CoChat
Sign Up
Cited by 81Gut microbiota and healthGenomics and Phylogenetic StudiesOral microbiology and periodontitis research

Global genetic diversity of human gut microbiome species is related to geographic location and host health

Open access
Andreu-Sánchez, Sergio, Blanco-Míguez, Aitor, Wang, Daoming, Golzato, Davide, Manghi, Paolo, Heidrich, Vitor, Fackelmann, Gloria, Zhernakova, Daria V. +6 more·2025·Cell
DOI 10.1016/j.cell.2025.04.014
Open this paper in CoChat — free account

Abstract

The human gut harbors thousands of microbial species, each exhibiting significant inter-individual genetic variability. Although many studies have associated microbial relative abundances with human-health-related phenotypes, the substantial intraspecies genetic variability of gut microbes has not yet been comprehensively considered, limiting the potential of linking such genetic traits with host conditions. Here, we analyzed 32,152 metagenomes from 94 microbiome studies across the globe to investigate the human microbiome intraspecies genetic diversity. We reconstructed 583 species-specific phylogenies and linked them to geographic information and species' horizontal transmissibility. We identified 484 microbial-strain-level associations with 241 host phenotypes, encompassing human anthropometric factors, biochemical measurements, diseases, and lifestyle. We observed a higher prevalence of a Ruminococcus gnavus clade in nonagenarians correlated with distinct plasma bile acid profiles and a melanoma and prostate-cancer-associated Collinsella clade. Our large-scale intraspecies genetic analysis highlights the relevance of strain diversity as it relates to human health.

AI snapshot

Written from the paper’s full text.

TL;DR

Analyzing 32,152 gut metagenomes from 94 studies across 42 countries, this study reconstructed strain-level phylogenies for 583 microbial species and found that intraspecies genetic diversity is geographically stratified and linked to host phenotypes. It reports 484 global strain–phenotype associations, including a *Collinsella aerofaciens* clade enriched in melanoma and prostate cancer and a *Ruminococcus gnavus* clade enriched in older individuals.

Key findings

  • Geographic distance correlated with phylogenetic distance for 456 of 583 species-level genome bins (SGBs; Mantel test, FDR < 0.05), while species with higher transmissibility showed weaker geographic stratification (rPearson = 0.21, p = 3 × 10⁻⁷; among the 456 geographically associated SGBs, rPearson = 0.20, p = 8.85 × 10⁻⁶).
  • The study identified 484 global strain-level associations, comprising 389 unique associations between 54 phenotypes and 272 unique SGBs, with 80 associations reproduced in more than one continent.
  • A *Bifidobacterium longum* infant-specific clade was found in both European (elpd_diff = 288) and African (elpd_diff = 9.1) samples, separating *B. longum* subsp. *infantis* and *suis*-like clades.
  • A *Ruminococcus gnavus* clade was enriched in nonagenarians and associated with plasma cholic acid (estimate = 0.09, FDR = 1.39 × 10⁻⁴) and chenodeoxycholic acid (estimate = 0.1, FDR = 2.33 × 10⁻⁴); a *Collinsella aerofaciens* clade was enriched in melanoma (123/181 clade samples had melanoma; OR = 131.843, p < 2 × 10⁻¹⁶) and prostate cancer (12/14 participants in the clade; Fisher p = 4.97 × 10⁻¹⁴).

Methods

  • Strain-resolved metagenomics was performed with MetaPhlAn 4 and StrainPhlAn 4 on 32,152 gut metagenomes from 24,829 individuals in 94 studies across 42 countries; 1,660 SGB phylogenies were reconstructed, and 583 SGBs with at least 300 samples were retained for downstream analysis.
  • Geographic associations used Mantel tests between phylogenetic and geographic distances; transmissibility was estimated from strain-sharing events among same-country, different-study samples; phenotype associations used phylogenetic generalized linear mixed models (PGLMMs) via the anpan R package, with continent-stratified models.
  • Clade-specific functional analyses included metagenome-assembled genomes (MAGs), CAZyme annotation, bile-acid-related gene profiling, and MetaCyc pathway enrichment.

