Six therapeutic areas. One engine. One underlying idea.
Every finding below comes from the same starting principle: gut bacteria aren't passengers, they're active chemists. They produce, modify and break down compounds that enter the bloodstream and act on organs far from the gut. Which specific bacteria are present varies enormously from person to person — but the chemistry they collectively perform is far more consistent, and that's the layer MANTRA is built to read. Each finding here is grouped by the therapeutic area it belongs to; for the same findings organised by clinical use case instead, see Applications.
Two conditions, one shared metabolic tell
Gut bacteria produce metabolites that either calm or inflame the gut lining. In UC and Crohn's, the specific bacteria present differ patient to patient, but the metabolic output — which enzymes are active, which byproducts accumulate — follows a more consistent pattern, which is what makes it possible to build a signature from it at all.
MANTRA distinguishes Ulcerative Colitis from Crohn's Disease from these signals alone (validated against HMP2/IBDMDB and Franzosa/PRISM), and separately flags disease complications earlier than endoscopy or imaging can, because metabolic shifts precede structural change. MANTRA also surfaced additional shared metabolic signals in public IBD data that earlier analyses of the same data had not reported.
Gut bacteria shape both how a tumour shows itself and how its treatment lands
A tumour reshapes its own neighbourhood before you can see it
Some gut bacteria can directly damage the colon lining or drive local inflammation that promotes tumour growth; others produce protective compounds that suppress it. As a tumour grows, it reshapes its local bacterial neighbourhood — which means that neighbourhood's fingerprint shows up in stool before the tumour is large enough to see on a colonoscopy or shed enough DNA to catch on a blood test.
MANTRA independently recovered a landmark study's own headline biomarkers, plus one organism nobody had flagged before — evidence the fingerprint is real and reproducible, not a statistical artefact of one dataset. That's the basis for a stool-based triage step ahead of ctDNA testing: cheap enough to run broadly, catching the people who need a colonoscopy while sparing everyone else an expensive liquid biopsy.
Bacteria can chemically reverse what the liver just deactivated
Gut bacteria carry enzymes that can chemically re-activate drugs after the liver has already deactivated them. A well-documented class of examples: the liver neutralises certain chemotherapy drugs, but a common bacterial enzyme (β-glucuronidase) can convert them back into their toxic active form inside the gut — a major driver of severe gut toxicity. Standard genetic tests only look at the human half of this equation (liver enzyme variants); they have no way to see the bacterial half. Blocking that bacterial enzyme protected mice from drug-induced toxicity (Wallace et al., Science 2010) — and it isn't an isolated case: of 271 common oral drugs screened against 76 gut bacterial strains, 176 were significantly depleted by at least one strain (Zimmermann et al., Nature 2019).
Combining a patient's genetics with their microbiome can flag chemotherapy toxicity risk before the first dose, using data standard panels don't capture. Gut bacteria also shape how much active drug reaches a tumour and how the immune system responds to dying tumour cells, which is why microbiome signatures may help separate likely responders from non-responders at the individual level, before treatment begins.
A single organism, one causal job, a clean on/off signal
A specific gut organism performs a metabolic function that reduces how much of a key stone-forming compound the body absorbs from the diet — whether or not a patient carries it is a causal, on/off signal for stone risk, not a loose correlation.
A point-of-care test for this organism — something most urology practices have no routine way to check for today — shows up to 70% lower recurrence when present. One in three kidney stone patients will form another one; right now there's no way to tell them which third they're in.
The gut has a direct pipeline straight to the liver
The gut and liver share a direct pipeline — the portal vein carries blood straight from the intestine to the liver before it reaches anywhere else in the body. Anything gut bacteria produce or leak gets delivered to the liver first, at concentration. Certain bacteria produce inflammatory fragments and altered bile acids that, delivered this way, drive fat accumulation and scarring in liver cells.
A causally-proven bacterial driver of disease progression powers a single stool signature that tracks the same disease from early fat buildup through fibrosis to cirrhosis — one non-invasive test at multiple decision points, instead of a separate liver biopsy at each stage.
The same meal, digested twice — once by you, once by your gut bacteria
When you eat, gut bacteria are digesting the same meal in parallel — fermenting fibre and starch into their own byproducts, which affect how fast and how high blood sugar rises afterward. That's a major reason two people can eat an identical meal and see very different blood sugar spikes, largely independent of their genetics.
During hard exercise, muscles produce lactate faster than they can immediately use it. Some gut bacteria can absorb and metabolise circulating lactate directly, acting as an extra clearance pathway alongside the liver and muscles — a trait that's been tied to specific bacterial genera in elite endurance athletes.
The same engine, unmodified, predicts individual glycemic response and identifies lactate-clearance signatures from gut composition alone — disease was just the first place it was pointed.
A promising area we haven't run through MANTRA yet
This one is well established in the published literature, though we haven't yet run it through MANTRA ourselves. Gut bacteria metabolise choline and L-carnitine — abundant in red meat and eggs — into trimethylamine (TMA), which the liver converts to TMAO, a compound repeatedly linked to atherosclerosis and cardiovascular risk in independent published cohorts. Bacterial metabolites more broadly are also implicated in blood pressure regulation and vascular inflammation.
We have no MANTRA-generated result here yet — this is an area we believe is promising based on the existing literature, not a validated application. We're open to exploring it with the right cohort and partner.
These six are where the evidence currently stands, not where it stops. We're open to exploring additional therapeutic areas with partners who bring the right cohort.