Pith. sign in

REVIEW 3 major objections 4 minor 75 references

Mapping the bacterial ways of life

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Bacterial metabolic niche space is a branched, filamentous structure whose diffusion-map coordinates describe real ecological strategies, allowing ecosystem-level metabolic fingerprints.

desk verdict A useful bacterial metabolic atlas with a convincing top-variable interpretation, but the branching niche-space claim is unanchored by any null model and should be read as a hypothesis. read the letter →

arxiv 1908.07631 v2 pith:HHBLLJNG submitted 2019-08-20 q-bio.PE

Ashkaan K Fahimipour , Thilo Gross This is my paper · ORCID
classification q-bio.PE
keywords bacterialnichespacediffusionmapsmanifoldlearningmetabolicnetworksmicrobiomefunctionalfingerprinttrait-basedecologyreconstructiongeometry
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to show that the bacterial niche space—the set of metabolic strategies by which bacteria persist—is not a filled hypervolume or a set of disconnected clusters, but a branching, filamentous structure. Using a nearly parameter-free manifold-learning method on predicted metabolic reaction sets from 2,621 bacterial genera, the authors construct a functional coordinate system whose axes are nonlinear combinations of metabolic capabilities. The first axes separate recognizable lifestyles: photosynthesis in Cyanobacteria, host association with amino-acid dependence, soil versus marine generalism, and gut colonization. The authors then show that environmental microbiome censuses map to characteristic regions of this space, so ecosystem types can be described by the metabolic strategies their communities carry. A sympathetic reader would care because this offers a taxonomy-free, capability-based language for comparing microbes and their communities.

What carries the argument

The central object is the diffusion map, a manifold-learning procedure that treats the 7,769-dimensional presence-absence matrix of predicted metabolic traits (directed substrate-product pairs) as a weighted graph among the 2,621 genus-level genome representatives: it computes an affinity matrix from k-nearest neighbors, interprets it as a graph Laplacian, and takes its eigenvectors as new diffusion variables, ordered by importance. These variables are nonlinear composites of metabolic capabilities, and the paper reads the taxa at each variable's extremes to interpret the axis. A second element is the low-dimensional embedding of all diffusion variables, which visualizes the filamentous branching geometry, and a third is the mapping of environmental 16S censuses to the space by matching sequence variants to the extremal genomes at 97% similarity.

What would settle it

A test would be to run the same diffusion map on a set of genera whose metabolic capabilities are known from experiments or curated high-quality models and check whether the same filamentous branches and the variable-1 photosynthetic separation reappear; if the geometry collapses or the axes no longer track known physiology, the predicted niche space is an artifact of the reconstruction pipeline.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a diffusion map of genome-predicted metabolic networks yields a metabolic niche space with a complex branching geometry, where branches constitute major bacterial strategies rather than arbitrary statistical artifacts. Variable 1 cleanly separates photosynthetic Cyanobacteria, and enrichment analysis ties it to photorespiratory and carbon-fixation metabolites; variables 2, 3, and 4 trace continuous contrasts among host-associated gamma-proteobacteria, soil Actinobacteria, gut-associated Clostridia, and marine generalists; variables 8 and 10 flag reliance on host-derived amino acids in tiny-genome endosymbionts and parasites. Diffusion distances correlate only weakly with phylogenetic distance (matrix correlation, r = 0.273, P < 0.001), and a two-dimensional embedding reveals quasi one-dimensional branches emerging from a common core. Microbiome censuses from similar ecosystem types are enriched for the same extremal strategies, yielding ecosystem-level metabolic fingerprints.

Load-bearing premise

The whole map rests on computer-predicted reaction sets for a single representative genome per genus, not on measured biochemistry; if those predictions are systematically wrong, the branching geometry and the ecosystem fingerprints inherit the error.

