How to Increase Gut Microbiome Diversity: The Dietary Evidence

What Microbiome Diversity Actually Means

When researchers discuss gut microbiome diversity, they typically refer to two distinct concepts: alpha diversity (diversity within a single individual's microbiome — typically measured by species richness, evenness, or the Shannon diversity index) and beta diversity (differences in microbiome composition between individuals). High alpha diversity — a greater number of different microbial species coexisting in balanced proportions — is consistently associated in epidemiological studies with better metabolic health, lower inflammatory markers, reduced risk of IBD and IBS, and healthier immune function.

The Shannon index is the most widely used diversity metric in gut microbiome research: it accounts for both the number of different species (richness) and the evenness of their distribution. A microbiome with 500 species where one species comprises 80% of the total has lower Shannon diversity than one with 200 species distributed more evenly. The healthiest microbiomes tend to show both high richness and high evenness.

The low diversity of Western industrialised microbiomes compared to traditional and hunter-gatherer communities is one of the most consistent findings in comparative microbiome research. The Sonnenburg lab's studies of Hadza hunter-gatherers in Tanzania showed 40% higher alpha diversity than age-matched Americans, with substantially different bacterial community structures — more plant-fiber-fermenting Firmicutes and Bacteroidetes, fewer opportunistic Proteobacteria. This diversity gap is attributed primarily to diet, antibiotic exposure, and mode of delivery/feeding in infancy.

The American Gut Project: 30 Plants Per Week

The American Gut Project, now part of the Earth Microbiome Project, is the largest community-sourced microbiome dataset ever assembled. Published in mSystems in 2018 by McDonald et al. (led by Rob Knight at UC San Diego), the study analysed 16S rRNA microbiome data from 11,336 participants across the US, UK, and Australia alongside detailed dietary questionnaires.

The most replicated finding: participants who consumed more than 30 different plant food types per week had significantly higher gut microbiome alpha diversity (measured by Faith's phylogenetic diversity and weighted UniFrac) compared to those consuming fewer than 10 plant types, regardless of whether they were omnivores, vegetarians, or vegans. The diversity advantage was present across all dietary categories and all demographics. The number of plant types mattered more than the quantity of any single plant food.

The 30-plants-per-week figure has since become a practical benchmark in nutritional science and clinical practice. The PREDICT 1 study (Spector, Gardner et al., involving 1,100 participants across the US and UK) subsequently confirmed this finding with deeper shotgun metagenomic sequencing, showing that dietary plant diversity predicted microbiome diversity more strongly than any other single dietary variable, including total fiber intake.

The Stanford Fermented Foods RCT: A Landmark Finding

The 2021 randomised controlled trial by Wastyk, Sonnenburg, Gardner and colleagues at Stanford, published in Cell, is the highest-quality human evidence directly testing dietary strategies for microbiome diversity and inflammation. The trial is notable for its rigorous design and mechanistically important findings.

36 healthy adults were randomised to a high-fermented-food diet or a high-fiber diet for 10 weeks, with a 4-week washout period. The fermented food group incrementally increased intake to an average of 6.3 servings per day of fermented foods (yoghurt, kefir, fermented cottage cheese, kimchi, sauerkraut, fermented vegetables, vegetable brine drinks, kombucha), providing an estimated 3 × 10¹⁰ CFU of live bacteria per day at peak consumption. The high-fiber group consumed 45.1g dietary fiber per day from diverse plant sources.

Results at 10 weeks: the fermented food group showed a significant increase in microbiome alpha diversity (Shannon index, p=0.016) and a decrease in 19 immunological proteins including interleukin-6, interleukin-10, interleukin-12p70, interferon-gamma, and monocyte chemoattractant protein-1. The high-fiber group showed no significant change in alpha diversity and no significant reduction in the inflammatory marker panel. Notably, a subset of high-fiber participants with higher baseline microbiome diversity did show anti-inflammatory responses to fiber — suggesting that the microbiome must have sufficient diversity to ferment fiber effectively before fiber increases produce anti-inflammatory benefits. This created an interesting cascade model: fermented foods improve diversity, which in turn may enable the anti-inflammatory benefits of higher fiber intake.