Limitations

  • The paper states the study is associative and observational, so it does not infer causal relationships; it focuses on the most abundant strain per SGB per sample and cannot profile strains below a certain coverage level; most phenotypes were absent from most cohorts, leading to overrepresentation of Dutch cohorts; large sample sizes are required for significant associations, limiting continent-stratified signals; isolation-by-distance analyses are noisy due to confounders such as travel, lifestyle, population mixture, and non-map distances; aging associations may be confounded by unmeasured behavioral and environmental differences.
  • Evident from the text: the sample set is heavily European (65%, with 35% Dutch), with much smaller representation from Africa (3.83%), Oceania and South America (0.5% combined), which limits global inference and likely contributes to the unequal number of associations detected across continents.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.cell.2025.04.014

Cites

146 of 146 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 20 cited sources checked · 0 supported · 0 flagged

  • Findingown data
    § SUMMARY · p. 2

    A Collinsella clade is associated with melanoma and prostate cancer.

    “and a melanoma and prostate-cancer-associated Collinsella clade.”
    Quote in the paperSubstantial claim
  • Findingown data
    § Figure 2. Geographic stratification of phylogenetic trees · p. 5

    The profiled strain of Bifidobacterium animalis was the same in all samples, independent of study and country.

    “We observed that the profiled strain was indeed the same in all samples, independent of study and country, which is in contrast to natural B. animalis strains in mice.”
    Quote in the paperSubstantial claim
  • Findingown data
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    Melanoma patients were enriched for the same C. aerofaciens clade in both European and North American samples.

    “Despite originating from different continents and studies, melanoma patients were en­ riched for the same clade in both European and North American samples.”
    Quote in the paperSubstantial claim
  • Findingown data
    § Phylogenetic association confirms three Bifidobacterium longum subspecies present in early life · p. 6

    For 14 SGBs, the age associations were driven by the presence of infant-specific clades.

    “For 14 SGBs, the age associations were driven by the presence of infant-specific clades.”
    Quote in the paperNot a substantial claim
  • Findingown data
    § Phylogenetic association confirms three Bifidobacterium longum subspecies present in early life · p. 8

    Two clades largely exclusive to infants were observed.

    “We also observed two clades largely exclusive to infants (Figure 3C).”
    Quote in the paperNot a substantial claim
  • May combine two claims. 30
  • Classified as probably not a substantial claim. 27

  • May combine two claims.

    Findingown data
    § SUMMARY · p. 2

    A higher prevalence of a Ruminococcus gnavus clade was observed in nonagenarians, correlated with distinct plasma bile acid profiles.

    “We observed a higher prevalence of a Ruminococcus gnavus clade in nonagenarians correlated with distinct plasma bile acid profiles”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § Intraspecies genetic distance correlates with the hosts' geographic distance · p. 5

    Continent-specific clades were evident in many SGBs, although European and North American samples tended to cluster together.

    “Continent-specific clades were evident in many SGBs, although European and North American samples tended to cluster together.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    CRC, hypertension, IBD, and melanoma exhibited associations across multiple continents.

    “Among these, CRC, hyperten­ sion, IBD, and melanoma exhibited associations across multiple continents.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § DISCUSSION · p. 12

    The majority of associations between subspecies' microbial variability and host anthropometrics were related to host age; BMI was also linked with subspecies diversity.

    “Although the majority of associations between subspecies' microbial variability and host anthropometrics were related to host age (Figures 3 and 4), BMI was also linked with subspe­ cies diversity.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § Phylogenetic association confirms three Bifidobacterium longum subspecies present in early life · p. 8

    The major infant-specific clade was characterized by multiple B. longum subsp. infantis references and included both European and African infants.

    “The major infant-specific clade was characterized by the presence of multiple B. longum subsp. infantis references and included both European and African infants (Figure 3C, in red).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    All individuals from PernigoniN_2021 had prostate cancer; 13 had either hormone-sensitive prostate cancer (four in the phylogeny, all also in the cancer-enriched clade) or castration-resistant prostate cancer (ten in the phylogeny, eight also in the cancer-enriched clade).

