CoChat

Taxa-function robustness in microbial communities

Open access
Eng, Alexander, Borenstein, Elhanan · 2018 · Microbiome

The DOI matches its Crossref record; 62 of the 62 references it cites were found in a registry or an index.

Abstract

BACKGROUND: The species composition of a microbial community is rarely fixed and often experiences fluctuations of varying degrees and at varying frequencies. These perturbations to a community's taxonomic profile naturally also alter the community's functional profile-the aggregate set of genes encoded by community members-ultimately altering the community's overall functional capacities. The magnitude of such functional changes and the specific shift that will occur in each function, however, are strongly dependent on how genes are distributed across community members' genomes. This gene distribution, in turn, is determined by the taxonomic composition of the community and would markedly differ, for example, between communities composed of species with similar genomic content vs. communities composed of species whose genomes encode relatively distinct gene sets. Combined, these observations suggest that community functional robustness to taxonomic perturbations could vary widely across communities with different compositions, yet, to date, a systematic study of the inherent link between community composition and robustness is lacking. RESULTS: In this study, we examined how a community's taxonomic composition influences the robustness of that community's functional profile to taxonomic perturbation (here termed taxa-function robustness) across a wide array of environments. Using a novel simulation-based computational model to quantify this taxa-function robustness in host-associated and non-host-associated communities, we find notable differences in robustness between communities inhabiting different body sites, including significantly higher robustness in gut communities compared to vaginal communities that cannot be attributed solely to differences in species richness. We additionally find between-site differences in the robustness of specific functions, some of which are potentially related to site-specific environmental conditions. These taxa-function robustness differences are most strongly associated with differences in overall functional redundancy, though other aspects of gene distribution also influence taxa-function robustness in certain body environments, and are sufficient to cluster communities by environment. Further analysis revealed a correspondence between our robustness estimates and taxonomic and functional shifts observed across human-associated communities. CONCLUSIONS: Our analysis approach revealed intriguing taxa-function robustness variation across environments and identified features of community and gene distribution that impact robustness. This approach could be further applied for estimating taxa-function robustness in novel communities and for informing the design of synthetic communities with specific robustness requirements.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s40168-018-0425-4