Editorial extensions

If this is right

  • A bacterial genome can be summarized by its coordinates in diffusion space, so extremal coordinates translate directly into candidate ecological roles such as photosynthesis, host dependence, or marine generalism.
  • Microbiome censuses can be described by which extremal strategies they contain, allowing soil, freshwater, marine, and host-associated communities to be compared without reference to taxonomy.
  • The filamentous geometry implies that large regions of metabolic niche space are empty; those empty regions are either unsampled bacteria or metabolisms that cannot persist, guiding isolation and culturing efforts.
  • Diffusion distance becomes a proxy for ecological similarity that complements phylogenetic distance, which the paper finds to be only weakly correlated with metabolic strategy (r = 0.273).
  • The same manifold-learning approach can be applied to other trait sets, such as species-level functional profiles from metagenomic or metatranscriptomic data, to map realized niches.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the branching structure is real, one testable prediction is that evolution and adaptation tend to move populations along existing branches rather than across empty niche space; engineered or experimentally evolved strains should track the branches.
  • The extremal strategies the paper identifies could double as growth-condition predictions: for instance, taxa at the amino-acid-dependent variables 8 and 10 should require those amino acids in culture, a direct laboratory test.
  • Because diffusion variables are nonlinear composites, they may be more robust to single-gene noise than individual reaction presence, making them promising functional markers for metagenomic screening.
  • The paper maps the fundamental niche (encoded capabilities); applying the same coordinates to transcriptomic data could map the realized niche and reveal condition-dependent strategy shifts.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The authors map the metabolic niche space of bacteria by reconstructing genome-scale metabolic networks for 2,621 genus-level representative genomes using CarveMe, encoding each network as a 7,769-dimensional binary vector of substrate-product reactions, and applying diffusion-map manifold learning. They report several interpretable diffusion variables that separate, for example, photosynthetic Cyanobacteria, host-associated gamma-proteobacteria, and soil Actinobacteria, and they support these interpretations with metabolite-enrichment analyses. They also compute a Mantel correlation between diffusion distance and phylogenetic distance, visualize the diffusion space with a PHATE embedding, and use that embedding to argue that the bacterial metabolic niche space has a branching, filamentous geometry. Finally, they map 16S rRNA sequence variants from Earth Microbiome Project communities onto the extremal diffusion-variable strategies and show that different ecosystem types have characteristic metabolic fingerprints.

Significance. If the claims hold, this paper would provide a large-scale, reproducible functional coordinate system for bacterial metabolism, a set of falsifiable metabolite associations for individual niche dimensions, and a new geometric hypothesis about the structure of bacterial niche space. The study has genuine strengths: the diffusion map is constructed unsupervised, with no parameters fit to ecological labels; the data and code are publicly available; the enrichment examples connect specific variables to concrete biochemistry; and the authors are candid about the distinction between fundamental and realized niches. The central geometry claim, however, is currently supported only by visual inspection of one embedding, and the interpretability of the variables is validated on the same metabolic features used to construct the coordinates. Both issues need quantitative or independent support before the manuscript's main conclusions can be accepted.

