Gut microbiome test results explained: diversity, phyla and what the scores mean

A gut microbiome test reads the bacterial DNA in a small stool sample and tells you which bacteria were in that sample and what share of the total each one made up, a measure called relative abundance. ¹ It also gives a diversity figure, and many companies add scores of their own. The percentages describe one sample on one day, there are no agreed healthy ranges for them, and a 2025 international consensus statement found too little evidence to recommend these tests in routine care. ¹ Read the report as a description of a community, then take any symptoms to a clinician, because the stool tests that diagnose infection, inflammation and bleeding are different tests.

What is a gut microbiome test?

Your large intestine holds a dense community of bacteria, along with yeasts and viruses, and that community helps break down fibre, trains the immune system and helps keep out germs that cause disease. ⁸ A gut microbiome test samples that community through stool. Consumer kits ask you to swab or scoop a small amount into a tube of preserving liquid and post it back. The lab pulls out the DNA, sequences it, and matches the sequences against a database of known bacteria to build a list of who was there. ¹

Searches for an at-home gut microbiome test, a gut microbiome test kit and a stool test for gut microbiome all describe this same process. Some clinicians also order microbiome sequencing, but most tests are bought directly by consumers. The 2025 consensus statement, written by an international panel of gut specialists and microbiologists, discourages patients from ordering one without a clinical recommendation, "as there is little evidence for the applicability of gut microbiome testing in clinical practice". ¹ It also advises keeping your usual diet and prescribed treatments unchanged before collecting a sample, unless your clinician advises otherwise. ¹

16S sequencing vs shotgun sequencing

Most report differences start with the sequencing method. 16S rRNA sequencing (also called amplicon sequencing) copies and reads a single bacterial gene. It is cheaper and needs less sample, but it usually identifies bacteria only to the genus level and says nothing direct about what genes they carry. Shotgun, or whole-genome, sequencing reads all the DNA in the sample, can reach species and strain level, and can list the genes present, at a higher cost. ¹ Both are accepted ways to profile the gut community. ¹

Relative abundance: why every number is a share

Sequencing counts DNA reads, not bacteria, and the report divides each group's reads by the total. So if Bacteroides makes up 30% of your report, that is 30% of the DNA read from that sample, not a number of cells. The consensus statement points out that, unlike most health biomarkers, these quantities are relative, asks labs to present them as percentages so they are not mistaken for absolute amounts, and states there is not enough information to set strict healthy reference ranges for them. ¹ That is why a microbiome report rarely has the high and low flags you see on a blood panel; when it does, the company has drawn them from its own comparison group. For how reference ranges are built on ordinary lab tests, see reference ranges explained.

Why one stool sample is a snapshot

The gut community responds quickly to what you eat. In a 2014 study in Nature, switching volunteers to a diet made entirely of animal products or entirely of plant products changed the community's structure within a short stretch: the animal diet raised bile-tolerant bacteria such as Bacteroides and lowered Firmicutes that ferment plant fibre, including Roseburia, Eubacterium rectale and Ruminococcus bromii. ⁷ Antibiotics, age, illness and travel shift it too. ⁸ Sampling and lab handling add their own noise: in the NIST evaluation, one company's three reports on identical reference stool disagreed so much that it labelled two of them healthy and the third unhealthy. ² The consensus panel suggests that testing at more than one time point may help in specific clinical situations. ¹ A single result is best read as one frame of a moving picture; see why trends beat single readings.

A phylum is a very broad branch of the bacterial family tree, roughly like grouping animals into mammals, birds and fish. Most reports open with a bar of four to six phyla. In 2021 the official names changed, so newer reports may use the new name beside, or instead of, the old one: Firmicutes became Bacillota, Bacteroidetes became Bacteroidota, Proteobacteria became Pseudomonadota, Actinobacteria became Actinomycetota and Verrucomicrobia became Verrucomicrobiota. ¹² The bacteria are the same; only the labels moved.

Firmicutes (Bacillota)

Firmicutes is one of the two giant phyla of the gut and contains more than 200 genera, including Lactobacillus, Clostridium and Ruminococcus. ⁸ It also holds the main butyrate producers, the bacteria that turn fibre into a fatty acid important to the health of the colon: the two most abundant groups are related to Eubacterium rectale and Roseburia and to Faecalibacterium prausnitzii. ¹⁰ Because the phylum mixes so many genera with different habits, a high or low Firmicutes share says little by itself. Early animal studies linked more Firmicutes to obesity, but human results have been contradictory. ³ Firmicutes shares shift quickly with diet, and the genus-level detail below the phylum is more informative. ⁷

Bacteroidetes (Bacteroidota)

Bacteroidetes is the other giant phylum, dominated in the gut by Bacteroides and Prevotella. ⁸ Which of the two leads tends to follow long-term eating patterns: Bacteroides-dominant communities are common with Western diets higher in fat and animal protein, and Prevotella-dominant ones with high-fibre, plant-heavy diets such as those of the Hadza hunter-gatherers of Tanzania. ⁸ The consensus panel uses Bacteroidetes to show why phylum-level labels mislead: a high Bacteroides share "can both mean a healthy Bacteroides-high community, an altered ecosystem, and a Prevotella-dominant ecosystem." ¹ A high or low Bacteroidetes share is not linked to a specific health problem on its own.