Cites

62 of 62 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 An illustration of the taxa-function relationship and response curves. a The functional profile of a community is a linear combination of the functional profile of each taxon in a community (the copy number of each gene in each taxon’s genome) weighted by the abundance of each taxon in that community. Note that here, a taxon need not be a species, but can instead represent any subpopulation of the community with shared genomic content (such as a single strain of a particular species). b The taxa-function relationship can be modeled as a high-dimensional landscape, linking each community composition to the corresponding functional profile. Here, we show an extremely simplified two-dimensional abstraction of this model to illustrate the impact of this landscape on taxa-function robustness. In this illustration, each coordinate on the plane corresponds to a specific taxonomic composition, with points close to one another corresponding to communities with similar taxonomic compositions. The height represents the functional profile of the community (and in this simplified illustration can denote, for example, the abundance of some function). Notably, the local topology of this landscape around a specific taxonomic composition (e.g., point 1) determines whether minor changes in that composition (represented as small movements on this plane; black arrows) will induce small (point 1) or large (point 2) functional shifts. c Depending on the exact distribution of genes across the genomes of species in a community, changes to the taxonomic composition of a community can produce functional shifts of varying magnitudes. For example, if the distribution of genes differs markedly between two communities (e.g., high vs. low redundancy), similar taxonomic composition perturbations to both communities may produce drastically different functional shifts. d To model the relationship between taxonomic perturbations and functional shifts in a given community, a taxa-function response curve is obtained by fitting a power function to a large array of measurements of functional shifts associated with many different taxonomic perturbations of varying magnitudes. e The taxa-function response curve can be decomposed into two factors: attenuation, which describes how quickly functional shifts increase in magnitude as taxonomic perturbations increase, and buffering, which indicates how well functional shifts are suppressed at smaller taxonomic perturbations
Fig. 1 An illustration of the taxa-function relationship and response curves. a The functional profile of a community is a linear combination of the functional profile of each taxon in a community (the copy number of each gene in each taxon’s genome) weighted by the abundance of each taxon in that community. Note that here, a taxon need not be a species, but can instead represent any subpopulation of the community with shared genomic content (such as a single strain of a particular species). b The taxa-function relationship can be modeled as a high-dimensional landscape, linking each community composition to the corresponding functional profile. Here, we show an extremely simplified two-dimensional abstraction of this model to illustrate the impact of this landscape on taxa-function robustness. In this illustration, each coordinate on the plane corresponds to a specific taxonomic composition, with points close to one another corresponding to communities with similar taxonomic compositions. The height represents the functional profile of the community (and in this simplified illustration can denote, for example, the abundance of some function). Notably, the local topology of this landscape around a specific taxonomic composition (e.g., point 1) determines whether minor changes in that composition (represented as small movements on this plane; black arrows) will induce small (point 1) or large (point 2) functional shifts. c Depending on the exact distribution of genes across the genomes of species in a community, changes to the taxonomic composition of a community can produce functional shifts of varying magnitudes. For example, if the distribution of genes differs markedly between two communities (e.g., high vs. low redundancy), similar taxonomic composition perturbations to both communities may produce drastically different functional shifts. d To model the relationship between taxonomic perturbations and functional shifts in a given community, a taxa-function response curve is obtained by fitting a power function to a large array of measurements of functional shifts associated with many different taxonomic perturbations of varying magnitudes. e The taxa-function response curve can be decomposed into two factors: attenuation, which describes how quickly functional shifts increase in magnitude as taxonomic perturbations increase, and buffering, which indicates how well functional shifts are suppressed at smaller taxonomic perturbationsCC BY · PubMed Central
Fig. 2 Examples of taxonomic perturbations, their corresponding functional shifts, and the associated response curves. The taxonomic perturbation and functional shift magnitudes generated for a single vaginal (a) and a single gut (b) community. Each point represents a single perturbation. The red lines indicate the taxa-function response curve fit to these points. The response curves for all vaginal (c) and gut (d) communities are overlaid to compare general body site trends. Green and blue lines represent the mean response curve for all vaginal and gut communities respectively. The robustness factor distributions associated with these response curves are shown as violin plots with inlaid boxplots for attenuation (e) and buffering (f). The width of the violin plot indicates the density of robustness factor values, the middle of the box displays the median robustness factor value, the upper and lower edges of the box represent the 75th and 25th percentiles, respectively, and the whiskers extend to 1.5 times the inner quartile range (range between the 75th and 25th percentiles) past either end of the box. Outliers are shown as individual black circles
Fig. 2 Examples of taxonomic perturbations, their corresponding functional shifts, and the associated response curves. The taxonomic perturbation and functional shift magnitudes generated for a single vaginal (a) and a single gut (b) community. Each point represents a single perturbation. The red lines indicate the taxa-function response curve fit to these points. The response curves for all vaginal (c) and gut (d) communities are overlaid to compare general body site trends. Green and blue lines represent the mean response curve for all vaginal and gut communities respectively. The robustness factor distributions associated with these response curves are shown as violin plots with inlaid boxplots for attenuation (e) and buffering (f). The width of the violin plot indicates the density of robustness factor values, the middle of the box displays the median robustness factor value, the upper and lower edges of the box represent the 75th and 25th percentiles, respectively, and the whiskers extend to 1.5 times the inner quartile range (range between the 75th and 25th percentiles) past either end of the box. Outliers are shown as individual black circlesCC BY · PubMed Central
Fig. 3 Comparison of attenuation and buffering between environments and subsites. Points represent the median attenuation and buffering values of samples from the indicated subsite, with error bars indicating the 95% confidence interval of the median in both dimensions (see “Methods”). Colors are shared by subsites from the same environment, while shape indicates the specific subsite within the environment
Fig. 3 Comparison of attenuation and buffering between environments and subsites. Points represent the median attenuation and buffering values of samples from the indicated subsite, with error bars indicating the 95% confidence interval of the median in both dimensions (see “Methods”). Colors are shared by subsites from the same environment, while shape indicates the specific subsite within the environmentCC BY · PubMed Central
Fig. 4 Attenuation of individual functions. a The density of attenuation values for each KEGG superpathway across all communities, with dots indicating the median attenuation. b A scatterplot of the median attenuation of each function in skin communities vs. gut communities. The color of each point indicates the difference in log median relative abundance of that function between skin and gut communities. The gray line indicates the 1:1 relationship in median attenuation. Only differentially abundant functions are show (FDR < 0.01; Wilcoxon rank-sum test). c Each point shows a function’s median attenuation in a particular environment with error bars displaying the 95% confidence interval and color indicating the environment
Fig. 4 Attenuation of individual functions. a The density of attenuation values for each KEGG superpathway across all communities, with dots indicating the median attenuation. b A scatterplot of the median attenuation of each function in skin communities vs. gut communities. The color of each point indicates the difference in log median relative abundance of that function between skin and gut communities. The gray line indicates the 1:1 relationship in median attenuation. Only differentially abundant functions are show (FDR < 0.01; Wilcoxon rank-sum test). c Each point shows a function’s median attenuation in a particular environment with error bars displaying the 95% confidence interval and color indicating the environmentCC BY · PubMed Central
Fig. 5 Associations between gene distribution features (GDFs) and taxa-function robustness. a Pearson correlations between GDFs and attenuation; blue and red bars indicate significant (p < 0.01) positive and negative correlation coefficients, respectively, whereas gray bars show non-significant correlation coefficients. Each panel corresponds to correlations when considering communities from all environments (pooled) or the subset of communities from a particular environment. b Communities plotted by the first two principal components (PCs) determined by principle component analysis (PCA) of the five GDFs. The percent variance explained by each PC is indicated on the axis labels. Loadings for GDFs are indicated by the direction and magnitude of the labeled vectors. Communities were binned along both axes, and the density of attenuation values is displayed by the width of the violin plots in the plots along the top and right margins of the PCA plot. Dots indicate the median attenuation of communities in each bin
Fig. 5 Associations between gene distribution features (GDFs) and taxa-function robustness. a Pearson correlations between GDFs and attenuation; blue and red bars indicate significant (p < 0.01) positive and negative correlation coefficients, respectively, whereas gray bars show non-significant correlation coefficients. Each panel corresponds to correlations when considering communities from all environments (pooled) or the subset of communities from a particular environment. b Communities plotted by the first two principal components (PCs) determined by principle component analysis (PCA) of the five GDFs. The percent variance explained by each PC is indicated on the axis labels. Loadings for GDFs are indicated by the direction and magnitude of the labeled vectors. Communities were binned along both axes, and the density of attenuation values is displayed by the width of the violin plots in the plots along the top and right margins of the PCA plot. Dots indicate the median attenuation of communities in each binCC BY · PubMed Central
Fig. 6 Robustness and metagenome-based functional shift trends across subsites. a Distribution of simulation-based attenuation estimates across five subsites, ordered by median attenuation in each subsite. b Distribution of the log scale ratio of metagenome-based functional profile dissimilarity to taxonomic profile dissimilarity observed between pairs of communities from the same subsite across five subsites. c Distribution of attenuation estimates from mixing community compositions between pairs of communities from the same subsite across five subsites
Fig. 6 Robustness and metagenome-based functional shift trends across subsites. a Distribution of simulation-based attenuation estimates across five subsites, ordered by median attenuation in each subsite. b Distribution of the log scale ratio of metagenome-based functional profile dissimilarity to taxonomic profile dissimilarity observed between pairs of communities from the same subsite across five subsites. c Distribution of attenuation estimates from mixing community compositions between pairs of communities from the same subsite across five subsitesCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 62 references