major comments (3)
  1. [Results, 'The 2-dimensional embedding...' and Fig. 3C] The central claim that the metabolic niche space has a branching, filamentous geometry is asserted from visual inspection of a single PHATE embedding of the diffusion variables. PHATE is deliberately designed to expose trajectory-like and branching structure, so the appearance of filaments in its output is not, by itself, evidence that the underlying data have such geometry. No null model is provided (e.g., matched randomized trait matrices preserving row/column margins, or a unimodal continuous cloud), no branch-point stability or skeleton-extraction analysis is reported, and no quantitative branchiness measure is given. Because the input matrix is a sparse, strongly clade-structured binary matrix, k-nearest-neighbor chains could produce visually structured embeddings even in the absence of true branching niche structure. This is load-bearing because the abstract and discussion present the branching geometry as the main conceptual result. Please add a quantitative test that distinguishes real branch structure from embedding artifacts and report the PHATE parameters used for Fig. 3C and Fig. S1.
  2. [Methods, 'Identifying associated metabolites' and 'Mapping environmental samples to diffusion space'] The interpretability of the diffusion variables is validated using the same metabolic reaction features that were used to construct the diffusion coordinates: the enrichment analysis identifies metabolites that are overrepresented in the networks of taxa at the extremes of variables derived from those networks, and the EMP fingerprints are defined by extremal taxa selected from those same variables. These analyses demonstrate internal consistency but do not independently establish that the variables correspond to ecologically meaningful metabolic strategies. To support the paper's broader claims, an out-of-sample or independent validation is needed, for example using experimentally characterized strains, independent trait annotations, held-out genomes, or transcriptomic/metabolomic data.
  3. [Methods, 'Diffusion map procedure'] The stated eigenvalue ordering is inconsistent with the standard diffusion-maps literature cited by the authors. The text says that the first (most important) variable is the eigenvector corresponding to the smallest non-zero eigenvalue of the row-normalized Laplacian, followed by the second smallest, but diffusion-map coordinates are conventionally ordered by the largest non-trivial eigenvalues of the diffusion operator, with the trivial eigenvalue 1 excluded. The exact matrix being diagonalized and the eigenvalue ordering matter because the paper's numbering of variables (variable 1, variable 2, etc.) and the associated 'importance' claims depend on this choice. Please clarify whether the implementation uses a graph Laplacian eigenmap rather than a diffusion map, and correct the ordering description or the implementation accordingly.
minor comments (4)
  1. [Methods, 'Diffusion map procedure'] The claim that the results are insensitive to the choice of k is not supported by any sensitivity analysis shown in the main text or supplement; please provide a quantitative check, such as the stability of the leading variables or the enrichment results across several values of k.
  2. [Methods, 'Mapping environmental samples to diffusion space'] The pipeline is described as nearly parameter-free with only the choice of k, but it also involves the BLAST 97% sequence similarity threshold, the extremal-genome count (10), the number of diffusion variables used (50), and the GSEA FDR threshold (0.05). Please list these as explicit parameters and discuss their influence on the main results.
  3. [Results, 'Phylogenetic relatedness...' and Fig. 3A] The Mantel test statistic is reported as r = 0.273, but the type of correlation (Pearson vs. Spearman) and the number of permutations used are not stated; please specify these details in the Methods.
  4. [Methods, 'Mapping environmental samples to diffusion space'] The EMP mapping uses a single representative genome per genus and assigns a niche as occupied if any ASV in a sample has at least 97% similarity to any extremal genome; the sensitivity of the ecosystem fingerprints to this threshold and to the genus-level representation is not assessed.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the diffusion coordinates are unsupervised and the EMP fingerprints are out-of-sample; the missing null model for the PHATE branches is a robustness gap, not a circular step.

full rationale

The derivation chain is: CarveMe genome-scale reconstructions (Methods, 'Metabolic networks') produce a 2,621 x 7,769 binary reaction-availability matrix; a k-nearest-neighbor affinity matrix (k = 10) is built; the row-normalized Laplacian's eigenvectors define diffusion variables; variables are ordered by eigenvalue; and taxa are interpreted by enrichment along variable orderings. No habitat, ecosystem, or trait label is used in constructing the coordinates, so the 'characteristic regions' of EMP communities (Fig. 4) are an out-of-sample mapping rather than a fitted prediction: ecosystem types enter only when communities are matched to extremal genomes and then clustered. The enrichment (GSEA) step does use the same metabolic networks that generated the coordinates, but it is a descriptive post-hoc labeling of variables, not a prediction forced by construction. The authors' claim that variables 'correspond to meaningful metabolic strategies' is therefore an interpretation supported by internal consistency, not a circular derivation. The filamentous-geometry claim rests on visual inspection of a PHATE embedding with no matched-null comparison; that is a missing robustness test, not a circular step, because PHATE is applied to the already-computed diffusion variables and the observed geometry is a property of the data rather than an input. Self-citations (Barter & Gross 2019 for the diffusion algorithm; Nyberg et al. 2015 for the term 'localized variables') are methodological or terminological and are not load-bearing: the diffusion-map mathematics is externally established (Coifman & Lafon), and no uniqueness theorem is invoked to forbid alternative interpretations. Hence no circularity; the score of 1 reflects only the presence of minor non-load-bearing self-citations, which do not affect the central claim.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central result is a re-representation of CarveMe reaction-presence data. All ecological interpretation rests on reconstruction accuracy and on the chosen diffusion and community-mapping parameters.

