308V Poster Online - Virtual Posters
Tuesday June 07, 11:00 AM - 3:00 PM

Authors:
Margarita Takou ^1,5; Daniel Balick ³; Kim Steige ^1,2; Hannes Dittberner ¹; Ulrike Göbel ¹; Holger Schielzeth ⁴; Juliette de Meaux ¹

Affiliations:
1) University of Cologne; 2) University of Hohenheim; 3) Harvard Medical School; 4) University of Jena; 5) Penn State University

Keywords:
Natural selection

The distinction between additive and non-additive components of genetic variance is crucial for predicting the adaptive potential of a variable trait. Yet we know surprisingly little about the genomic and evolutionary factors that influence non-additive genetic variance in natural populations. Here, we use a quantitative genetic breeding design to decompose the additive and non-additive components of variance of 17,657 gene transcripts in the outcrossing plant Arabidopsis lyrata. We find that the major fraction of gene expression variance corresponds to non-additive variance in our data. Transcripts showing the highest levels of non-additive variance tend to be longer, cluster in larger groups of co-expressed genes, and are particularly enriched among genes involved in cellular differentiation and epigenetic reprogramming of gene expression. Furthermore, we find that genes with the highest fraction of non-additive variation are exposed to stronger stabilizing selection than genes with higher additive variance. As a source of many evolutionary novelties, gene expression variation plays a key role in adaptive evolution, yet the amount of non-additive genetic variation will limit its capacity to respond to selection. Our study is the first to show that stabilizing selection on amino-acid sequences associates with a decrease in a population’s potential for adapting gene expression levels.