentation of genes differentially methylated by sex in several broad Gene Ontology biological processes, including RNA splicing and DNA repair.

Conclusions: This study represents a comprehensive analysis of sex-specific methylation patterns. We demonstrate the existence of sex-specific methylation profiles and report a large number of novel DNA methylation differences in autosomal CpG sites between sexes.

Show MeSH
Related in: MedlinePlus