free parameters (4)
  • k (number of nearest neighbors in diffusion map) = 10
    The only parameter of the diffusion map; the authors state results are insensitive to k but provide no quantitative sensitivity analysis.
  • BLAST sequence identity threshold for EMP mapping = 97%
    Used to match 16S variants to extremal genomes; changing this threshold changes niche occupancy calls.
  • Extremal genome count and number of diffusion variables = 10 genomes; first 50 variables
    Niche fingerprints use the 10 most extreme genomes per variable and the first 50 variables, producing 100 presence/absence strategies.
  • GSEA FDR threshold = 0.05
    Standard cutoff for retaining metabolites associated with variable extrema; a choice that affects the interpretation tables.
assumptions (5)
  • domain assumption Genome-encoded biochemical reactions represent feasible metabolic strategies and define the niche space.
    Introduction operationalizes the niche as the set of biochemical reactions encoded by genomes; this modeling choice underpins all downstream claims.
  • domain assumption The BiGG universal bacterial model and CarveMe reconstruction provide an accurate and sufficient summary of bacterial metabolic capabilities.
    Methods build all networks from CarveMe against BiGG; if the universal model lacks relevant metabolism, the input trait matrix is incomplete.
  • domain assumption Diffusion map eigenvectors are meaningful nonlinear composites of metabolic traits, and ordering taxa by extreme entries reveals ecological strategies.
    The authors acknowledge the procedure gives no interpretation and rely on post hoc enrichment to assign ecological meaning.
  • domain assumption 16S rRNA gene sequence similarity at 97% maps an environmental taxon to the same metabolic niche as its closest sequenced relative.
    Used to transfer extremal-genome strategies to EMP community censuses; gene-level similarity need not imply metabolic equivalence.
  • standard math Diffusion map theory as presented by Coifman-Lafon and Barter-Gross is correct.
    Background mathematical machinery for the Laplacian eigenvectors; not proved in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mapping the bacterial ways of life." pith.science (2026). https://pith.science/paper/HHBLLJNG

@misc{pith2026190807631,
  author       = {Pith},
  title        = {Pith review of: Mapping the bacterial ways of life},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HHBLLJNG}},
  note         = {Machine review of arXiv:1908.07631}
}
read the original abstract

The rise in the availability of bacterial genomes defines a need for synthesis: abstracting from individual taxa, to see larger patterns of bacterial lifestyles across systems. A key concept for such synthesis in ecology is the niche, the set of capabilities that enables a population's persistence and defines its impact on the environment. The set of possible niches forms the niche space, a conceptual space delineating ways in which persistence in a system is possible. Here we use manifold learning to map the space of metabolic networks representing thousands of bacterial genera. The results reveal a metabolic niche space with a complex branching geometry, whose branches constitute major strategies spanning life in different habitats and hosts. We further demonstrate that communities from similar ecosystem types map to characteristic regions of this new functional coordinate system, permitting ecological descriptions of microbiomes in terms of large scale metabolic roles that may be filled.

Figures

Figures reproduced from arXiv: 1908.07631 by the authors.

Figure 1
Figure 1. The diffusion map identifies variables describing discrete yes-or-no bacterial metabolic strate￾gies. Variable entries for each genome are visualized as colored tiles near the tips of a phylogenetic tree. Large negative or positive values (saturated reds and blues) indicate strong overlap with the focal strategy, whereas white indicates an absence of these capabilities. Circles are collapsed clades with near-zero en… view at source ↗
Figure 2
Figure 2. A broad spectrum of class-level capabilities indicated by variables 2, 3, and 4. A) Variable entries for each genome are shown as tiles near the tips of a phylogenetic tree. Darker red and blue tiles mark genomes receiving larger (in magnitude) negative and positive variable entries. B) The ordering of taxa defined by variable 2 entries, from negative to positive (left to right). The taxonomic compositions correspon… view at source ↗
Figure 3
Figure 3. Metabolic and phylogenetic similarities are roughly correlated. A) The relationship between distance in diffusion space, and cophenetic distance along branches of the phylogenetic tree between genome pairs. A large range of diffusion distances are observed for most given cophenetic distances. (B) Some variables such as 19 show similarities across the tree, with similar functional capabilities shared by remotely rela… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Bacterial communities across different ecosystem types (rows) in the Earth Microbiome Project [57] exhibit characteristic metabolic niche profiles. Columns correspond to different diffusion variable ex￾trema. Darker tiles indicate that a larger fraction of community ce…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 75 canonical work pages

