How longitudinal data can contribute to our understanding of host genetic effects on the gut microbiome

Abstract

A key component of microbiome research is understanding the role of host genetic influence on gut microbial composition. However, it can be difficult to link host genetics with gut microbial composition because host genetic similarity and environmental similarity are often correlated. Longitudinal microbiome data can supplement our understanding of the relative role of genetic processes in the microbiome. These data can reveal environmentally contingent host genetic effects, both in terms of controlling for environmental differences and in comparing how genetic effects differ by environment. Here, we explore four research areas where longitudinal data could lend new insights into host genetic effects on the microbiome: microbial heritability, microbial plasticity, microbial stability, and host and microbiome population genetics. We conclude with a discussion of methodological considerations for future studies.

Keywords: Gut microbiome; heritability; host genetics; host-microbiome interactions; plasticity; time series.

Figure 1.

Genetic and non-genetic variance components of the microbiome for four highly heritable taxa and one metric of overall microbiome composition from up to 585 baboon subjects measured (A) at one, two, five, ten, and twenty timepoints per subject; and (B) with timepoints grouped by hydrological year, with varying numbers of timepoints per subject across a mean of 220 subjects per year. These data were generated from a data set published as part of. Figure 1A emphasizes that increasing the number of samples per subject affects our ability to detect host genetic effects, while Figure 1B shows that the relative contribution of genetic and environmental components to overall microbiome variance can change over time.

Figure 2.

Stability and reproducibility of the microbiome. (A) A horizon plot shows how a taxon’s abundance does not change in consistent ways across time points between individual subjects. Band colors represent quartiles relative to the median. (B) A violin plot depicts the mean ± SD relative abundance of a single taxon in each individual subject, highlighting that stability is personalized. These data are from the demonstration data set from the BiomeHorizon package.