Integrating genetics and social science: genetic risk scores

Abstract

The sequencing of the human genome and the advent of low-cost genome-wide assays that generate millions of observations of individual genomes in a matter of hours constitute a disruptive innovation for social science. Many public use social science datasets have or will soon add genome-wide genetic data. With these new data come technical challenges, but also new possibilities. Among these, the lowest-hanging fruit and the most potentially disruptive to existing research programs is the ability to measure previously invisible contours of health and disease risk within populations. In this article, we outline why now is the time for social scientists to bring genetics into their research programs. We discuss how to select genetic variants to study. We explain how the polygenic architecture of complex traits and the low penetrance of individual genetic loci pose challenges to research integrating genetics and social science. We introduce genetic risk scores as a method of addressing these challenges and provide guidance on how genetic risk scores can be constructed. We conclude by outlining research questions that are ripe for social science inquiry.

Figure 1. Polygenic risk for complex health conditions is continuously and normally distributed

Figure 2. Three models of genetic contributions to social gradients in health

In each panel, the densities graph hypothesized distributions of polygenic risk (left-side y-axis) and the slopes graph hypothesized associations between polygenic risk and morbidity (right-side y-axis). Red lines show cases in which the distribution of polygenic risk (density) or the genetic gradient in disease risk (slope) are the same in socially advantaged and socially disadvantaged population strata. For example, in Models 1 and 3, the distribution of polygenic risk is shared across social strata. Blue lines show where the distribution of polygenic risk or the genetic gradient in disease risk are different in socially advantaged and socially disadvantaged population strata. For example, in Model 2, the distribution of polygenic risk is shifted to the right in the socially disadvantaged population stratum and to the left in the socially advantaged population stratum (the socially disadvantaged population stratum carries a higher burden of genetic risk as compared to the socially advantaged stratum).