    “Notably, all individuals from this study had prostate cancer, 13 either hormone-sensitive prostate cancer (four in the phylogeny, all also found within the cancer-enriched clade) or castration-resistant prostate cancer (ten in the phylogeny, eight also within the cancer-enriched clade).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given
    § DISCUSSION · p. 12

    Subspecies microbial variation was linked not only to geography but also to a wide range of host characteristics.

    “Subspecies microbial variation was linked not only to geogra­ phy but also to a wide range of host characteristics.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    332 UniRef90 annotations were differentially enriched between the melanoma clade and the rest of the phylogeny, with 218 genes more frequently present in the melanoma-enriched clade.

    “From these reconstructed MAGs we identi­ fied 332 UniRef90 annotations as differentially enriched between the melanoma clade and the rest of the phylogeny, with 218 genes more frequently present in the melanoma-enriched clade (Table S5, 4).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § Phylogenetic association confirms three Bifidobacterium longum subspecies present in early life · p. 6

    Age was the phenotype most commonly associated with SGB intraspecies diversity (253/583 SGBs), followed by BMI (12/583) and sex (6/583).

    “Among anthropometric associations, age was the phenotype most commonly associated with SGB intraspecies diversity (253/583 SGBs), followed by BMI (12/583) and sex (6/583).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingreasoning
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    2 SGBs in Asian samples and 13 SGBs in European samples were linked to IBD.

    “2 SGBs in Asian samples and 13 SGBs in European samples were linked to IBD”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § Supplemental figures · p. 28

    The number of associations decreases when all continents are downsampled to equal numbers of samples.

    “(D) The number of associations decreases when all continents are downsampled to equal numbers of samples.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingown data
    § Species transmissibility and geographic effect · p. 22

    For SGBs without a previous threshold value, no longitudinal samples were available in the phylogenies, so the 3rd percentile was used as the threshold.

    “For the SGBs that did not have a previous threshold value, we observed that no longitudinal samples were available in the phylogenies and used the 3 rd percentile as the threshold.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data
    § Microbial phylogenetic tree reconstruction · p. 20

    A median of two samples were removed per phylogeny due to TreeShrink, with a maximum of 24 samples removed.

    “A median of two samples were removed per phylogeny due to TreeShrink, with a maximum of 24 samples being removed.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § Supplemental figures · p. 28

    The number of samples in each phenotype-SGB pair tested in the phylogenetic framework is different between continents.

    “(C) Number of samples in each phenotype-SGB pair tested on the phylogenetic framework is different between continents.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § Supplemental figures · p. 28

    The linear relationship between sample number and elpd_diff is not seen in associations with no signal.

    “This linear relationship is not seen in associations with no signal.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Phylogenetic relations between samples follow a Brownian motion model.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data
    § Microbial characteristics and geographic effect · p. 22

    The Lachnospiraceae family was highly enriched in geographically stratified species and might act as a confounder.

    “as this family was highly enriched in geographically stratified species and might act as a confounder.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § Geographic association with microbial phylogeny · p. 22

    Correlation estimates between phylogenetic and geographic distances using Pearson's and Spearman's coefficients were highly correlated (rPearson=0.93).

    “A comparison of the correlation estimates using the two different coefficients (Pearson's and Spearman's) showed high correlation (rPearson=0.93).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § Microbial phylogenetic tree reconstruction · p. 20

    For nine random species, a Mantel test (Pearson correlation) between phylogenies reconstructed with RAxML and IQTree showed high correspondence between distance metrics (0.72-0.99, mean of 0.93).

    “A Mantel test (Pearson correlation) between each showed high correspondence between distance metrics (0.72-0.99, mean of 0.93).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § Comparison of dominant strains of StrainPhlAn with GT-Pro · p. 21

    For all remaining 243 common taxa (median 1,316 samples), correlations between StrainPhlAn phylogenetic distances and GT-Pro core-genome SNV dissimilarities were significantly different from 0 (FDR<0.05), with a median correlation of 0.7.

