BGPT: Paper Review: Global chemical effects of the microbiome include new bile-acid conjugations

Fuel Your Discoveries

Quick Explanation Copied

Paper review (skeptical, evidence-based)

This Nature study reports that microbiota presence (vs germ-free) globally reshapes host metabolomes across organs, with a notable discovery: three previously uncharacterized amino-acid–amide conjugates of cholic acid (phenylalanocholic, tyrosocholic, and leucocholic acid) that are microbiome-dependent, detectable in humans, enriched in inflammatory bowel disease and cystic fibrosis, and activate human FXR with downstream suppression of bile-acid synthesis gene expression in mice.

Long Explanation

Global chemical effects of the microbiome include new bile-acid conjugations

DOI: 10.1038/s41586-020-2047-9 ; Published: 26 Feb 2020

Figure-style schematic (what they did what they found)

System-wide profiling: Untargeted LC-MS/MS metabolomics across 29 mouse organ sites, with microbiome profiling by 16S rRNA gene sequencing, comparing germ-free (GF) vs specific-pathogen-free (SPF) states.
Chemical signatures: Microbiota-derived chemical signatures appear in many organs, with strong gut contributions.
New bile-acid conjugations: Three novel amino-acid amide conjugates of cholic acid (Phe-chol, Tyr-chol, Leu-chol) were microbiome-dependent and validated with synthesized standards.
Human detection + disease enrichment: The conjugates were detected in human-associated MS datasets and enriched in inflammatory bowel disease or cystic fibrosis cohorts in the paper’s analyses.
Mechanistic hook via FXR: Phe-chol and Tyr-chol act as human FXR agonists in vitro; in mice gavaged with the compounds, FXR target gene expression changes and bile-acid synthesis gene expression is reduced in a time-dependent manner.

Key potency values (from the paper’s reported FXR assays)

EC50/EC50-like values are taken directly from the paper’s FXR reporter description.

Phe-chol spatial abundance across GI subsections (mouse)

Values shown are the paper’s reported mean and SD for Phe-chol in SPF mouse GI sections.

FXR downstream liver bile-acid synthesis gene suppression (reported relative levels)

The plot uses the paper’s explicitly stated liver transcript relative levels for Cyp7a1 and Cyp8b1 at 24 h and 72 h.

Network-graph of claimed causal chain (evidence type annotated)

Nodes represent biological assertions in the paper; edges are evidence types: (A) GF vs SPF detection; (B) human dataset mining; (C) producer culture/labeling; (D) receptor activation + in vivo gene regulation.

Skeptical critique (what is strong vs what remains uncertain)

Strengths

Multi-organ sampling + microbiome perturbation logic: GF vs SPF design supports the claim that observed metabolite differences reflect microbiota presence rather than host genetics, and mapping across organ systems supports the “global chemical effects” premise.
Chemical identification workflow: The paper does not rely solely on spectral clustering; it describes tandem MS analysis for the candidate nodes and validation with synthesized standards with retention time and MS/MS matching.
Human relevance uses multiple angles: (i) cross-dataset spectral mining (MASST on GNPS datasets), (ii) pediatric CF cohort prevalence, and (iii) IBD cohort reanalysis; this triangulates detection beyond the mouse system.
Function is tested at receptor level and downstream physiology: in vitro human FXR luciferase assays and in vivo gavage with qPCR readouts.

Key uncertainties / failure modes

Microbial enzyme identity remains unresolved: the paper explicitly says that homologues of BAAT exist in clostridial genomes but the microbial enzyme remains unknown.
Endogenous physiology vs exogenous gavage: the FXR gene-expression changes after gavage show biological plausibility but do not by themselves prove endogenous causal roles in humans; the paper states that further studies are required to confirm physiological roles and whether these compounds contribute to gut diseases associated with dysbiosis.
Alternative explanation for FXR signaling: the paper notes that the possibility that observed in vivo effects are due to FXR agonism with release of cholate from amide conjugate hydrolysis cannot be excluded.
Public dataset heterogeneity: human detection depends on the availability and comparability of MS datasets (GNPS cohorts, American Gut Project, etc.). The paper uses MASST prevalence ranges, but cross-study confounding (diet, medications, antibiotics, disease stage) is always a risk when interpreting prevalence enrichment.

Data availability & computational traceability (reviewer lens)

Component	Where described / available	Why it matters
Mouse metabolomics (GF vs SPF)	GNPS under MassIVE IDs listed in the paper’s Data availability section	Enables independent reprocessing and compound reassignment checks
Mouse microbiome sequencing	Qiita study ID 10801 and ENA accession ERP109688	Supports replicating microbiome-attribution steps
Cross-dataset searching	MASST accessible at masst.ucsd.edu; MASST code/workflow linked in paper	Lets others reproduce spectrum matching / prevalence claims

What would disprove the main mechanistic story?

Missing conjugates in the microbiome-dependent comparison: if the reported amino-acid conjugates did not appear (or did not differ) in colonized vs germ-free mice under comparable analysis, the “microbiome-dependent reconjugation” claim would be undermined.
Loss of FXR signaling specificity: if FXR agonism and downstream gene regulation did not follow, or if the in vivo effects were fully attributable to released cholate rather than intact conjugates, the mechanistic link would be weakened.
Failure of human detection and enrichment: independent datasets (or re-analyses) failing to show detection/prevalence in humans and enrichment with the discussed diseases would argue against translational relevance.

Next-step BGPT actions

You can also run a fully independent science agent for iterative checks and deeper figure/number extraction from the deposited data, if needed.

Author review links (click through)

Feedback:

Updated: March 22, 2026