235T Poster - Population Genetics
Thursday June 09, 9:15 PM - 10:00 PM

Authors:
Xinyi Li ¹; John Novembre ^2,3; Jeremy Berg ²

Affiliations:
1) Committee on Genetics, Genomics & Systems Biology, University of Chicago, Chicago, IL; 2) Department of Human Genetics, University of Chicago, Chicago, IL; 3) Department of Ecology and Evolution, University of Chicago, Chicago, IL

Keywords:
Natural selection

How different evolutionary processes maintain phenotypic variation is an important question in human genetics. While the importance of background selection in shaping patterns of neutral genetic diversity is well-studied, its influence on the genetic architecture and the prevalence of complex disease is not well understood.

First, we investigated how background selection influences negatively selected mutations. By approximating the effect of background selection as a reduction in effective population size, we derived the predicted genetic diversity reduction of deleterious variants. We found while background selection reduces the genetic diversity of weakly selected sites (4Ns < 5), it has minimal effect on strongly selected sites. We performed forward simulation to confirm our theoretical predictions.

Second, as a direct consequence of genetic diversity reduction by background selection, when the causal variants of a complex trait are effectively neutral or weakly selected, background selection reduces heritability of the traits, confirmed by simulations.

Third, to study the disease prevalence, we extended the liability threshold model to include background selection. In the model, mutational pressure increases the liability while selection acts to reduce it, and the equilibrium disease prevalence arises due to a balance between the two forces. We found that when background selection is included, it alters this equilibrium by reducing the genetic variance for liability, which drives an increase in disease prevalence. To validate our theoretical results, we performed forward-time SLIM simulations and measure disease prevalence with varying intensity of background selection. We found scenarios where, for example, when the diversity reduction due to background selection is roughly 20% (as estimated in humans), the disease prevalence increases 10%. In addition, we also examined whether the distortion of site frequency spectrum changes the disease prevalence and found no meaningful effects.

From these investigations, we concluded that background selection reduces genetic diversity of weakly selected variants, but not strongly selected variants. As a result, background selection shifts the distribution of genetic variance more toward rare alleles with large effect than it otherwise would be. In addition, background selection increases disease prevalence primarily through overall levels of diversity. This project addresses the importance of population genetic models in understanding phenotypic variation.