— October 7, 2026 · Ancient Gut Microbes Traveled the World With Humans

Table of Contents

1. At a Glance
2. What Was Done
3. Microbial Genomics in Plain Language
4. What Was Found, Number by Number
5. Which Microbes Travelled and Which Spread Later
6. How the Ages Were Estimated, and How Sure They Are
7. How It Fits With Earlier Research
8. What It Cannot Show and What Remains Unknown
9. What Experts Said
10. Sources and Primary Documents
11. Key Research Papers
12. Connections

At a Glance

• Paper — Carter MM, Liu Z, Olm MR, et al. “Prehistoric global migration of vanishing gut microbes with humans.” Nature, published online October 7, 2026. Open access.
• Who — 15 researchers from Stanford University and other institutions in the United States, France and Bolivia, working with the Tsimane Health and Life History Project.
• Design — comparative metagenomics (sequencing all the microbial DNA in stool) and population genetics; observational; no health outcomes measured.
• Samples — 133 stool samples from 85 Tsimane people, collected in 2009 and 2012–2013, compared with 32,034 bacterial genomes recovered earlier from 137 Hadza. A further 8,855 genomes from Europe, Asia and North America served as the industrialised comparison.
• Main result — 1,231 species shared, which is 87.4% of the species found in the Tsimane; 848 of them (60.2%) are rare in or absent from industrialised populations.
• Dating — the median estimated time since Tsimane and Hadza strains last exchanged DNA was 17,090 years, against 1,109 years for Europe–North America and 4,637 years for Asia–North America.
• Health claim — none tested; the authors call the health consequences “uncertain.”
• Funding — US and French public science agencies, university fellowships and endowments, and private philanthropy. The authors declare no competing interests.

Back to Table of Contents

What Was Done

The Tsimane are an Indigenous people of the Bolivian Amazon whom the paper describes as forager-horticulturalists: people who combine foraging with small-scale farming. The samples came from nine villages: six along the Maniqui River and three in a nearby forest region. The paper records that at the time of collection none of the participating families had plumbing, used pit toilets or nappies, or had refrigeration. The samples were collected in 2009 and in 2012–2013 by the Tsimane Health and Life History Project, a Bolivian–US team that has worked in Bolivia for more than two decades and provides primary health care to the communities. Participants in the stool study were at least 3 years old, with a mean age of 28.4 years.

The same samples had been studied before with a cheaper method that identifies bacteria from a single marker gene (16S ribosomal RNA). For this paper they were sequenced in depth: every fragment of DNA in each sample was read, to a median of 31.9 gigabases per sample and 5.25 trillion DNA letters in total. Deep sequencing of this kind lets researchers rebuild the genomes of individual bacteria, including rare ones, and compare them strain by strain.

The Tsimane genomes were then set against the team’s earlier ultra-deep collection from the Hadza of Tanzania, whose stool samples were collected in 2013–2014 and published in 2023. To see how strains behave in industrialised societies, the team added 8,855 published genomes from people in Europe, Asia and North America.

The paper describes consent obtained at three levels: the Gran Consejo Tsimane (the Tsimane governing body), community leaders and the individual participants. Individual-level data are held under restricted access following the CARE Principles for Indigenous Data Governance, which give the Tsimane people authority over how their data are shared.

Back to Table of Contents

Microbial Genomics in Plain Language

Rebuilding genomes from stool

A stool sample holds DNA from hundreds of kinds of bacteria at once. In shotgun metagenomics all of that DNA is broken into short pieces and read; computers then sort and stitch the pieces back into draft genomes called metagenome-assembled genomes (MAGs). From the Tsimane samples the team rebuilt 12,746 such genomes, belonging to 1,408 species of bacteria and archaea (a separate branch of single-celled life). The gut microbiome page explains what these communities do in the body.

Species and strains

A species is one kind of microbe. Within a species there are strains, family lines that differ by small changes in their DNA. Two people can carry the same species, but whether they carry closely related strains says something about whether the microbe passed between them, or their ancestors, recently or long ago.

The molecular clock

As bacteria divide, copying errors slowly change their DNA. The paper uses an estimate that gut bacteria accumulate mutations at a rate of at least 10⁻⁷ per DNA letter per year. The share of DNA letters two genomes have in common is called average nucleotide identity (ANI). Tsimane and Hadza genomes of the same species typically matched at about 98%, which the authors translate into a shared ancestor within the past 100,000–200,000 years. That figure is an upper limit, because it ignores gene swapping.

