3. Microbial target index

The Food Sources v1.5 book provides a per-food, in-depth description across its 196 chapters. This appendix offers the reverse view: it organizes the foods by microbial targets. A clinician, dietitian or donor-diet planner can thus get a quick answer to the question of which chapters contain the best foods for achieving a given target.

The appendix presents eight ranked lists by target. Only those foods are included for which the Food Sources v1.5 references contain at least moderate-strength evidence (★★). Entries at the ★ level (preclinical / hypothetical only) and empty cells are omitted — for the sake of credibility.

The ★ and empty rows are omitted. The full 196-chapter evidence matrix is detailed within the individual chapters of the book.

What does it target? Fructan, FOS, GOS, inulin prebiotics → selective increase in the Bifidobacterium proportion in the colon.

★★★ — Strong human RCT evidence:

★★ — Moderate evidence:

What does it target? Intake of live Lactobacillus / Lactiplantibacillus / lactic acid bacteria via fermented carriers.

★★★ — Strong human RCT evidence:

★★ — Moderate evidence:

What does it target? Polyphenols, ellagitannins, anthocyanins → support of the Akkermansia muciniphila mucin matrix.

★★★ — Strong human RCT evidence:

★★ — Moderate evidence:

What does it target? Roseburia, Faecalibacterium prausnitzii, Eubacterium rectale SCFA substrates.

★★★ — Strong human RCT evidence:

★★ — Moderate evidence:

What does it target? Bacteroidetes-mediated propionate production → GLP-1, PYY, appetite regulation, reduction of liver fat.

The human propionate-specific RCT evidence is more modest than for butyrate — therefore there is no ★★★ rating here. Propionate always appears in a butyrate + propionate matrix in human fermentation studies.

★★ — Moderate evidence:

What does it target? Dietary nitrate → Veillonella / Rothia / oral commensals nitrate reduction → nitrite → systemic NO → blood pressure, vascular function, athletic performance.

★★★ — Strong human RCT evidence:

★★ — Moderate evidence:

What does it indicate? Bilophila wadsworthia proliferates with high animal fat (especially saturated) + high taurine-conjugated bile-acid intake. A chronically high proportion is pro-inflammatory and colitis-associated. A risk requiring control, not an absolute contraindication.

⚠ — Risk category (each is high animal fat or high SFA):

What does it indicate? High fructan, GOS, raffinose, fructose or sorbitol is the gas-producing substrate of Methanobrevibacter smithii + Methanosphaera. In IBS- / SIBO-sensitive individuals it causes symptomatic worsening. It can be dosed (5—10 g/day start, titration), but control is required.

⚠ — Risk category (high FODMAP / fermentable substrate):

Clinical decision-making: Choose a target (e.g. "strong Akkermansia support"), look up the ★★★ list, and from there every food is ranked, with evidence. The contraindication matrix (Appendix B) can further filter for an individual patient.

Donor-diet planning: For weekly rotation, choose 2—3 foods from each target in the ★★★ group. Diversity supports not only taste but also the microbiome response.

Risk management: The ⚠ marking is not a ban but a question of dosing. In IBS / SIBO cases, titrate the items in list 8 starting at 5 g/day, gradually increasing over 2—4 weeks.

Updating with new evidence: The tables are dynamic — as new RCTs appear, the evidence levels are updated. They reflect the state of the v1.5 edition as of 2026-06-02.

The classification was carried out according to strict audit criteria developed during the internal review of Food Sources v1.5:

• 7a) Journal-attribution accuracy: PMC, Wikipedia and other indexing / popular sources are NOT considered peer-reviewed sources.
• 7b) Context match: the cited RCT applies to the food-specific form being examined (e.g. chicory inulin ≠ artichoke VLCI, garlic ≠ black garlic).
• 7c) Prohibition of non-peer-reviewed sources: only peer-reviewed journals, EFSA / EMA / Cochrane / WHO / FDA official documents may be included.
• 7d) Year consistency: the year in the text and the reference number must match.
• 7e) Meta vs RCT disambiguation: a meta-analysis cannot be cited as an individual RCT.

If a statement does not meet the above criteria, the food is omitted from the list or assigned a lower evidence level. In the full review of the 196 chapters, 33 genuine factual errors and 54 attribution inconsistencies were corrected for the v1.5 version (see `SCI-REVIEW-FIXES.md`).