    “For all the remaining 243 common taxa (median of 1,316 samples) we identified correlations significantly different than 0 for all taxa (FDR&lt;0.05), with a median correlation value of 0.7, suggesting a robust estimation of dominant strains [Table S2, 3].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Table S2, 3

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § Intraspecies genetic distance correlates with the hosts' geographic distance · p. 5

    Clostridiales is part of the phylum Bacillota, which was previously observed to be the top phylum showing co-diversification with human populations.

    “Interestingly, Clostridiales is part of the phylum Bacillota (formerly known as Firmicutes), which was previously observed to be the top phylum showing co-diversification with human populations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [15] Suzuki, 2022, Codiversification of gut microbiota with humans, Science source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation
    § INTRODUCTION · p. 2

    The advancement of computational tools for shotgun metagenomic sequencing has enabled profiling of microbes taxonomically and quantitatively and access to genetic information at subspecies and strain resolutions.

    “The advancement of computational tools for the analysis of shotgun metagenomic sequencing has enabled researchers to not only profile microbes taxonomically and quantitatively but also to access their genetic information at subspecies and strain resolutions.”
    Quote in the paperSubstantial claim0 of 4 sources checked
    • cites [9] Blanco Míguez, 2023, Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn, Nat. Biotechnol. source

      Not checked in the free run.

    • cites [10] Olm, 2021, inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains, Nat. Biotechnol. source

      Not checked in the free run.

    • cites [11] Van Rossum, 2022, metaSNV v2: detection of SNVs and subspecies in prokaryotic metagenomes, Bioinformatics source

      Not checked in the free run.

    • cites [12] Shi, 2022, Fast and accurate metagenotyping of the human gut microbiome with GT-Pro, Nat. Biotechnol. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisenone given
    § INTRODUCTION · p. 3

    The high resolution required for strain profiling can only be achieved for species with high coverage, and the substantial variability in microbiomes across individuals necessitates very large sample sizes for well-powered association studies.

    “This is partially due to the high resolution required for strain profiling, which can only be achieved for species with high coverage, and the substantial variability in microbiomes across individuals, which necessitates very large sample sizes for well-powered association studies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation
    § DISCUSSION · p. 12

    Kumbhari et al. describe a larger number of associations with a similar IBD sample size (20% of their associations), these associations are generally in small clades, and their top StrainPhlAn association agrees with one of the top associations described in our study, Fusicatenibacter saccharivorans.

    “Although Kumbhari et al. describe a larger number of associations with a similar IBD sample size (20% of their associations), these associations are generally in small clades, and their top StrainPhlAn association agrees with one of the top associations described in our study, Fusicatenibacter saccharivorans .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [30] Kumbhari, 2024, Discovery of disease-adapted bacterial lineages in inflammatory bowel diseases, Cell Host Microbe source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § DISCUSSION · p. 12

    In a recent association study between bacterial SNVs and human BMI, most of the SNVs that could be replicated between Israeli and Dutch populations (15/17 SNV) belonged to Faecalibacterium and F. prausnitzii.

    “In a recent association study between bacterial single-nucleotide variants (SNVs) and human BMI, the authors observed that most of the SNVs that could be replicated between Israeli and Dutch populations (15/17 SNV) belonged to Faecalibacte­ rium and F. prausnitzii . 23”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] Zahavi, 2023, Bacterial SNPs in the human gut microbiome associate with host BMI, Nat. Med. source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 11

    Based on the data investigated in this study, the presence of this clade does not appear to be related to clinical outcomes in melanoma or to testosterone production in prostate cancer.

    “Based on the data investigated in this study, the presence of this clade does not appear to be related to clinical outcomes in melanoma or to testosterone production in prostate cancer”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § Enrichment of a Ruminococcus gnavus clade in older Asian and European individuals · p. 8

    R. gnavus has previously been related to inflammation, but it was also associated with longevity in a Chinese study.

    “R. gnavus has previously been related to inflammation, 42 but it was also associated with longevity in a Chinese study. 43”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [42] Crost, 2023, Ruminococcus gnavus: friend or foe for human health, FEMS Microbiol. Rev. source

      Not checked in the free run.

