Analysis of the Genetic Variation and Age Effects on Gene Expression Using RNA-seq Data from Multiple Tissues. A. Vi˝uela1, A. A. Brown2, A. Buil3, M. N. Davies1, P. Tsai1, J. T. Bell1, K. S. Small1, E. T. Dermitzakis3, R. Durbin2, T. D. Spector1 1) Department of Twin Research & Genetic Epidemiology, Kings College London, London, London, United Kingdom; 2) Wellcome Trust Sanger Institute, Hinxton, United Kingdom; 3) Department of Genetic Medicine, University of Geneva, Switzerland.
Aging can be seen as the cumulative sum of all environments an organism is exposed to over the course of its life. Effects of aging on gene expression are diverse, manifesting on mean and variance in expression, splicing, and environmental changes to genetic regulation (GxE). We analysed RNA-seq data from adipose, skin, whole blood, and lymphoblastoid cell lines (LCLs) from ~850 female adult twins from the TwinsUK cohort (39-85 years old). We identified 905 (fat), 4307 (skin), 480 (blood), and 7 (LCLs) genes where the level of expression was associated with age. Only 430 genes were associated in two or more tissues, suggesting the aging process acts in tissue specific ways. Using fat methylome data available for 552 of the twins (Grundberg, 2013) and applying Bayesian Networks, we tested whether changes in expression with age were mediated by epigenetic markers. In most cases we found little evidence that epigenetic markers were involved in differential expression. To further understand whether these changes were due to GxE or environmental factors we investigated changes in discordance of expression within monozygous twin (MZ) pairs with age. We found 2 genes in adipose, 152 in skin, 1 in blood and 26 in LCLs where discordance was age-dependent. As MZ twins are genetically identical, age-dependent differences must be due to a changing environmental component. Decomposition of variance showed that age related genes had a larger genetic component, and that the sources of variation were highly tissue specific. A more sophisticated approach looked at changes in heritability of expression with age; observing in general that genetic factors explained decreasing proportions of variance as people aged. While this could be due to increased stochasticity, it is plausible that some of this effect is due to eQTL being modified as individuals age. To identify such SNPs whose effects on gene expression changed with age we performed a genome-wide scan looking for age-genotype interactions (GxA). One gene, CD82, was genome-wide significant in fat, showing a concrete example of the how genetic control of expression is modified over time. Interestingly this gene, a metastasis suppressor, showed increased expression with age in individuals with a particular, potentially protective, allele. In summary, we have produced a comprehensive description of how aging affects expression and its genetic control, observing that these effects are frequently tissue specific.
You may contact the first author (during and after the meeting) at