A simulation study on power of epigenome-wide association analysis using disease-discordant twin design

Publikation: Konferencebidrag uden forlag/tidsskriftPosterForskningpeer review


Identical twin pairs discordant for a disease have been widely used in epigenome-wide association studies (EWAS) of complex diseases because the underlying genetic background is cancelled out. Although the twin design is deemed to have enriched power as compared with ordinary case-control design, its statistical power has not been well investigated.
We perform a computer simulation study on the power of the disease discordant twin design assuming that both genetic and environmental factors contribute to the liability of a disease phenotype. Power is assessed under different levels of genetic contribution or heritability for various effect sizes of environmental exposure that induce change in the level and variance of the normal-distributed M value of DNA methylation.
In general, the discordant twin design outperforms ordinary case-control design especially for CpG sites with moderate effect and for diseases with heritability>0.3. The power advantage in discordant twin design becomes more obvious with increasing heritability. For diseases with no genetic background, the twin design is slightly underpowered, while for high heritability (0.5), high intrapair methylation correlation (0.5) and small effect CpG sites (R2 = 0.1), sample size required for power over 0.8 is 215 for ordinary case-control design, and only 59 for discordant twin design, which is an over three folds reduction.
The twin design is powerful in detecting epigenetic alterations induced by environmental exposure especially for complex diseases with a strong genetic background. The use of twin design should be encouraged in EWAS using microarray and high-cost bisulfite sequencing techniques.
Publikationsdato28. maj 2017
StatusUdgivet - 28. maj 2017
Begivenhed50th European Human Genetics Conference: Anniversary Meeting - Bella Center, Center Blvd. 5, Copenhagen, Danmark
Varighed: 27. maj 201730. maj 2017
Konferencens nummer: 50


Konference50th European Human Genetics Conference
LokationBella Center, Center Blvd. 5


  • Epigenetic study
  • Simulation
  • Twins


Dyk ned i forskningsemnerne om 'A simulation study on power of epigenome-wide association analysis using disease-discordant twin design'. Sammen danner de et unikt fingeraftryk.