    • cites [43] Wang, 2019, Enriched taxa were found among the gut microbiota of centenarians in East China, PLoS One

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given
    § Limitations of the study · p. 12

    Large sample sizes are required for identification of significant associations, which hampers the identification of continent-stratified signals.

    “Large sample sizes are required for identification of sig­ nificant associations, which hampers the identification of conti­ nent-stratified signals.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premiseown data
    § DISCUSSION · p. 12

    The authors report high inflation in the quantile-quantile (Q-Q) plots obtained from F. prausnitzii associations, suggesting tight linkage of significant genetic variations associated with BMI.

    “In the same vein, the authors report high inflation in the quantile-quantile (Q-Q) plots obtained from F. prausnitzii associations, suggesting tight linkage of sig­ nificant genetic variations associated with BMI.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] Zahavi, 2023, Bacterial SNPs in the human gut microbiome associate with host BMI, Nat. Med. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    Some bacteria could produce androgens, such as testosterone, which are associated with adverse cancer outcomes.

    “The original study indicated that some bacteria could produce androgens, such as testosterone, which are associated with adverse cancer outcomes. 54”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [54] Pernigoni, 2021, Commensal bacteria promote endocrine resistance in prostate cancer through androgen biosynthesis, Science source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § DISCUSSION · p. 11

    Multiple bacterial species' phylogenies resemble their human host populations' phylogenies, highlighting that geographical stratification in those species is unlikely to be driven by lifestyle alone.

    “Moreover, it has also been observed that multiple bac­ terial species' phylogenies resemble their human host popula­ tions' phylogenies, highlighting that geographical stratification in those species is unlikely to be driven by lifestyle alone. 15”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [15] Suzuki, 2022, Codiversification of gut microbiota with humans, Science source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductreasoning
    § Limitations of the study · p. 12

    This study is associative and observational by design and does not try to infer causal relationships between gut microbial subspecies and human health.

    “Because this study is associative and observational by design, it does not try to infer causal relationships between gut microbial subspecies and human health.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § A gut microbial intraspecies phylogenetic signal in melanoma and prostate cancer patients · p. 10

    CRC has often been associated with Fusobacterium species and clades, as have several other bacterial species.

    “It is worth noting that CRC has often been associated with Fusobac­ terium species and clades, as have several other bacterial spe­ cies, 31,47,48”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [31] Piccinno, 2025, The gut microbiome in colorectal cancer: a cross-stage strain-level pooled analysis of 3,741 individuals from 18 cohorts, Nat. Med. source

      Not checked in the free run.

    • cites [47] Zepeda-Rivera, 2024, A distinct Fusobacterium nucleatum clade dominates the colorectal cancer niche, Nature source

      Not checked in the free run.

    • cites [48] Borozan, 2022, Molecular and Pathology Features of Colorectal Tumors and Patient Outcomes Are Associated with Fusobacterium nucleatum and Its Subspecies animalis, Cancer Epidemiol. Biomarkers Prev. source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductreasoning
    § Limitations of the study · p. 12

    The analysis focuses on the most abundant strain for each SGB in each sample and cannot profile strains below a certain coverage level, which depends both on species abundance and total per-sample sequencing depth.

    “In exploring the trade-off between accuracy and profiling intra-sample and intersample comprehensiveness, our analysis focuses on the most abundant strain for each SGB in each sample and cannot profile strains below a certain coverage level, which depends both on species abundance and total per-sample sequencing depth.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data
    § Limitations of the study · p. 12

    Most phenotypes were absent from most cohorts, leading to an overrepresentation of associations found in Dutch cohorts.

    “In addition, most phenotypes were absent from most cohorts, leading to an overrepresentation of associations found in Dutch cohorts.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § Enrichment of a Ruminococcus gnavus clade in older Asian and European individuals · p. 8

    Previous research highlighted the involvement of R. gnavus in bile acid metabolism, particularly the production of iso-bile acids, which were elevated in Japanese centenarians.

