Different patients. Different microbes. Shared disease pathways. MANTRA reveals the functional signatures behind disease.
MANTRA is a validated multi-omic discovery engine built on that principle: species-level signatures rarely reproduce across cohorts, while the underlying metabolic pathways do. Tested head-to-head against standard bioinformatics tools across seven clinical and applied verticals, MANTRA has already outperformed them — and no non-invasive, India-validated tool currently exists for any of these indications.
What the microbiome actually is — and why it belongs in your differential
A bioinformatics background isn't required to use MANTRA's findings — the three principles below are sufficient context, and the foundation everything else on this site builds on.
It's a network, not a species list
Trillions of bacteria, viruses, fungi and archaea act as one interconnected network — not a checklist of "good" or "bad" organisms. Most tools that just count species miss this entirely.
It's dynamic — a lever, not a fixed trait
Reshaped continuously by diet, drugs and disease. Unlike a patient's genome, it's something you can act on — which is exactly why it matters for treatment decisions, not just diagnosis.
Function generalises. Species don't.
The specific bacteria in a patient's gut vary person to person, even with the same disease. Function is often more conserved than species composition — which is why MANTRA is built to score at that functional level directly, alongside the species-level picture rather than in place of it.
Function-first isn't a hunch. It's what the field's largest studies found.
Eight published results, from independent groups, that MANTRA's design rests on.
Your gut bacteria carry roughly 150 times more genes than you do.
The human gut microbial gene catalogue holds about 3.3 million genes. Much of the biochemistry happening in the gut is encoded there, not in the patient's own genome.
Qin et al., Nature 2010Which species you carry varies from person to person. What the community does stays remarkably stable.
Across hundreds of healthy people, the Human Microbiome Project found that microbial composition varied widely while the metabolic pathways performed stayed stable. This is the observation MANTRA is built around.
Human Microbiome Project Consortium, Nature 2012About half of the bacteria linked to a disease are linked to other diseases too.
Re-analysing 28 case–control studies across 10 diseases, researchers found roughly half of disease-associated genera respond to more than one condition: a shared response to being unwell, not a specific marker. It is one reason species-only signatures often fail to carry over between cohorts.
Duvallet et al., Nature Communications 2017Gut bacteria chemically modify most of the drugs we prescribe.
When 76 human gut bacterial strains were screened against 271 common oral drugs, 176 of those drugs were significantly depleted by at least one strain. Treatment response and toxicity are partly a microbial question.
Zimmermann et al., Nature 2019The same meal can send two people's blood sugar in opposite directions.
In a cohort of roughly 800 people, identical foods produced widely different glucose responses, and a model that included gut microbiome features predicted those responses better than standard carbohydrate counting.
Zeevi et al., Cell 2015Whether a cancer immunotherapy works can depend on what lives in the gut.
Patients who had recently taken antibiotics responded less well to PD-1 blockade, and the gut bacterium Akkermansia muciniphila was enriched in responders. Transferring responder microbiota into mice improved their response.
Routy et al., Science 2018Environment shapes the gut microbiome far more than genetics does.
In over 1,000 people, host genetics explained only a small share of microbiome variation, while diet, medication and lifestyle explained much more. That is why a microbiome-based readout carries information a genome test cannot.
Rothschild et al., Nature 2018In inflammatory bowel disease, the gut's chemistry shifts with disease activity.
Following people over time with metagenomics, metabolomics and more, the second Human Microbiome Project (HMP2) found that functional and metabolic changes track disease activity, beyond what species counts alone show.
Lloyd-Price et al., Nature 2019The institution that validates this first in Indian patients sets the evidentiary standard for it.
Every signature on this site has already cleared the field's toughest public benchmarks, and every one remains unvalidated in an Indian population. This is not a limitation so much as an opening: a clinical partner who leads that validation study becomes the originating site, with joint IP and a meaningful head start on the rest of the country.
Two ways in: by clinical function, or by therapeutic area
Diagnosis, prognosis, toxicity & more — organised by what you're trying to do
The same findings, organised by therapeutic area — IBD, oncology, nephrology & more
Peer-reviewed and published work from the team behind MANTRA
Who's building MANTRA, and what they bring from biology and AI
Maha is the vast: an ecosystem of trillions of cells. Anim is the minute: the single molecule that tips a patient toward disease. The microbiome is both at once, and MANTRA is built to read across that whole range, from the whole community down to the one pathway that matters.