  1. [1]

    Hutchinson, G. E. Cold spring harbor symposium on quantitative biology. Concluding Remarks 22, 415–427 (1957)

    work page 1957

  2. [2]

    Coexistence of species

    MacArthur, R. Coexistence of species. in ’challenging biological problems: Directions toward their solution’ (ed. ja behnke.) pp. 253–259 (1972)

    work page 1972

  3. [3]

    Chase, J. M. & Leibold, M. A. Ecological Niches: Linking Classical and Contemporary Approaches (University of Chicago Press, 2003)

    work page 2003

  4. [4]

    Holt, R. D. Bringing the hutchinsonian niche into the 21st century: ecological and evolutionary perspectives. Proceedings of the National Academy of Sciences 106, 19659–19665 (2009)

    work page 2009

  5. [5]

    O., Fitzgerald, D

    Winemiller, K. O., Fitzgerald, D. B., Bower, L. M. & Pianka, E. R. Functional traits, convergent evolution, and periodic tables of niches. Ecology Letters 18, 737–751 (2015)

    work page 2015

  6. [6]

    R., Vitt, L

    Pianka, E. R., Vitt, L. J., Pelegrin, N., Fitzgerald, D. B. & Winemiller, K. O. Toward a periodic table of niches, or exploring the lizard niche hypervolume.The American Naturalist 190, 601–616 (2017). Page 15

    work page 2017

  7. [7]

    & Enquist, B

    Blonder, B., Lamanna, C., Violle, C. & Enquist, B. J. The n-dimensional hypervolume. Global Ecology and Biogeography 23, 595–609 (2014)

    work page 2014

  8. [8]

    Hoogenboom, M. O. & Connolly, S. R. Defining fundamental niche dimensions of corals: synergistic effects of colony size, light, and flow. Ecology 90, 767–780 (2009)

    work page 2009

Show all 75 references
  1. [9]

    Porter, W. P. & Kearney, M. Size, shape, and the thermal niche of endotherms. Proceedings of the National Academy of Sciences 106, 19666–19672 (2009)

    2009

  2. [10]

    J., Godoy, O

    Kraft, N. J., Godoy, O. & Levine, J. M. Plant functional traits and the multidimensional nature of species coexistence. Proceedings of the National Academy of Sciences 112, 797–802 (2015)

    2015

  3. [11]

    Hypervolume concepts in niche-and trait-based ecology

    Blonder, B. Hypervolume concepts in niche-and trait-based ecology. Ecography 41, 1441–1455 (2018)

    2018

  4. [12]

    Stevenson, B. G. The hutchinsonian niche: multivariate statistical analysis of dung beetle niches.The Coleopterists’ Bulletin246–249 (1982)

    1982

  5. [13]

    J., Davies, R

    Inward, D. J., Davies, R. G., Pergande, C., Denham, A. J. & Vogler, A. P. Local and regional ecological morphology of dung beetle assemblages across four biogeographic regions.Journal of Biogeography 38, 1668–1682 (2011)

    2011

  6. [14]

    The global spectrum of plant form and function

    D ´ıaz, S.et al. The global spectrum of plant form and function. Nature 529, 167–171 (2016)

    2016

  7. [15]

    L., Bohannan, B

    Green, J. L., Bohannan, B. J. & Whitaker, R. J. Microbial biogeography: from taxonomy to traits. science 320, 1039–1043 (2008)

    2008

  8. [16]

    Fierer, N., Bradford, M. A. & Jackson, R. B. Toward an ecological classification of soil bacteria. Ecology 88, 1354–1364 (2007)

    2007

  9. [17]

    Horner-Devine, M. C. & Bohannan, B. J. Phylogenetic clustering and overdispersion in bacterial communities. Ecology 87, S100–S108 (2006)

    2006

  10. [18]

