MS1 · Four-Dimensional Evidence Matrix

How the Evidence Monitor scores metabolite-level evidence: each candidate metabolite node is rated on four dimensions — forward pathway, reverse pathway, coupling-verification depth and measurement depth.

Applies to: the gut microbiome co-metabolism channel (Evidence Monitor) only. The eight bioinformatics channels are graded by AI significance rating (High / Medium / Low) instead, and do not use this matrix.

1. The Compartmentalized Co-Metabolism Network Model

The model divides the porcine gut into three functionally distinct but interconnected compartments. The compartments are conceptual constructs, not strict anatomical divisions, acknowledging that biological parameters exist along continuous gradients. A molecule has different concentrations, functions, and dynamics in each compartment; interactions occur only at compartment interfaces.

Luminal Compartment

Site of microbial metabolism. SCFAs produced at millimolar concentrations (butyrate ~4–20 mM in porcine cecal and proximal colonic contents). Carrier-independent passive diffusion across the apical membrane. Timescale: minutes to hours.

Epithelial Compartment

Site of host epigenetic regulation. Intracellular butyrate concentrations ~3 orders of magnitude lower than luminal (micromolar range). MCT1-mediated active transport across the apical membrane; majority of imported butyrate is rapidly metabolically consumed. Timescale: hours to days.

Microenvironment Compartment

Site of feedback pathways. Host-secreted immune factors (IgA, defensins, cytokines), pH, oxygen concentration, and mucus layer thickness. Chemical and physical gradients regulated by both microbial metabolism and host epithelial activity. Timescale: days to weeks.

Core principle: The forward and reverse pathways are not independent causal chains but are connected through shared metabolites acting as interaction nodes bridging compartments at concentrations differing by three orders of magnitude and on timescales differing by days. This spatial and temporal asynchrony may explain inconsistent SCFA intervention effects in weaned piglets.

2. Four-Dimensional Matrix Rating System

All operational definitions are given below. Dimensions I–III are scored Level 0–4 (five levels); Dimension IV is scored Level 0–2 (three levels), reflecting the current technological ceiling.

DimensionLevelOperational Definition
I. Forward Pathway
(Microbe → Host)
0No study has tested whether microbial metabolites regulate host epigenetic marks in any model system.
1Association inferred from co-occurrence network analysis or statistical correlation in porcine gut metagenomic/metabolomic data.
2aIn vitro porcine model validation (e.g., porcine ileal organoids, IPEC-J2 cells) demonstrating that specific metabolites regulate epigenetic marks.
2bIn vivo porcine model validation (e.g., oral metabolite supplementation) demonstrating that metabolites regulate epigenetic marks in the porcine gut.
3Causal validation in murine models (e.g., mono-colonization, gene knockout) confirming the forward pathway causal chain.
4Causal validation in porcine models with independent replication by a separate laboratory.
II. Reverse Pathway
(Host → Microbiome)
0No study has tested whether host epigenetic status regulates the gut microbiota in any model system.
1Association inferred from host genetic–microbiome correlation studies, maternal nutritional programming studies, or germ-free pig transcriptome comparisons.
2aIn vitro porcine model validation demonstrating that host epigenetic changes alter microbial community composition or function.
2bIn vivo porcine model validation (e.g., HDAC inhibitor treatment, dietary methyl-donor manipulation) demonstrating that host epigenetic changes alter the gut microbiota.
3Causal validation in murine models (e.g., SIRT1, HDAC6, Tet2, or KDM5 knockout) confirming the reverse pathway causal chain.
4Causal validation in porcine models with independent replication by a separate laboratory.
III. Coupling Verification Depth0Neither the forward nor the reverse pathway has been verified for this node in any model system.
1Forward and reverse pathways have been independently verified, but not within the same experimental system.
2Forward and reverse effects have been simultaneously observed within the same experiment, but without interventional proof of causal connection between the two pathways.
3An intervention experiment demonstrates that blocking one pathway (e.g., forward) affects the effect of the other pathway (e.g., reverse), establishing causal coupling.
4Cross-scale coupling verified through simultaneous multi-compartment, multi-timepoint measurement within a single experiment, demonstrating that forward effects temporally precede reverse effects.
IV. Measurement Depth0Single compartment measured at a single timepoint; no simultaneous measurement of metabolites and epigenetic marks across multiple compartments.
1Multiple compartments measured OR multiple timepoints sampled.
2Multiple compartments measured AND multiple timepoints sampled within a single experimental system.

3. Research Priority Matrix & Falsifiable Predictions

PriorityNodeFeasibilityImportanceKey ExperimentTimeline
P1Butyrate (lumen–epithelium interface)HighHighIntracellular butyrate → HDAC → histone acetylation, with simultaneous luminal + epithelial + microenvironment sampling2–3 yr
P2Bile acids; Tryptophan metabolitesMediumHighFXR/TGR5 and AhR pathway validation with multi-compartment sampling—
P3Polyamines; Vitamin B12LowMediumCross-species extrapolation needed; porcine models absent at any level for polyamines and B12—
Falsifiable prediction (core): Intracellular butyrate concentration is a better predictor of host epigenetic changes than luminal butyrate concentration. Specifically: (a) intracellular butyrate concentration is predicted to correlate with H3K9 acetylation at r ≥ 0.7, while (b) luminal butyrate concentration is predicted to correlate with H3K9 acetylation at r ≤ 0.3. The core claim—that intracellular concentration is the better predictor—survives the failure of any specific numerical threshold; the thresholds are empirically informed benchmarks (Xiong et al. 2016 effect sizes) proposed as testable hypotheses.

4. Methods

Systematic evidence mapping following PRISMA-ScR guidelines. Primary search: PubMed (MEDLINE), May 20, 2026 (n = 12 records). Supplementary searches: CNKI and Wanfang Data (Chinese-language studies). Reference list checking (snowballing) identified 8 additional empirical studies and 6 framework/review articles. After deduplication, 26 studies were included in qualitative synthesis (20 empirical + 6 framework/review). Three reviewers with doctoral-level expertise in porcine gut microbiome or epigenetics independently completed the matrix rating.

5. Relationship to MS2

MS2 provides a complementary framework. MS1 operates at the metabolite-node level (diagnosing evidence distribution across a co-metabolism network), while MS2 operates at the strain level (assessing translational readiness of specific probiotic candidates). Together they form the methodological foundation for the Gut Microbiome Co-Metabolism Evidence Monitor channel on bioTriage-Open. The seven papers that founded this channel are listed in Foundation Papers.

6. Methodology References

The rating conventions are not ad hoc: each component is anchored to an established methodological standard, as mapped below.

Anchored to established standards. Dimension scoring follows the hierarchy-of-evidence tradition (Evans 2003; GRADE — Guyatt 2008, Balshem 2011); the forward/reverse pathway and coupling-verification dimensions follow causal-inference criteria (Hill 1965; Neville 2018, commensal Koch’s postulates); the evidence mapping follows PRISMA-ScR and PRISMA 2020 (Tricco 2018; Page 2021), with AMSTAR 2 for review quality (Shea 2017).