Tiny microbes, enormous impacts: what matters in gut microbiome studies?

Justine Debelius,Yoshiki Vazquez-Baeza,Rob Knight,Se Jin Song,Antonio Gonzalez,Zhenjiang Zech Xu

doi:10.1186/s13059-016-1086-x

Justine Debelius, Yoshiki Vazquez-Baeza + Show 4 more

Open Access

https://doi.org/10.1186/s13059-016-1086-x

Copy DOI

Abstract

Many factors affect the microbiomes of humans, mice, and other mammals, but substantial challenges remain in determining which of these factors are of practical importance. Considering the relative effect sizes of both biological and technical covariates can help improve study design and the quality of biological conclusions. Care must be taken to avoid technical bias that can lead to incorrect biological conclusions. The presentation of quantitative effect sizes in addition to P values will improve our ability to perform meta-analysis and to evaluate potentially relevant biological effects. A better consideration of effect size and statistical power will lead to more robust biological conclusions in microbiome studies.

Highlights

The human microbiome is a virtual organ that contains >100 times as many genes as the human genome [1]
The Flemish Gut Flora Project, which used 16S rRNA amplicon sequencing on a cohort of 1106 individuals, identified 69 variables relating to the subjects that correlated with the microbiome, including use of 13 drugs ranging from antibiotics to antidepressants, and explained 7.7 % of the variation in the microbiome
The American Gut Project, with over 10,000 samples processed, is a crowd-sourced microbiome study that expands on the effects considered by the Human Microbiome Project (HMP) to evaluate microbial diversity across Western populations with fewer restrictions on health and lifestyle

Summary

Introduction

The human microbiome is a virtual organ that contains >100 times as many genes as the human genome [1]. Despite problems in generalizing findings across some microbiome studies that result from the factors noted above, we are beginning to understand how the effect sizes of specific biological and technical variables in community profiling are structured relative to others. We argue that by explicitly considering and quantifying effect sizes in microbiome studies, we can better design experiments that limit confounding factors This principle is well established in other fields, such as ecology [35], epidemiology (see for example [36]), and genome-wide association studies (their relationship to microbiome studies is reviewed in [37]).

Findings

Conclusions