Gene swapping, and how the team read it

Bacteria do not only inherit DNA from their parent cell; they also swap stretches of DNA with other strains (homologous recombination, a form of horizontal gene transfer). That scrambles simple family trees. The team used two signals designed with this in mind. First, pairs of genomes with more than 10% of their genes completely identical were treated as near-clones that shared an ancestor within roughly the last 5,000 years. Second, after a swap, two strains carry a long identical stretch of DNA that is worn down over time by new mutations. Measuring the longest identical stretches (the 99th percentile, called L99) within and between populations shows how long ago the two populations last swapped genes. A third, independent method (a software package called Moments) fitted the frequencies of single-letter DNA variants to a model of two populations splitting apart.

The animated explainer The Gut Barrier and Your Microbiome shows where these bacteria live in the intestine.

Back to Table of Contents

What Was Found, Number by Number

Shared species

• Of the 1,408 species recovered from the Tsimane, 1,231 (87.4%) were also found in the Hadza, despite the large distance between them and what the paper calls a “lack of historical direct contact.”
• Of the shared species, 848 (60.2%) are rare in or absent from the stool of people in industrialised populations.
• Matching sequencing depth between the two groups changed the shared count only from 1,231 to 1,227. A second check using a different detection method found that 703 of 759 species judged absent from industrialised samples were present in fewer than 1% of them.
• Statistical estimates of how many species each population holds in total came to 2,443 for the Hadza (95% confidence interval 2,385–2,503) and 1,590 for the Tsimane (1,561–1,620).

Genetic distance

• 636 shared species had at least four genomes in each population and were analysed in detail.
• Genomes from the same population were, on average, more alike than genomes from different populations (P < 0.001), but the difference was small next to the variation within each population: the mean fixation index (F_ST, a standard measure of how distinct two populations are) was 0.10. For the human genomes of the two peoples the paper cites a value of about 0.25.
• The authors read this to mean that most of the strain diversity in the Tsimane microbes was already present before the two populations were separated.

Recent sharing

• Across the 636 species, 78,746 genome pairs (7.8%) were near-clones (ancestor within roughly 5,000 years).
• 93.8% of those near-clone pairs were found within the same population (P < 10⁻⁴).
• 545 of the 636 species (85.6%) showed no recent strain sharing at all between the Tsimane and the Hadza (P < 10⁻⁴).

Gene swapping and split times

• Most species had shorter identical stretches between the Tsimane and Hadza than within each group (P < 10⁻⁴), the signature of populations that stopped swapping genes.
• That barrier was stronger than between industrialised populations (Asia–North America, P = 0.01; Europe–North America, P = 0.41). The same held among species common in every lifestyle (P = 0.018).
• Median estimated time since strains last exchanged DNA: Tsimane–Hadza 17,090 years (standard error 863 years); Europe–North America 1,109 years (377); Asia–North America 4,637 years (1,267); P < 10⁻⁴.
• Only 15 species had enough data for the third, independent method. Its results “broadly align” with the main estimates and favour an ancient split over no separation; the two methods’ dates for those species correlated at r = 0.581 (P = 0.022).

Back to Table of Contents

Which Microbes Travelled and Which Spread Later

The paper discusses the shared microbes in three groups, using two labels from earlier work by the same research group and one new one.

VANISH species: the long-term companions

VANISH stands for “volatile and/or associated negatively with industrialized societies of humans.” Many of the shared species had been identified as VANISH taxa in earlier work, based on their enrichment in the Hadza. The new genetic analysis shows that many VANISH species were also present in the common ancestors of the Tsimane and Hadza. The authors suggest that “widespread losses might explain the absence of these species from industrialized populations.” Many of them have only laboratory code names so far; two the paper uses as worked examples are Bulleidia sp905194705 and Vescimonas sp900551995.

BloSSUM species: recent arrivals

A few species showed as much Tsimane–Hadza strain sharing as sharing within each group, which points to recent spread rather than ancient travel. They include Akkermansia muciniphila and Bacteroides ovatus, species the paper says “tend to be enriched in industrialized populations.” Close relatives of these strains turned up in genomes from Europe, Asia and North America, which the authors read as a sign that such species reached the Tsimane and Hadza through more recent contact with other human populations. In their discussion they describe examples of BloSSUM species (“bloom or selected in societies of urbanization or modernization”) with extensive strain sharing across lifestyles, which “probably resulted from species transmission from industrialized populations to the Tsimane and Hadza.” Escherichia coli also showed high recent sharing across populations.

StIL species: common everywhere, but still isolated

A third group, named StIL (“stable independent of lifestyle”), is found in people of every lifestyle, including industrialised ones. Examples are Agathobacter rectalis and Ruminococcus bromii. Even so, the Tsimane and Hadza strains of these species showed little recent sharing, and their isolation was generally stronger than between industrialised populations separated for similar lengths of time. The authors propose that ongoing exchange of microbes in industrialised populations has “probably degraded signals of co-migration,” which “may partly explain why previous observations of these signals have been limited.”