    T., Aanderud, Z

    Lennon, J. T., Aanderud, Z. T., Lehmkuhl, B. & Schoolmaster Jr, D. R. Mapping the niche space of soil microorganisms using taxonomy and traits. Ecology 93, 1867–1879 (2012)

    2012

  11. [19]

    K., Mora, T

    Fisher, C. K., Mora, T. & Walczak, A. M. Variable habitat conditions drive species covariation in the human microbiota. PLoS computational biology 13 (2017)

    2017

  12. [20]

    Prosser, J. I. et al. The role of ecological theory in microbial ecology. Nature Reviews Microbiology 5, 384–392 (2007). Page 16

    2007

  13. [21]

    Borenstein, E., Kupiec, M., Feldman, M. W. & Ruppin, E. Large-scale reconstruction and phylogenetic analysis of metabolic environments. Proceedings of the National Academy of Sciences 105, 14482– 14487 (2008)

    2008

  14. [22]

    Humphries, M. M. & McCann, K. S. Metabolic ecology. The Journal of Animal Ecology 83, 7–19 (2014)

    2014

  15. [23]

    Chase, J. M. Ecological niche theory. The Theory of Ecology 93–107 (2011)

    2011

  16. [24]

    & Ostling, A

    D’Andrea, R. & Ostling, A. Challenges in linking trait patterns to niche differentiation. Oikos 125, 1369–1385 (2016)

    2016

  17. [25]

    & Gross, T

    Barter, E. & Gross, T. Manifold cities: social variables of urban areas in the uk. Proceedings of the Royal Society A 475, 20180615 (2019)

    2019

  18. [26]

    Coifman, R. R. et al. Geometric diffusions as a tool for harmonic analysis and structure definition of data: Diffusion maps. Proceedings of the National Academy of Sciences 102, 7426–7431 (2005)

    2005

  19. [27]

    Coifman, R. R. & Lafon, S. Diffusion maps. Applied and Computational Harmonic Analysis 21, 5–30 (2006)

    2006

  20. [28]

    & Patil, K

    Machado, D., Andrejev, S., Tramontano, M. & Patil, K. R. Fast automated reconstruction of genome- scale metabolic models for microbial species and communities.Nucleic Acids Research46, 7542–7553 (2018)

    2018

  21. [29]

    D., Tatusova, T

    Pruitt, K. D., Tatusova, T. & Maglott, D. R. Ncbi reference sequences (refseq): a curated non- redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Research 35, D61– D65 (2006)

    2006

  22. [30]

    Mendes-Soares, H., Mundy, M., Soares, L. M. & Chia, N. Mminte: an application for predicting metabolic interactions among the microbial species in a community. BMC Bioinformatics 17, 343 (2016)

    2016

  23. [31]

    Eisenhut, M. et al. The photorespiratory glycolate metabolism is essential for cyanobacteria and might have been conveyed endosymbiontically to plants.Proceedings of the National Academy of Sciences 105, 17199–17204 (2008)

    2008

  24. [32]

    & Forchhammer, K

    Watzer, B. & Forchhammer, K. Cyanophycin synthesis optimizes nitrogen utilization in the unicellular cyanobacterium synechocystis sp. strain pcc 6803. Appl. Environ. Microbiol. 84, e01298–18 (2018). Page 17

    2018

  25. [33]

    E., Borel, F

    Fieulaine, S., Lunn, J. E., Borel, F. & Ferrer, J.-L. The structure of a cyanobacterial sucrose- phosphatase reveals the sugar tongs that release free sucrose in the cell.The Plant Cell 17, 2049–2058 (2005)

    2005

  26. [34]

    & Bassler, K

    Nyberg, A., Gross, T. & Bassler, K. E. Mesoscopic structures and the laplacian spectra of random geometric graphs. Journal of Complex Networks 3, 543–551 (2015)

    2015

  27. [35]

    & Yamada, Y

    Komagata, K., Iino, T. & Yamada, Y. The family acetobacteraceae. The Prokaryotes: Alphapro- teobacteria and Betaproteobacteria 3–78 (2014)

    2014

  28. [36]

