The Bacteria Your Probiotic Doesn't Contain
Six months in, Maree had stopped counting the supplements. The kefir in the morning, the multi-strain capsule at lunch, the sachets of inulin her gastroenterologist had waved off as "probably worth a try". Her gut still dictated the shape of her week: the cramping after rice, the bloating that arrived by 3pm, the careful mental map of every public toilet between home and the office. Maree is a composite of dozens of patients I have spoken to in clinic visits, in workshops, and over the years as a science writer. Her story matters here because a new Cambridge-led meta-analysis of more than 11,000 gut microbiomes points to a possible reason that none of those products helped her: the bacteria most consistently linked to a healthy gut may be the ones the supplement industry has, until now, been unable to see at all [1].
The organism in question is called CAG-170. It is not a single species but a previously uncultured genus in the family Oscillospiraceae, comprising thirteen species, eleven of which are strongly associated with health [1][5]. Two unnamed members, catalogued as sp. 900553545 and sp. 900549635, show the strongest negative correlation with Crohn's disease and ulcerative colitis of any microbes in the entire dataset [5]. Yet almost no researcher has ever grown a CAG-170 cell in a dish. Until recently, the genus was known only as a smudge of DNA in stool samples, a fingerprint without a face.
How do you study a bacterium you cannot grow?
That is the central methodological question the new paper, published in Cell Host & Microbe in February 2026, tries to answer [1]. Lead author Alexandre Almeida, of the University of Cambridge's Department of Veterinary Medicine, and his team pulled together 11,115 human gut metagenomes from 62 studies across 39 countries, comprising 13 non-communicable diseases, including Crohn's disease, ulcerative colitis, obesity, colorectal cancer, Parkinson's disease, multiple sclerosis and chronic fatigue syndrome [1][2]. Roughly 36% of samples came from Europe, 29% from North America and 25% from Asia, with the remainder drawn from Africa, South America and Oceania [5].
To make sense of these enormous datasets, the researchers relied on the Unified Human Gastrointestinal Genome (UHGG) catalogue, which Almeida and colleagues first published in Nature Biotechnology in 2020 [4]. That reference set, 204,938 non-redundant prokaryotic genomes, is what allows a computer to recognise a CAG-170 read-out in a 2026 stool sample even though the organism itself has never been coaxed out of a petri dish. The team then ran three independent analyses: a case-control comparison against the UHGG, an ecological-network study of more than 6,000 healthy gut microbiomes, and a dysbiosis analysis that tracked CAG-170 abundance against measures of microbial imbalance [2][6]. CAG-170 was the strongest signal in all three.
A companion paper, posted to bioRxiv in September 2025, makes the broader point explicit: more than 60% of the bacterial species in the human gut remain uncultured and therefore inaccessible to direct experimental manipulation [3]. That single statistic should give any consumer of microbiome science a moment of pause. The probiotics lining pharmacy shelves in Sydney, in London and in New York have been formulated almost entirely from the cultivable minority.
Why the probiotic aisle is behind the science
Most commercial probiotics still belong to two well-trodden genera, Bifidobacterium and Lactobacillus, both of which were characterised decades ago and grow happily in oxygen-tolerant culture [5]. CAG-170 belongs to a different world. Like many strict anaerobes, it is sensitive to oxygen, which makes it a casualty of the very inflammation it may help to prevent. In conditions such as inflammatory bowel disease, the inflamed gut becomes more oxygenated, and oxygen-intolerant commensals are the first to drop out [5]. The signal Almeida's team sees in their case-control comparison, lower CAG-170 in people with Crohn's, with ulcerative colitis, with Parkinson's, may therefore be partly cause and partly consequence, a question the observational design cannot fully resolve [1][10].
It is an important nuance for clinicians and patients to hold in mind. "The link between CAG-170 and health does not yet prove causation, and factors like diet and medication were not consistently controlled for," notes Dr Alena Pribyl, lead scientist at the Brisbane-based microbiome company Microba, in her own commentary on the study [10]. Her team's read of the data also surfaces a striking quantitative observation: uncultured bacteria account for only around 12% of the total microbial abundance in a typical gut sample, yet they carry nearly as much diagnostic information as the cultured 88% [10]. In other words, the parts of the microbiome that the supplement industry has historically ignored may, in aggregate, be the parts most worth knowing about.
Almeida's own framing is blunter. "The probiotic industry hasn't really kept up with gut microbiome research," he said when the paper was released. "People are still using the same probiotic species that were being used decades ago. We're now discovering new groups of bacteria like CAG-170 with important links to our health, and probiotics aimed at supporting them could have a much greater health benefit" [2][9]. For a patient like Maree, that line captures something her gastroenterologist rarely says out loud: the over-the-counter capsules she has been dutifully swallowing were designed for a version of gut science that has since moved on.
The vitamin B12 clue and a possible cross-feeding story
The mechanistic hook, the part that makes CAG-170 a candidate keystone rather than a passenger, is its functional profile. Genomic analysis shows CAG-170 species are enriched for genes involved in vitamin B12 biosynthesis and for a wide array of carbohydrate-active enzymes capable of breaking down fibres, sugars and complex carbohydrates that the human small intestine cannot digest on its own [1][2][9]. They are, in this sense, equipped to do exactly the work the modern Western diet hands off to the colon: liberate energy and micronutrients from the fibrous leftovers of plant-based meals.
There is a twist, though. The researchers interpret the B12 that CAG-170 produces as being more likely to feed other beneficial bacteria than to feed the human host directly, a form of microbial cross-feeding in which the host benefits indirectly, by way of a healthier microbial community [6][7]. The same group of species also appears to lack key genes for arginine biosynthesis, a possible molecular signature of why their absence is associated with dysbiosis [1][3]. Dysbiosis itself is a recognised feature of irritable bowel syndrome, rheumatoid arthritis, anxiety and depression, which is why lower CAG-170 levels in chronic disease are biologically plausible, even if causality remains unproven [2][8][9].
In ecological terms, CAG-170 is also the most interconnected member of the global healthy-microbiome network, the genus with the most links to other microbial partners in a healthy gut [1][6]. That kind of centrality, the data scientists call it "hub status", is exactly what you would expect of a keystone species. The genus was also stable over a year-long follow-up in healthy individuals, a property that makes it a credible candidate for a future diagnostic readout [10].
What happens next
For now, the most honest answer to "should I take a CAG-170 probiotic?" is that you cannot. The organism has not been cultured, which means there is nothing to put in a capsule. Before any CAG-170-based product could reach a pharmacy in Australia, the United States or Europe, the bacterium would need to be isolated, characterised, screened for virulence factors and approved by regulators such as FSANZ, the FDA or EFSA as a novel food ingredient [5]. The realistic near-term applications are therefore diagnostic rather than therapeutic: stool metagenomic tests that report CAG-170 abundance alongside the more familiar Bifidobacterium counts.
Two implications are worth carrying away. First, the next generation of microbiome-targeting supplements is likely to look quite different from the one on today's shelves, anchored in organisms the field has only just learned to name. Second, the dominant signal in this paper, that an uncultured, oxygen-sensitive, B12-producing genus of the Oscillospiraceae family is the single strongest correlate of gut health across continents and diseases, is a reminder that the most important players in the microbiome are not necessarily the most famous ones. For patients like Maree, that message, frustrating as it is, is also the first sign that better-targeted help is on the way.