Proteobacteria (Pseudomonadota)

Proteobacteria is normally a small part of the gut community. ³ It is a huge and varied phylum that includes many familiar disease-causing bacteria, such as Salmonella, Shigella and some strains of Escherichia coli, alongside harmless residents. ⁹ Reviews describe an expanded Proteobacteria share as a possible signature of a disturbed gut, with the strongest evidence in metabolic disorders and inflammatory bowel disease, and note that inflammation is the common thread. ⁹ A higher Proteobacteria share on a consumer report is still not a diagnosis: 16S sequencing cannot tell a harmless E. coli from a harmful one, and the share also rises after antibiotics and with age. ¹⁸ If you have diarrhea, fever or blood in the stool, the right test is a stool infection test ordered by a clinician, not a repeat microbiome kit. See Escherichia / Shigella.

Actinobacteria (Actinomycetota)

In the gut, the Actinobacteria phylum is a small, subdominant group represented mainly by Bifidobacterium. ⁸ Its share is highest in breastfed babies, whose Bifidobacterium ferment sugars from breast milk, and tends to fall in older age; macrolide antibiotics such as clarithromycin lower it too. ⁸ Diet and geography matter as well: Actinobacteria are nearly absent in some hunter-gatherer populations. ⁸ A low Actinobacteria share in an adult is common and does not by itself point to a problem.

Verrucomicrobia (Verrucomicrobiota)

Verrucomicrobia is another small, subdominant phylum. ³ On a gut report it almost always means one bacterium: Akkermansia muciniphila, a mucus-eating species first described in 2004 as a new member of this phylum. ¹¹ Akkermansia has been found to fall in obese mice, which is one reason reports highlight it, and the consensus panel lists it among the health-relevant bacteria worth reporting despite scarce evidence of cause and effect. ¹⁸ A zero or very low Verrucomicrobia share is common and is not a diagnosis.

The Firmicutes/Bacteroidetes ratio and why it fell out of favour

The Firmicutes/Bacteroidetes ratio, or F/B ratio, is the Firmicutes share divided by the Bacteroidetes share on the same report. ⁴ To calculate it, divide one percentage by the other: Firmicutes at 60% and Bacteroidetes at 30% gives a ratio of 2.0. The ratio became famous after studies in mice found obese animals had relatively more Firmicutes and fewer Bacteroidetes than lean ones, and it was widely cited as a hallmark of obesity. ³⁴

Human data did not hold up. A 2016 pooled analysis of 10 human studies found no significant association between obesity and either the Bacteroidetes/Firmicutes ratio or the individual shares of the two phyla. ⁴ A 2020 review in Nutrients blamed the contradictory findings on differences in how samples were processed and sequenced and on poorly described study volunteers, including their diets and lifestyles. ³ Its conclusion is the clearest summary of where the ratio stands:

It is currently difficult to associate the Firmicutes/Bacteroidetes ratio with a determined health status and more specifically to consider it as a hallmark of obesity.Magne F, Gotteland M, et al., Nutrients, 2020 ³

In 2025 the international consensus panel went further, agreeing with 86% support that "the reporting of Firmicutes-to-Bacteroidetes ratio in the microbiome testing is discouraged." ¹ So what is a good Firmicutes to Bacteroidetes ratio? There is no validated target, high or low. Searches for how to increase or lower the ratio rest on the same outdated idea; diet does move it, but no ratio has been shown to be healthier than another. ¹⁷

Diversity: alpha diversity, Shannon index and beta diversity

Diversity is the most established number on a microbiome report. Alpha diversity describes the variety within your one sample, combining richness (how many different kinds of bacteria were found) and evenness (how evenly the sample is shared between them); the consensus panel recommends that every test reports it. ¹ The Shannon index is the most common alpha diversity score: it rises when there are more kinds of bacteria and when no single group dominates. A Shannon value only makes sense against the company's own scale and comparison group, because sequencing depth and method change it. ²

Beta diversity is different: it measures how unlike one sample is from another, for example your stool compared with a healthy comparison group, or your stool this month compared with last month. ¹ The NIST study used one such measure, Bray-Curtis dissimilarity, to show that the same stool analysed by different companies looked as different as stools from different people. ²

What high or low diversity has been linked to

Lower diversity has been associated in research with obesity and other conditions, and it falls after antibiotic courses. ⁴⁸ The size of the difference is often small. Across 10 obesity studies, Shannon diversity was only 2.07% lower in people with obesity, and only one study had enough participants to detect even a 5% difference. ⁴ Higher diversity is generally read as a sign of a more varied, resilient community, but there is no single optimal profile. One review puts it plainly: "there is not a unique optimal gut microbiota composition since it is different for each individual." ⁸ A low diversity score is a reason to look at context, such as recent antibiotics, illness or a narrow diet, not a diagnosis.

