Genotype × environment interactions in gene regulation and complex traits

Abstract

Genotype × environment interactions (GxE) have long been recognized as a key mechanism underlying human phenotypic variation. Technological developments over the past 15 years have dramatically expanded our appreciation of the role of GxE in both gene regulation and complex traits. The richness and complexity of these datasets also required parallel efforts to develop robust and sensitive statistical and computational approaches. Although our understanding of the genetic architecture of molecular and complex traits has been maturing, a large proportion of complex trait heritability remains unexplained. Furthermore, there are increasing efforts to characterize the effect of environmental exposure on human health. We therefore review GxE in human gene regulation and complex traits, advocating for a comprehensive approach that jointly considers genetic and environmental factors in human health and disease.

Fig. 1 ∣. Molecular mechanisms of GxE.

a, Different gene-regulatory mechanisms can be impacted by GxE, such as (1) DNA methylation (Me), (2) chromatin accessibility, (3) DNA acetylation (Ac), (4) transcription factor (TF) binding, (5) alternative splicing and (6) transcription. b, Examples of GxE for each of these gene-regulatory mechanisms. The two boxes represent different environmental contexts, with each genomic region affected by a gene-regulatory mechanism containing a SNP with a heterozygous genotype (A and B alleles are indicated in blue and yellow, respectively). Both genetic and environmental contexts contribute to the outcome measured in the molecular readout, typically by high-throughput sequencing. c, An example plot for how data obtained from these molecular readouts in a population sample are visualized. The horizontal axis in each plot indicates the genotype at one of the SNPs, while the vertical axis gives a value for the molecular readout. The example plots indicate a GxE for environment 1, but not for environment 2, as the mean of the molecular readout is similar across all genotypes.

Fig. 2 ∣. Summary of popular methods to detect molecular GxE.

a, QTL mapping following in vivo exposures In a cohort. Here, the cohort’s exposure to an environment is measured, biological samples are collected, and a molecular readout (for example, gene expression) and genotypes are measured to perform GxE QTL mapping. b, QTL mapping following in vitro exposures in cells from a cohort. Here, biological samples are collected from a cohort, the samples are exposed to an environment, and a molecular readout and genotypes are measured to perform GxE QTL mapping. c, MPRAs. Here, constructs (often containing a minimal promoter, a reporter gene and a sequence of interest) are transfected into cells, and a molecular readout is measured and compared for constructs with different alleles and across environmental contexts. d, An example of a GxE eQTL is shown, in which both environmental and genetic context determine the resulting gene expression measured by RNA sequencing. e, Examples of graphical representations of GxE QTL data, showing a scatterplot with regression lines for the three genotypes across a quantitative environmental exposure or a box plot for a qualitative environmental exposure. f, Example of a graphical representation of cASE, where a bar graph shows the measured activity of each allele–environment combination.

Fig. 3 ∣. Effect of latent environments (unmeasured environmental variables) on gene expression.

a, Here, the blue and red environments (Exposure A and B, respectively) are measured, but the yellow environment (Unknown exposure (X)) is not; however, all of these factors contribute to the observed gene expression levels. Using computational methods, effects of latent environments can be inferred from gene expression data or other molecular phenotypes. b, Chromatin accessibility data can be used to generate an annotation of environment-responsive regulatory elements; if these are enriched for eQTL, this suggests that there is a latent environment.