King's College London

Research portal

Collapsed methylation quantitative trait loci analysis for low frequency and rare variants

Research output: Contribution to journalArticle

Tom G. Richardson, Hashem A. Shihab, Gibran Hemani, Jie Zheng, Eilis Hannon, Jonathan Mill, Elena Carnero-Montoro, Jordana T. Bell, Oliver Lyttleton, Wendy L. McArdle, Susan M. Ring, Santiago Rodriguez, Colin Campbell, George Davey Smith, Caroline L. Relton, Nicholas J. Timpson, Tom R. Gaunt

Original languageEnglish
Pages (from-to)4339-4349
Number of pages11
JournalHuman Molecular Genetics
Volume25
Issue number19
Early online date24 Aug 2016
DOIs
E-pub ahead of print24 Aug 2016
Published1 Oct 2016

Documents

King's Authors

Abstract

BACKGROUND: Single variant approaches have been successful in identifying DNA methylation quantitative trait loci (mQTL), although as with complex traits they lack the statistical power to identify the effects from rare genetic variants. We have undertaken extensive analyses to identify regions of low frequency and rare variants that are associated with DNA methylation levels.

METHODS: We used repeated measurements of DNA methylation from five different life stages in human blood, taken from the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort. Variants were collapsed across CpG islands and their flanking regions to identify variants collectively associated with methylation, where no single variant was individually responsible for the observed signal. All analyses were undertaken using the sequence kernel association test.

RESULTS: For loci where no individual variant mQTL was observed based on a single variant analysis, we identified 95 unique regions where the combined effect of low frequency variants (MAF ≤ 5%) provided strong evidence of association with methylation. For loci where there was previous evidence of an individual variant mQTL, a further 3 regions provided evidence of association between multiple low frequency variants and methylation levels. Effects were observed consistently across 5 different time points in the lifecourse and evidence of replication in the TwinsUK and Exeter cohorts was also identified.

CONCLUSION: We have demonstrated the potential of this novel approach to mQTL analysis by analysing the combined effect of multiple low frequency or rare variants. Future studies should benefit from applying this approach as a complementary follow up to single variant analyses.

Download statistics

No data available

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454