Limits...
Social networks predict gut microbiome composition in wild baboons.

Tung J, Barreiro LB, Burns MB, Grenier JC, Lynch J, Grieneisen LE, Altmann J, Alberts SC, Blekhman R, Archie EA - Elife (2015)

Bottom Line: Social relationships have profound effects on health in humans and other primates, but the mechanisms that explain this relationship are not well understood.Using shotgun metagenomic data from wild baboons, we found that social group membership and social network relationships predicted both the taxonomic structure of the gut microbiome and the structure of genes encoded by gut microbial species.Our results argue that social interactions are an important determinant of gut microbiome composition in natural animal populations-a relationship with important ramifications for understanding how social relationships influence health, as well as the evolution of group living.

View Article: PubMed Central - PubMed

Affiliation: Department of Evolutionary Anthropology, Duke University, Durham, United States.

ABSTRACT
Social relationships have profound effects on health in humans and other primates, but the mechanisms that explain this relationship are not well understood. Using shotgun metagenomic data from wild baboons, we found that social group membership and social network relationships predicted both the taxonomic structure of the gut microbiome and the structure of genes encoded by gut microbial species. Rates of interaction directly explained variation in the gut microbiome, even after controlling for diet, kinship, and shared environments. They therefore strongly implicate direct physical contact among social partners in the transmission of gut microbial species. We identified 51 socially structured taxa, which were significantly enriched for anaerobic and non-spore-forming lifestyles. Our results argue that social interactions are an important determinant of gut microbiome composition in natural animal populations-a relationship with important ramifications for understanding how social relationships influence health, as well as the evolution of group living.

No MeSH data available.


Related in: MedlinePlus

Enrichment of low p-values in the data vs an empirical : between group analyses.To confirm that our modeling approach (quantile normalization of species relative abundances, followed by mixed effects modeling in GEMMA) did not bias us towards detecting false positives, we compared the signal in our true data set against an empirically derived . The histogram distribution of p-values for the true data (gold) is plotted against the distribution of p-values from 10 permutations (blue). In each permutation, group membership was scrambled across the data set while keeping the modeling approach, kinship structure, and all other covariates constant. The inset shows a quantile–quantile plot of the same data, with clear enrichment of differentially abundant species in the actual data vs the empirical . No differentially abundant species are detected at a 10% FDR in the permuted data sets, while 64 are discovered in the true data set.DOI:http://dx.doi.org/10.7554/eLife.05224.012
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4379495&req=5

fig3s1: Enrichment of low p-values in the data vs an empirical : between group analyses.To confirm that our modeling approach (quantile normalization of species relative abundances, followed by mixed effects modeling in GEMMA) did not bias us towards detecting false positives, we compared the signal in our true data set against an empirically derived . The histogram distribution of p-values for the true data (gold) is plotted against the distribution of p-values from 10 permutations (blue). In each permutation, group membership was scrambled across the data set while keeping the modeling approach, kinship structure, and all other covariates constant. The inset shows a quantile–quantile plot of the same data, with clear enrichment of differentially abundant species in the actual data vs the empirical . No differentially abundant species are detected at a 10% FDR in the permuted data sets, while 64 are discovered in the true data set.DOI:http://dx.doi.org/10.7554/eLife.05224.012

Mentions: Here, y is the n by 1 vector of normalized taxon abundances for the n individuals in the sample; μ is the intercept; x is the n by 1 vector denoting social group membership; and βx is the effect size of social group membership. For the other covariates, a is the n by 1 vector denoting age and βa describes its effects on taxon abundance; s is the n by 1 vector denoting sex and βs its effect size; and r is the n by 1 vector denoting read depth and βr its effect size. The n by 1 vector of u is a random effects term to control for relatedness, and the n by n matrix K provides pedigree-based estimates of relatedness. Residual errors are represented by ε, an n by 1 vector, and MVN denotes the multivariate normal distribution. We interpreted significantly non-zero βx values as support for differences in taxon abundance between social groups, using a false discovery rate threshold of 10% (Storey and Tibshirani, 2003) after checking that an empirically derived distribution of p-values for this analysis was uniform (Figure 3—figure supplement 1).