    “Previous research also highlighted the involvement of R. gnavus in bile acid metabolism, 44 particularly the production of iso-bile acids, which were elevated in Japanese centenarians. 45”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [44] Devlin, 2015, A biosynthetic pathway for a prominent class of microbiota-derived bile acids, Nat. Chem. Biol. source

      Not checked in the free run.

    • cites [45] Sato, 2021, Novel bile acid biosynthetic pathways are enriched in the microbiome of centenarians, Nature source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given
    § Comparison of dominant strains of StrainPhlAn with GT-Pro · p. 21

    Species with fewer than 50 paired samples between distances were discarded before FDR estimation.

    “We then discarded species with fewer than 50 paired samples between distances and estimated a Benjamini-Hochberg (BH) FDR.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data
    § Microbial characteristics and geographic effect · p. 22

    For FDR<0.05 results, the family Lachnospiraceae was added as a covariate in a subsequent model because it was highly enriched in geographically stratified species and might act as a confounder.

    “For FDR &lt; 0.05 results, we further added whether the family of the SGB was Lachnospiraceae as a covariate in a subsequent model, as this family was highly enriched in geographically stratified species and might act as a confounder.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given
    § Microbial characteristics and geographic effect · p. 22

    Only annotations where the prediction of phypat and the phypat hypat + PGL classifiers matched were kept.

    “We kept only annotations where the prediction of phypat and the phypat hypat + PGL classifiers (including additional evolutionary information on phenotype gains and losses) matched.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning
    § Phylogenetic generalized mixed models in anpan · p. 22

    Associations where the diagnostic Pareto K value was ≥0.7 in at least 1% of samples were removed.

    “In addition, we removed associations where the diagnostic Pareto K value was ≥ 0.7 in at least 1% of the samples.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Phylogenetic relations between samples follow a Brownian motion model.

  • Classified as probably not a substantial claim.

    Study design and conductnone given
    § Functional analysis of MAGs · p. 25

    High-quality MAGs were selected as those with at least 90% completeness and maximum 5% contamination.

    “We then selected the high-quality MAGs, i.e., MAGs with at least 90% completeness and maximum 5% contamination.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning
    § Species transmissibility and geographic effect · p. 22

    To avoid possible confounders from same-study biases, species transmissibility rates were computed in samples from different studies.

    “To avoid possible confounders due to same-study biases (unaccounted-for social networks, contamination events, other technical biases), we computed these rates in samples from different studies.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Species transmissibility, computed as the fraction of same-strain pairs among samples from the same country but different studies, is a proxy for bacterial migration rates (dispersion).

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning
    § Limitations of the study · p. 12

    Some associations, such as the clade of R. gnavus found in older individuals, might be driven by other unknown processes for which the study does not account or measure.

    “Some of the associations we identified, such as the clade of R. gnavus found in older individuals, might be driven by other un­ known processes for which we do not account or which we do not measure.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data
    § Simulation analysis to assess model performance · p. 23

    In 200 simulations with no phylogenetic signal, the anpan phylogenetic model was never chosen over the null model, indicating that the model's regularization makes the false positive rate approach 0 as signal-to-noise ratio increases.

    “out of the 200 simulations, in none of them the phylogenetic model was chosen above the model with no phylogenetic term, indicating that the regularization of the model (where the phylogenetic term σ phylo is shrunk towards 0) makes the false positive rate asymptotically approaching 0 as the signal-to-noise ratio increases.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data
    § Phylogenetic generalized mixed models fit to data · p. 23

    For individual phylogeny-phenotype pairs, additional analyses controlled for specific covariates that might confound the phylogenetic signal and/or stratified per study.

    “For individual phylogeny-phenotype pairs, we performed additional analysis controlling for specific covariates that might confound the phylogenetic signal and/or stratifying per study.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Phylogenetic relations between samples follow a Brownian motion model.
44 of 44

  • The paper relies on this without support.

    Some bile-acid-related genes are usually found in operons.

    “Because some of these genes are usually found in operons, we also identified biosynthetic gene clusters related to metabolism.”

  • The authors state this as an assumption.

    Species transmissibility, computed as the fraction of same-strain pairs among samples from the same country but different studies, is a proxy for bacterial migration rates (dispersion).