How to increase gut microbiome diversity: what diet studies found

Two findings come up most often. In the American Gut Project, a citizen-science study that had sequenced more than 15,000 samples from more than 11,000 people by 2017, the number of different plant types people reported eating was associated with microbial diversity, more than labels such as vegan or omnivore. ⁵ People eating more than 30 types of plants a week were compared with those eating 10 or fewer, and the higher-variety group carried more bacteria that ferment fibre into short-chain fatty acids, including Faecalibacterium prausnitzii, and fewer of some antibiotic resistance genes. ⁵ The data were self-reported and observational, so they show an association, not proof that adding plants raises diversity.

The second is a randomised trial. In a 2021 study published in Cell, 36 healthy adults followed either a high-fibre or a high-fermented-food diet for 17 weeks. The fermented-food diet steadily increased microbiome diversity and lowered several inflammation markers; the high-fibre diet changed which fibre-digesting enzymes the bacteria carried while diversity stayed stable over the study. ⁶ It is a small trial, and it measured diversity and immune markers, not disease outcomes.

Why vendor scores are proprietary and unvalidated

Many reports add their own scores: an overall gut health number, an inflammation score, a butyrate or short-chain fatty acid score, sometimes a list of good and bad bacteria. Each company chooses which bacteria feed each score, how to weight them and which comparison group defines normal. The NIST team found that comparison groups vary between companies, from in-house healthy cohorts to public datasets such as the Human Microbiome Project or American Gut. ² The consensus panel found "insufficient evidence to include any dysbiosis index" in a report, discouraged estimating a sample's metabolic potential from its list of bacteria, and discouraged testing companies from giving treatment advice. ¹ There are no regulator-approved clinical microbiome tests in the US. ²

Inflammation balance score

An inflammation balance score, sometimes called an inflammatory or pro- versus anti-inflammatory balance, typically compares bacteria linked in research with inflammation, such as an expanded Proteobacteria share, against bacteria linked with a calmer gut, such as butyrate producers. ⁹¹⁰ It is a model built on associations, and no version has been validated as a measure of inflammation. ¹ A low inflammation balance score does not mean your gut is inflamed, and a good score does not rule inflammation out. Gut inflammation is measured directly with fecal calprotectin or fecal lactoferrin, proteins from white blood cells that leak into stool when the bowel lining is inflamed; blood markers are covered in inflammation markers explained.

Butyrate potential score

Butyrate is a short-chain fatty acid that gut bacteria make by fermenting fibre, and butyrate-producing bacteria have an important role in the health of the colon. ¹⁰ A butyrate potential score estimates how much butyrate your community could make, either from how many known butyrate producers were found, such as Faecalibacterium prausnitzii, Roseburia and Eubacterium, or, on shotgun tests, from butyrate-making genes. ¹¹⁰ Either way it measures potential, not the butyrate your gut produced, and the consensus panel discourages inferring metabolic potential from a list of bacteria. ¹ Populations of F. prausnitzii are reported to be lower in Crohn's disease, and Roseburia relatives are particularly sensitive to diet, which helps explain why this score swings between samples. ¹⁰

What a gut microbiome test cannot do

A microbiome report cannot diagnose irritable bowel syndrome, inflammatory bowel disease (Crohn's disease and ulcerative colitis), an infection, celiac disease, food intolerance or colon cancer. ¹ The consensus panel states there is "insufficient evidence to widely recommend the routine use of microbiome testing in clinical practice". ¹ The tests clinicians use instead look for something specific:

IBS is diagnosed from belly pain linked with changes in bowel habits, at least weekly over the past three months, and in most cases no test is needed beyond those that rule out other problems. ¹³ Anemia, rectal bleeding, bloody or black tarry stools, or weight loss suggest something other than IBS and need a clinician's prompt attention. ¹³ Some wellness reports also include stool markers such as secretory IgA, zonulin or stool pH; each has its own page and its own limits.

Every genus page in the LifeFrom library

Below the phyla, most reports list genera, the level where 16S sequencing usually stops. ¹ Each of these has its own page covering what a higher or lower share has been linked to and how much it varies.

What to ask your clinician

Microbiome results rarely change medical care on their own, and the consensus panel states that any change in management based on them should be made by your own clinician. ¹ If you bring a report to an appointment, these questions keep the conversation useful:

• Do my symptoms need a test that looks for infection, inflammation or bleeding, such as a stool infection test, calprotectin or FIT?
• Could a recent antibiotic course, illness or diet change explain my low diversity or unusual shares?
• Is any genus flagged on my report a known disease-causing germ, or a normal resident that sequencing cannot tell apart?
• Does anything on this report change what you would do, or is it background?
• If I want to change my diet for a gut condition, can you refer me to a registered dietitian?

For a wider view of reading any result, see how to read any lab result, and for all gut tests in one place, the gut topic hub. This page is general education, not medical advice, and it cannot interpret your report for you. Your clinician, who knows your history and symptoms, is the right person to do that.

Free in LifeFromAdd the lab results you already have, and LifeFrom reads each marker next to the others it moves with, in plain language, with the sources it used. Free, on the web or in the app.

Start on the web Download the app