Social networks predict gut microbiome composition in wild baboons.

Tung J, Barreiro LB, Burns MB, Grenier JC, Lynch J, Grieneisen LE, Altmann J, Alberts SC, Blekhman R, Archie EA - Elife (2015)

Enrichment of low p-values in the data vs an empirical : between group analyses.To confirm that our modeling approach (quantile normalization of species relative abundances, followed by mixed effects modeling in GEMMA) did not bias us towards detecting false positives, we compared the signal in our true data set against an empirically derived . The histogram distribution of p-values for the true data (gold) is plotted against the distribution of p-values from 10 permutations (blue). In each permutation, group membership was scrambled across the data set while keeping the modeling approach, kinship structure, and all other covariates constant. The inset shows a quantile–quantile plot of the same data, with clear enrichment of differentially abundant species in the actual data vs the empirical . No differentially abundant species are detected at a 10% FDR in the permuted data sets, while 64 are discovered in the true data set.DOI:http://dx.doi.org/10.7554/eLife.05224.012
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4379495&req=5

fig3s1: Enrichment of low p-values in the data vs an empirical : between group analyses.To confirm that our modeling approach (quantile normalization of species relative abundances, followed by mixed effects modeling in GEMMA) did not bias us towards detecting false positives, we compared the signal in our true data set against an empirically derived . The histogram distribution of p-values for the true data (gold) is plotted against the distribution of p-values from 10 permutations (blue). In each permutation, group membership was scrambled across the data set while keeping the modeling approach, kinship structure, and all other covariates constant. The inset shows a quantile–quantile plot of the same data, with clear enrichment of differentially abundant species in the actual data vs the empirical . No differentially abundant species are detected at a 10% FDR in the permuted data sets, while 64 are discovered in the true data set.DOI:http://dx.doi.org/10.7554/eLife.05224.012
Mentions: Here, y is the n by 1 vector of normalized taxon abundances for the n individuals in the sample; μ is the intercept; x is the n by 1 vector denoting social group membership; and βx is the effect size of social group membership. For the other covariates, a is the n by 1 vector denoting age and βa describes its effects on taxon abundance; s is the n by 1 vector denoting sex and βs its effect size; and r is the n by 1 vector denoting read depth and βr its effect size. The n by 1 vector of u is a random effects term to control for relatedness, and the n by n matrix K provides pedigree-based estimates of relatedness. Residual errors are represented by ε, an n by 1 vector, and MVN denotes the multivariate normal distribution. We interpreted significantly non-zero βx values as support for differences in taxon abundance between social groups, using a false discovery rate threshold of 10% (Storey and Tibshirani, 2003) after checking that an empirically derived distribution of p-values for this analysis was uniform (Figure 3—figure supplement 1).

Bottom Line: Social relationships have profound effects on health in humans and other primates, but the mechanisms that explain this relationship are not well understood.Using shotgun metagenomic data from wild baboons, we found that social group membership and social network relationships predicted both the taxonomic structure of the gut microbiome and the structure of genes encoded by gut microbial species.Our results argue that social interactions are an important determinant of gut microbiome composition in natural animal populations-a relationship with important ramifications for understanding how social relationships influence health, as well as the evolution of group living.

View Article: PubMed Central - PubMed

Affiliation: Department of Evolutionary Anthropology, Duke University, Durham, United States.

ABSTRACT
Social relationships have profound effects on health in humans and other primates, but the mechanisms that explain this relationship are not well understood. Using shotgun metagenomic data from wild baboons, we found that social group membership and social network relationships predicted both the taxonomic structure of the gut microbiome and the structure of genes encoded by gut microbial species. Rates of interaction directly explained variation in the gut microbiome, even after controlling for diet, kinship, and shared environments. They therefore strongly implicate direct physical contact among social partners in the transmission of gut microbial species. We identified 51 socially structured taxa, which were significantly enriched for anaerobic and non-spore-forming lifestyles. Our results argue that social interactions are an important determinant of gut microbiome composition in natural animal populations-a relationship with important ramifications for understanding how social relationships influence health, as well as the evolution of group living.

No MeSH data available.


Related in: MedlinePlus