Back to Table of Contents

How the Ages Were Estimated, and How Sure They Are

The study gives three different kinds of age, and they answer different questions:

• 100,000–200,000 years — the window within which typical Tsimane and Hadza strains shared an ancestor, from overall DNA similarity. The paper describes this as an upper bound.
• About 5,000 years — the cut-off used to call two strains near-clones. It varies by species: about 4,000 years for Bulleidia sp905194705, and a median of about 2,000 years across species (95% range 500–5,000 years).
• 17,090 years (median) — the estimated time since Tsimane and Hadza strains last exchanged DNA. This is the “split time” the authors compare with human migration.

For comparison, the paper places the first migration of humans out of Africa at around 60,000–70,000 years ago and the settlement of the Americas at around 16,000–23,000 years ago. It says the Tsimane–Hadza estimate “roughly aligns with the timeframe between” those events, “considering the uncertainties in our molecular clock estimates.” The reported standard error of 863 years reflects variation among the individual species estimates; the separate uncertainty in the mutation-rate clock is set out in the paper’s supplementary methods, which were not read for this report.

Some coverage said the strains “shared a common ancestor” about 17,000 years ago. In the paper, the 17,090-year figure is the time since the strains last swapped genes; their common ancestry, by the paper’s overall similarity estimate, goes back further.

The second, independent method could only be run on 15 species, and the authors note that applying it to microbes has “challenges, such as small sample sizes and complex population structures.” For those 15 species, the dates from the two methods correlated at r = 0.581.

Back to Table of Contents

How It Fits With Earlier Research

Bacteria that track human history

The idea that microbes can record the movements of their hosts is not new. A 2007 study of the stomach bacterium Helicobacter pylori found that its genetic diversity falls with distance from east Africa, as human diversity does, and simulations suggested it spread from east Africa around 58,000 years ago (Linz 2007). A 2016 study found that lineages of the gut bacterial families Bacteroidaceae and Bifidobacteriaceae arose by “cospeciation” with humans, chimpanzees, bonobos and gorillas over the past 15 million years (Moeller 2016). A 2023 study of 9,460 primate and human genomes reported that about 44% of bacterial groups that co-diversified with African apes were absent from available human data, and about 54% were absent from industrialised humans, against about 3% of other groups (Sanders 2023).

Within modern humans, a 2022 study of paired human and microbial genomes from 1,225 people in Europe, Asia and Africa reported a parallel evolutionary history for people and their gut microbes (Suzuki 2022). In July 2026, Benjamin Good, one of the senior authors of the new paper, re-analysed such data and reported that the links between human and microbial family trees are often weaker than those between unlinked human chromosomes (ones inherited independently of each other), and that similar links can arise from geography alone (Good 2026). The new paper asks a related but different question, whether microbes moved with people, and uses methods built to account for bacterial gene swapping.

Microbes missing from industrialised guts

DNA rebuilt from eight preserved human stools 1,000–2,000 years old, from the southwestern United States and Mexico, yielded 498 microbial genomes, 39% of the most reliable ones from species not described before. Those ancient microbiomes looked more like present-day non-industrialised microbiomes than industrialised ones (Wibowo 2021). In the Hadza, 350 stool samples collected over more than a year showed some bacteria disappearing and returning with the seasons. The most changeable of these were the same groups that separate industrialised from traditional populations (Smits 2017). The team’s 2023 ultra-deep study estimated that the average Hadza adult gut holds 730 species (standard error 14.5), against 277 (standard error 32) in the average Californian adult, and identified 124 gut species vanishing in industrialised populations (Carter 2023). The “750 versus 250” figures in some coverage of the new paper appear to be rounded from that 2023 study; they are not in the Nature paper. A 2019 review by two of the authors names antibiotics, sanitation and processed food among the lifestyle factors linked to these shifts (Sonnenburg and Sonnenburg 2019).

Diet studies from the same laboratory

In mice carrying human gut bacteria, a diet low in microbiota-accessible carbohydrates (the parts of dietary fibre that gut bacteria can ferment) caused changes that were largely reversible within one generation. Over several generations, though, it led to a progressive loss of diversity that adding the fibre back did not recover; restoring the community required giving back the missing bacteria together with the fibre (Sonnenburg 2016). The site’s Dietary Fiber and Prebiotics pages cover this area.