    Meadows, J. A. & Wargo, M. J. Carnitine in bacterial physiology and metabolism.Microbiology 161, 1161 (2015)

    2015

  29. [37]

    & M ¨uller, H

    K ¨ampfer, P., Meyer, S. & M ¨uller, H. E. Characterization of buttiauxella and kluyvera species by analysis of whole cell fatty acid patterns. Systematic and Applied Microbiology 20, 566–571 (1997)

    1997

  30. [38]

    Parsons, J. B. & Rock, C. O. Bacterial lipids: metabolism and membrane homeostasis. Progress in lipid research 52, 249–276 (2013)

    2013

  31. [39]

    Foster, D. B. et al. Phosphatidylethanolamine recognition promotes enteropathogenic e. coli and en- terohemorrhagic e. coli host cell attachment. Microbial pathogenesis 27, 289–301 (1999)

    1999

  32. [40]

    & Boos, W

    Mayer, C. & Boos, W. Hexose/pentose and hexitol/pentitol metabolism. EcoSal Plus 1 (2005)

    2005

  33. [41]

    Reimer, L. C. et al. Bac dive in 2019: bacterial phenotypic data for high-throughput biodiversity analysis. Nucleic acids research 47, D631–D636 (2019)

    2019

  34. [42]

    Kaur, D., Brennan, P. J. & Crick, D. C. Decaprenyl diphosphate synthesis in mycobacterium tubercu- losis. Journal of bacteriology 186, 7564–7570 (2004)

    2004

  35. [43]

    L., Buchmeier, N

    Newton, G. L., Buchmeier, N. & Fahey, R. C. Biosynthesis and functions of mycothiol, the unique protective thiol of actinobacteria. Microbiol. Mol. Biol. Rev. 72, 471–494 (2008)

    2008

  36. [44]

    & Tan, X

    Wei, Y., Zhu, X., Zhang, S. & Tan, X. Structural and functional insights into corrinoid iron-sulfur protein from human pathogen clostridium difficile. Journal of inorganic biochemistry 170, 26–33 (2017)

    2017

  37. [45]

    & DuPont, H

    Darkoh, C., Plants-Paris, K., Bishoff, D. & DuPont, H. L. Clostridium difficile modulates the gut mi- crobiota by inducing the production of indole, an interkingdom signaling and antimicrobial molecule. mSystems 4, e00346–18 (2019)

    2019

  38. [46]

    & Moran, M

    Luo, H. & Moran, M. A. How do divergent ecological strategies emerge among marine bacterioplank- ton lineages? Trends in Microbiology 23, 577–584 (2015). Page 18

    2015

  39. [47]

    & Goto, S

    Kanehisa, M. & Goto, S. Kegg: kyoto encyclopedia of genes and genomes. Nucleic acids research 28, 27–30 (2000)

    2000

  40. [48]

    A., Kiyota, E

    Neshich, I. A., Kiyota, E. & Arruda, P. Genome-wide analysis of lysine catabolism in bacteria reveals new connections with osmotic stress resistance. The ISME journal 7, 2400–2410 (2013)

    2013

  41. [49]

    Chang, H.-H. et al. Complete genome sequence of âĂIJcandidatus sulcia muelleriâĂİ ml, an obligate nutritional symbiont of maize leafhopper (dalbulus maidis).Genome Announc. 3, e01483–14 (2015)

    2015

  42. [50]

    & Gil, R

    L ´opez-Madrigal, S., Latorre, A., Moya, A. & Gil, R. The link between independent acquisition of intracellular gamma-endosymbionts and concerted evolution in tremblaya princeps. Frontiers in Mi- crobiology 6, 642 (2015)

    2015

  43. [51]

    & Moran, N

    Dale, C. & Moran, N. A. Molecular interactions between bacterial symbionts and their hosts. Cell 126, 453–465 (2006)

    2006

  44. [52]

    & Coifman, R

    Nadler, B., Lafon, S., Kevrekidis, I. & Coifman, R. R. Diffusion maps, spectral clustering and eigen- functions of fokker-planck operators. InAdvances in neural information processing systems, 955–962 (2006)

    2006

  45. [53]

