Why Stool Sequencing Cannot Tell You What Bacteria Do

Consumer microbiome tests report which bacteria a sample contains and what they are said to do. The gap between those two statements is where most of the interpretive difficulty sits.

The sample is an exit stream

Stool represents material that has completed its passage through the intestine. It is a reasonable proxy for the colon's contents but a poor one for the small intestine.

Communities differ substantially along the length of the gut, shaped by oxygen availability, transit speed and acidity. A single exit sample averages over those distinct environments.

The material also separates into organisms living freely in the interior and those attached to the mucus layer. Attached populations are underrepresented, and they are often the most relevant ones.

Sequencing detects genes, not activity

Most tests amplify a marker gene region that varies between bacterial groups, allowing identification without culturing anything. This is efficient but limited to the question of who is present.

Genetic material persists after cells die. A sequence in the sample does not establish that the organism was alive, let alone that it was metabolically active.

Establishing activity requires measuring gene expression or the chemical products themselves. These are different laboratory procedures, and a composition report does not include them.

Functional claims are usually inferred

Reports that assign functions to a community typically infer them, matching identified organisms against reference databases describing what related organisms can do.

Capability and behaviour are not the same. An organism carrying the genes for a pathway will only run it if the substrate is available and conditions favour it.

Reference databases also skew towards organisms that have been studied intensively, which are the ones easiest to grow in a laboratory. Less tractable organisms are described less completely.

Variability between samples is large

Composition changes with recent meals, transit time and where in the stool the sample was taken. A single specimen captures one point in a moving distribution.

This matters for anyone comparing two tests taken months apart. A difference may reflect natural fluctuation rather than any response to an intervening change.

Research designs handle this with repeated sampling and control groups. Individual testing rarely includes either, which weakens the conclusions that can be drawn from a personal result.

What the results can reasonably support

Composition data is genuinely useful for population research, where averaging across many people reveals patterns that no individual sample could show on its own.

For an individual, the reasonable reading is descriptive: this is roughly what was present at one moment. Treating it as a diagnosis extends it past what the method supports.

Digestive symptoms that persist are assessed clinically through history, examination and targeted testing. A sequencing report may add context but does not replace that process.