Cited by

See all 92

Metadata

Authors
Eng, A., & Borenstein, E.
Affiliations
  • University of Washington (US)
  • Santa Fe Institute (US)
Corresponding author
Borenstein, Elhanan
Venue
Microbiome
Published
2018-03-02
Volume
6
Issue
1
Pages
45-45
ISSN
2049-2618 (ISSN-L)
OpenAlex
W2794489850
Cited by
89
References
62
Citations by year
2026: 10 · 2025: 10 · 2024: 8 · 2023: 7 · 2022: 14 · 2021: 22 · 2020: 8 · 2019: 6 · 2018: 4
FWCI
3.56
Type
journal article
Language
en
Open access
Licence
Locations
  • Microbiome (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Institutes of Health (DP2 AT007802-01)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gut microbiota and health
  • Physical Sciences › Environmental Science › Ecology › Microbial Community Ecology and Physiology
  • Physical Sciences › Environmental Science › Nature and Landscape Conservation › Ecology and Vegetation Dynamics Studies
Keywords
host-associated microbiome, gut microbiota, vaginal microbiota, body site differences, community composition, synthetic microbial communities
MeSH terms
Bacteria, Computer Simulation, Environment, Female, Humans, Phylogeny, RNA, Ribosomal, 16S, Vagina, Gastrointestinal Tract, Biodiversity, Metagenomics, Cellular Microenvironment, Microbiota

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Taxa-function robustness in microbial communities. https://app.cochat.ai/papers/10.1186/s40168-018-0425-4
The paper
Eng, A., & Borenstein, E. (2018). Taxa-function robustness in microbial communities. Microbiome. https://doi.org/10.1186/s40168-018-0425-4

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.