Polyphenols as Selective Prebiotics

Polyphenols — the diverse class of plant secondary metabolites including flavonoids, stilbenes, lignans, and tannins — reach the colon largely undigested (80–90% of ingested polyphenols are not absorbed in the small intestine) and are fermented by colonic bacteria. This fermentation is bidirectional: bacteria modify polyphenols into bioactive metabolites, while the polyphenols selectively feed specific bacterial taxa, acting as prebiotics.

The best-studied example is pomegranate ellagitannins (punicalagin) and ellagic acid: these are metabolised by intestinal bacteria — specifically Gordonibacter urolithinfaciens and Ellagibacter isourolithinifaciens — into urolithin A, a compound that activates mitophagy (selective autophagy of damaged mitochondria) in mammalian cells, extends lifespan in C. elegans, and improves muscle function in aging humans in clinical trials. Critically, only approximately 40% of people harbour the specific bacteria capable of producing urolithin A from dietary ellagitannins — illustrating how the clinical effect of a food depends on the individual's microbiome composition.

Similarly, resveratrol (from red grapes and berries) acts as a selective growth promoter for Akkermansia muciniphila — a keystone species in mucus layer homeostasis and metabolic health — in addition to its SIRT1-activating properties. Green tea catechins (particularly EGCG) promote Bifidobacterium and inhibit Clostridium perfringens. These relationships between specific polyphenols and specific bacterial taxa represent a rapidly developing area of personalised nutrition research.

Inulin-Type Fructans: The Most Evidence-Supported Prebiotic

Among specific prebiotic substrates, inulin-type fructans (ITF) — including inulin, fructooligosaccharides (FOS), and chicory-derived oligofructose — have the most extensive human clinical trial evidence for selectively increasing Bifidobacterium and Lactobacillus populations. ITFs are non-digestible dietary fibers composed of fructose chains with a terminal glucose unit, and are highly selectively fermented by beneficial bacteria in the colon.

Dosing context matters when reading that table. Clinical trials demonstrating reliable Bifidobacterium increases typically administered 5–15g of ITF daily — achievable with roughly 50–100g of Jerusalem artichoke or a generous serving of cooked garlic — and higher doses above 20g per day produce significant gas and bloating in individuals with dysbiosis or small intestinal bacterial overgrowth (SIBO). The table reflects raw inulin content, but cooking hydrolyzes a portion of the fructan chains, reducing the effective prebiotic dose delivered to the colon; this is relevant for foods like onion and leek that are rarely eaten raw. For people with existing gut symptoms, the practical approach is to start with the lower-inulin items on the list — asparagus, leek, onion — and increase intake gradually over 4–6 weeks, rather than beginning with chicory root or Jerusalem artichoke. Bloating experienced during this ramp-up reflects rapid fermentation of substrates by bacterial populations that were previously substrate-starved; it is not a harmful reaction. In most cases the discomfort resolves within 2–4 weeks as the microbiome adapts and the fermentative capacity of Bifidobacterium and Lactobacillus populations expands to match the new substrate load.

The Practical 30-Plants Protocol

Reaching 30+ plant types per week requires a structured approach for most people in Western dietary patterns. All plant food types count separately: each vegetable variety, fruit variety, whole grain variety, legume variety, nut type, seed type, herb, and spice is one distinct plant point. A mixed-seed sprinkle on a salad containing four different seeds counts as four plant types. A spice blend can add multiple points per meal. Dark chocolate (>70% cocoa) counts as a plant food and contributes polyphenols beneficial to Bifidobacterium.

Practical strategy: build a weekly plant food inventory, deliberately rotating through different varieties within food categories (swap white potato for sweet potato, then for cassava; rotate between lentils, chickpeas, black beans, and edamame; alternate between oats, rye, barley, and quinoa in grains). The benefit is ecological — diversity of dietary inputs drives diversity of microbiome. Even small amounts of new plant types introduce new prebiotic substrates and polyphenol profiles that support underrepresented bacterial taxa. The 30-plant target is a minimum threshold, not a ceiling — people averaging 50+ plant types per week show even higher diversity in the American Gut Project data.