For about fifteen years now, the microbiome has been one of the most reliably exciting stories in biology. Trillions of bacteria, viruses, and fungi living in and on us, quietly steering our immune systems, our metabolism, even our moods. The headlines are everywhere; what has been much harder to find is the part where a doctor actually does something with all of it – orders a test, reads a result, prescribes a treatment, and a patient gets better because of what their microbes were or weren’t doing.
The microbiome in clinical practice, however, remains far less common than the headlines suggest.
That gap, between a field bursting with discovery and a field that changes clinical practice, is the subject of a recent review in Nature Medicine by Jack Gilbert and colleagues.¹ It is worth reading not because it announces a breakthrough, but because it takes an inventory of where the microbiome stands as a clinical tool, what has crossed the line into medicine, and what still remains in the lab.
What It Takes to Bring the Microbiome into Clinical Practice
The review organizes the whole question around a tidy idea, which is that for the microbiome to matter at the bedside, you usually need one of three things.
The first is a mechanism – knowing not just that some bacterium is associated with a disease, but how it does what it does. The second is a metric – a measurement reliable enough that a clinician can read it off a result and make a decision, the way they read a cholesterol or blood glucose panel. The third is a modifier – some way to actually change the microbiome on purpose and get a predictable result, whether that’s a diet, a drug, a probiotic, or a transplant.
It sounds almost obvious once it is spelled out like that. The reason it had to be spelling out is because the microbiome field has produced an enormous pile of associations (this bug goes with that disease), and associations, on their own, do pretty much nothing for a patient. An association tells where to look. Mechanisms, metrics, and modifiers tell what to do.
Where It’s Already Working
There is one clean success story, and it shows what “clinical translation” looks like when it actually lands.
For years, people with recurrent C. difficile infection (a brutal, relapsing gut infection that antibiotics often can’t shake), were treated, somewhat desperately, with fecal microbiota transplants. The logic was crude but sound: the patient’s gut ecosystem had been wrecked, so you replace it wholesale with a healthy donor’s. It worked surprisingly well. And over the last few years that crude procedure matured into something a regulator could approve: two FDA-licensed products, REBYOTA and VOWST, both for preventing recurrent C. diff. VOWST is even an oral capsule of purified bacterial spores rather than a raw stool preparation.
The review walks through this development as a kind of proof of concept, and the timeline is sobering. The science behind that capsule took more than a decade of trials to become a prescription. This is remarkably fast by drug development standards, but it is painfully slow for those hoping the microbiome revolution would already be in the medicine cabinet.
Beyond C. diff, the pipeline is busy- dozens of defined bacterial consortia, single strains, and engineered cocktails are in mid- to late-stage trials for everything from graft-versus-host disease to ulcerative colitis to cancer immunotherapy support. Even though most are not approved yet, the trend is real.
The Part Nobody Likes to Admit
To start, we still cannot confidently define a “healthy” microbiome. There is no reference range. What counts as a normal, healthy gut community in one population can look quite different in another, and the differences track with geography, age, diet, and ancestry rather than with health or sickness. A bacterium that’s a villain in one person can be perfectly benign in another. The review specifically points out how much of this is down to the strain-level, where two strains of the same species can do opposite things, and most of our measurements are too coarse to tell them apart.
Even the field’s favorite piece of advice comes in for some nuance. Fiber is good for you; everyone knows this. But the review notes that certain dietary fibers can actually drive inflammation in some patients with inflammatory bowel disease, and that some traditional, low-fiber diets have sustained populations for thousands of years. The lesson isn’t that fiber is bad. It’s that “good for the microbiome” is not a universal constant, and a clinical recommendation has to account for who, exactly, is sitting in the chair. Not all fiber types are created equal and are not equally good for everyone. For research studies, stool-based inflammation biomarkers can be combined with metabolomics to examine microbial function alongside intestinal inflammation.
Microbiome Composition vs. Function
Most microbiome measurements are a census. The DNA in a stool sample is sequenced to get a list of who’s present and in what proportions. That’s genuinely useful for discovery, but it has a built-in limitation: knowing which microbes are present tells very little about what they’re doing. The same species can sit quietly or work overtime depending on what they are being fed, who their neighbors are, and what the gut environment looks like that week. A list of names is not a list of activities.
The activities show up as chemistry. Microbes earn their influence on us by producing molecules – short-chain fatty acids that control inflammation, secondary bile acids that hold pathogens in check, neurotransmitter precursors, and thousands of compounds some of which we haven’t even discovered yet. Those molecules are the actual interface between the microbiome and the body, they are what metabolomics measures. For studies focused on host–microbiome signaling, bile acid analysis can quantify primary, secondary, and microbially conjugated bile acids.
This is the practical case for reading function, not just composition. A microbiome profile might find that the fiber-fermenting bacteria are present; a metabolomic readout is whether the short chain fatty acids those bacteria are supposed to produce are actually there. The review itself leans on exactly this kind of functional thinking. Composition alone is rarely something a clinician can act on, but the chemistry is the thing that’s closer to a decision.
The Future of the Microbiome in Clinical Practice
This story is currently being written. The summary is that the microbiome field is aspiring toward clinical utility and making real progress, but that broad, routine clinical use remains a work in progress. The barriers it lists are not minor: the lack of quantitative, reproducible metrics; the strain-level variation we can’t yet resolve; the geographic variability that breaks our reference ranges; and the simple fact that, outside a few clear cases, we still can’t reliably modify a microbiome to order.
What’s encouraging is the direction. The microbiome went from correlation to causation to a handful of approved therapies in about twenty years, faster than cancer immunotherapy or mRNA vaccines made the same journey from basic discovery to the clinic. The foundation has already been laid, microbial measurements are getting more quantitative and more functional, As the readouts shift from who is there toward what they are doing, the microbiome moves a little closer to being something a doctor can prescribe.
There’s one more piece to that future, and it’s a practical one. For functional readouts to matter clinically, they can’t stay locked inside a handful of specialized academic labs. The chemistry has to become measurable at scale – routinely, affordably, and for everyone. Semi-targeted metabolomics offers one practical approach by combining broad metabolic coverage with more reliable measurement of known compounds. Democratizing metabolomics, in other words, is not a side quest, it’s a step toward the microbiome becoming a real clinical tool. This is the problem Arome is focused on: making metabolomics broadly accessible, so that reading what microbes are actually doing becomes something within reach of everyday testing rather than a privilege of the research bench.
Microbiome therapy is not in the medicine cabinet yet. But it’s no longer only in the lab, either.