    Langille, M. G. et al. Predictive functional profiling of microbial communities using 16s rrna marker gene sequences. Nature biotechnology 31, 814 (2013)

    2013

  46. [54]

    Louca, S. et al. Function and functional redundancy in microbial systems.Nature Ecology & Evolution 2, 936 (2018)

    2018

  47. [55]

    Douglas, G. M. et al. Picrust2: An improved and extensible approach for metagenome inference. BioRxiv 672295 (2019)

    2019

  48. [56]

    Moon, K. R. et al. Visualizing structure and transitions in high-dimensional biological data. Nature Biotechnology 37, 1482–1492 (2019)

    2019

  49. [57]

    Thompson, L. R. et al. A communal catalogue reveals earthâĂŹs multiscale microbial diversity. Nature 551, 457 (2017)

    2017

  50. [58]

    & Knight, R

    Lozupone, C. & Knight, R. Unifrac: a new phylogenetic method for comparing microbial communi- ties. Appl. Environ. Microbiol. 71, 8228–8235 (2005)

    2005

  51. [59]

    Franzosa, E. A. et al. Species-level functional profiling of metagenomes and metatranscriptomes. Nature methods 15, 962–968 (2018)

    2018

  52. [60]

    & Ishida, K.-i

    Shiratori, T., Suzuki, S., Kakizawa, Y. & Ishida, K.-i. Phagocytosis-like cell engulfment by a planc- tomycete bacterium. Nature Communications 10, 1–11 (2019). Page 19

    2019

  53. [61]

    King, Z. A. et al. Bigg models: A platform for integrating, standardizing and sharing genome-scale models. Nucleic Acids Research 44, D515–D522 (2015)

    2015

  54. [62]

    Lee, M. D. Gtotree: a user-friendly workflow for phylogenomics. Bioinformatics 1, 3 (2019)

    2019

  55. [63]

    Hug, L. A. et al. A new view of the tree of life.Nature Microbiology 1, 16048 (2016)

    2016

  56. [64]

    Eddy, S. R. Accelerated profile hmm searches. PLoS Computational Biology 7, e1002195 (2011)

    2011

  57. [65]

    Edgar, R. C. Muscle: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32, 1792–1797 (2004)

    2004

  58. [66]

    Capella-Guti ´errez, S., Silla-Mart´ınez, J. M. & Gabald´on, T. trimal: a tool for automated alignment trimming in large-scale phylogenetic analyses.Bioinformatics 25, 1972–1973 (2009)

    2009

  59. [67]

    N., Dehal, P

    Price, M. N., Dehal, P. S. & Arkin, A. P. Fasttree 2–approximately maximum-likelihood trees for large alignments. PloS One 5, e9490 (2010)

    2010

  60. [68]

    & Bork, P

    Letunic, I. & Bork, P. Interactive tree of life (itol) v4: recent updates and new developments.Nucleic Acids Research (2019)

    2019

  61. [69]

    Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences 102, 15545– 15550 (2005)

    2005

  62. [70]

    R: A Language and Environment for Statistical Computing

    R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria (2019). URL https://www.R-project.org/

    2019

  63. [71]

    F., Gish, W., Miller, W., Myers, E

    Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. Journal of Molecular Biology 215, 403–410 (1990)

    1990

  64. [72]

    Wright, E. S. Using decipher v2.0 to analyze big biological sequence data in r. R Journal 8 (2016)

    2016

  65. [73]

    Ward Jr, J. H. Hierarchical grouping to optimize an objective function. Journal of the American statistical association 58, 236–244 (1963)

    1963

  66. [74]

    Maaten, L. v. d. & Hinton, G. Visualizing data using t-sne. Journal of machine learning research 9, 2579–2605 (2008)

    2008

  67. [75]

    An algorithm for fast preranked gene set enrichment analysis using cumulative statistic calculation

    Sergushichev, A. An algorithm for fast preranked gene set enrichment analysis using cumulative statistic calculation. BioRxiv 060012 (2016). Page 20 Acknowledgments We thank Jonathan A. Eisen and James P. O’Dwyer, for comments. A.K.F and T.G. were supported by the Dept. of Com...

    2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.