    “This rate was used as a proxy for bacterial migration rates (dispersion).”

  • The authors state this as an assumption.

    Phylogenetic relations between samples follow a Brownian motion model.

    “assuming a Brownian motion”
3 of 3

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?

  • Geographical origin is a major driver of intraspecific diversity, shaping host traits from diet to antibiotic exposure.

    “Indeed, geographical origin is a major driver of intraspecific diversity, shaping host traits from diet to antibiotic exposure [94-97] .”

    From Genome-scale insights into the metabolic landscape and evolutionary development of Bifidobacterium bifidum (2026) · cites this paper as [97]

    The section ‘Intraspecies genetic distance correlates with the hosts’ geographic distance’ (p. 5) reports significant geographic–genetic associations for 456 of 583 gut microbial species groups. The Discussion (p. 11) treats geography as correlated with strain variation but leaves its causes open. These passages do not establish that geographic origin shapes host diet or antibiotic exposure. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

Self-check

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

Each contradiction between two passages listed here was confirmed by an independent check.

  • Summary says associations span 241 phenotypes, but results identify 54

    Major

    The summary describes the identified strain-level associations as encompassing 241 host phenotypes, whereas the results say the associations involved 54 different phenotypes. The larger number could describe phenotypes surveyed, but not phenotypes with identified associations as stated in the summary.

  • Association scores have the opposite sign from the methods' selection criterion

    Minor

    The methods define a supported phylogenetic association as having an elp_diff below −4, but the results report supported associations with positive elpd_diff values, including 288. As written, those examples cannot meet the stated criterion.

2 of 2

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Andreu-Sánchez, S., Blanco-Míguez, A., Wang, D., Golzato, D., Manghi, P., Heidrich, V., Fackelmann, G., Zhernakova, D. V., Kurilshikov, A., Valles-Colomer, M., Weersma, R. K., Zhernakova, A., Fu, J., & Segata, N.
Affiliations
  • University Medical Center Groningen (NL)
  • University of Trento (IT)
  • King's College London (GB)
Corresponding author
  • Fu, Jingyuan
  • Segata, Nicola
ORCID
Venue
Cell
Publisher
Elsevier BV
Published
2025-07
Volume
188
Issue
15
Pages
3942-3959.e9
ISSN
0092-8674 (ISSN-L)
DOI
10.1016/j.cell.2025.04.014
OpenAlex
W4410017394
Cited by
81
References
146
Citations by year
2026: 58 · 2025: 23
FWCI
24.01
Type
journal article
Language
en
Open access
Licence
Locations
  • Cell (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • Repositori digital de la UPF (Universitat Pompeu Fabra) (publishedVersion, cc-by-nc-nd)
  • Iris (University of Trento) (submittedVersion, cc-by-nc-nd)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (submittedVersion, cc-by-nc-nd)
Funders
  • Hartstichting (IN-CONTROL CVON 2012-03)
  • HORIZON EUROPE Framework Programme (101094099)
  • European Molecular Biology Organization (10263)
  • Ministerie van Onderwijs, Cultuur en Wetenschap (222.016)
  • European Commission (101152592)
  • European Research Council (MetaPG-716575)
  • National Cancer Institute
  • Horizon 2020 Framework Programme (825410, IHMCSA-964590)
  • HORIZON EUROPE European Research Council (ERC-CoG microTOUCH-101045015)
  • National Institutes of Health (1U01CA230551)
  • Nederlandse Organisatie voor Wetenschappelijk Onderzoek (VI.C.202.022)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gut microbiota and health
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Health Sciences › Dentistry › Periodontics › Oral microbiology and periodontitis research
Keywords
human gut microbiome, intraspecies genetic diversity, regional variation, host phenotype, Ruminococcus gnavus, Collinsella, bile acids, melanoma, prostate cancer, nonagenarians, horizontal transmission, metagenomics
MeSH terms
Eubacteriales, Gastrointestinal Microbiome, Bile Acids and Salts, Female, Humans, Male, Phylogeny, Genetic Variation, Metagenome

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

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