In a 17-week randomised study in 36 healthy adults (18 per arm), participants on a high-fermented-food diet went from 0.4 to 6.3 servings a day of foods such as yogurt, kefir, fermented cottage cheese, kombucha, vegetable brine drinks and fermented vegetables like kimchi. Their gut microbial diversity rose steadily, and 19 of 93 inflammatory proteins measured in blood fell, including interleukin-6. The high-fibre arm, which went from 21.5 to 45.1 grams of fibre a day, showed stable diversity but an increase in bacterial enzymes that break down fibre; in that arm the primary outcome, a cytokine response score, was unchanged. The new microbes that appeared on the fermented-food diet were largely not the ones in the foods themselves. The authors list a modest number of participants and no control arm among the study’s limits (Wastyk 2021). That trial is covered on Fermented Foods for Gut Microbiome Diversity. None of these diet studies involved the Tsimane or the Hadza, and the Nature paper did not test any food.

Back to Table of Contents

What It Cannot Show and What Remains Unknown

Limits stated in the paper

• No health data. The study shows where microbes came from and how they spread, not what they do for health. The abstract says the consequences of losing them “remain uncertain.”
• Present-day people as a window on the past. The Tsimane and Hadza are contemporary populations; the ancestral picture is inferred from their microbes’ DNA, not sampled from ancient people. The paper notes that existing preserved ancient stools are only about 1,000–2,000 years old, which, with the difficulty of analysing degraded ancient DNA, makes it hard to tell whether microbes were passed down over 10,000–100,000 years.
• Recovery of rare genomes. The authors write that differences in strain variety and sequencing depth between the two cohorts “may have contributed to the differential recovery of some rare MAGs.”
• Clock uncertainty. Every date depends on an estimated mutation rate, and gene swapping complicates the arithmetic. The overall-similarity dates are upper bounds, and the independent method covered only 15 species.
• Strain-sharing counts are shaped by who has contact with whom and by sweeps of single strains through a population, which the authors say “can be difficult to compare across populations.”

Open questions

• Whether the losses matter for health. In their discussion the authors note that the loss of VANISH species has been linked to a higher risk of several chronic metabolic and inflammatory disorders, and write that this “raises the possibility that reintroducing these species may help to restore microbial functions that co-evolved with human biology.” That is a possibility they raise, not something this study tested. They add that such work “will require thoughtful and inclusive dialogue between researchers, ethicists and Indigenous communities.”
• Why the species were lost. The paper does not test causes. Earlier work points to diet, antibiotics and sanitation; the paper does not identify which factor removed which species.
• Other populations. Only two non-industrialised groups were compared, so the pattern in other rural and Indigenous populations was not examined.
• Timing since collection. The Tsimane samples date from 2009–2013; the paper does not report how the communities’ microbiomes may have changed since.

Back to Table of Contents

What Experts Said

A News & Views commentary published in Nature the same day, by Carina Schlebusch and Maja Vukovikj of the Department of Organismal Biology at Uppsala University, opens: “Humans have never travelled alone.” It summarises the paper’s report that “hundreds of microbial species are present in both groups” and adds: “Many of these shared microbes are now rare or missing in populations from industrialized regions.” (Only the opening of the commentary could be read for this report.)

In Stanford Medicine’s news release, as republished by News-Medical, senior author Justin Sonnenburg said: “Our study establishes that the hundreds of bacterial species that are rare or missing in industrialized microbiomes were ancient companions of ours as we migrated around the globe, likely passed from generation to generation for millennia.” Co-senior author Benjamin Good said: “We see evidence that these bacteria were present in the ancestors of both the Tsimane and the Hadza, and therefore presumably the ancestors of all of us.”

Speaking to NPR, Sonnenburg said: “If we were inhabited by very similar microbes over a long period of time, it suggests that our human genomes may have come to expect a certain set of microbes and functions.” Good told NPR: “Once the land bridge went away, we had a clear picture that it was going to be very hard for them to exchange gut bacteria through traditional human contact means.”

Two researchers not involved in the study offered different emphases to NPR. Andrew Moeller of Princeton University said: “Another alternative is that the industrialized microbiome profile may actually be adapted to an industrialized lifestyle,” and: “From a health perspective, is it good or deleterious that industrialized humans have lost a lot of ancestral microbial taxa? We actually don’t know.” He added: “With regards to the evolutionary questions like, which of these microbes have we co-evolved with? Which ones have we lost that would be beneficial to regain? It’s still early days for that line of questioning.” Taichi Suzuki of Arizona State University said: “If the bacteria and the humans had a shared history for many, many generations, there are more opportunities for selection.”

Back to Table of Contents

Sources and Primary Documents

Every figure attributed to the study above was taken from the open-access full text of the Nature paper (main text, methods and extended-data legends). The supplementary methods were not read. The species counts per person come from the 2023 Cell paper, as stated.

1. Carter MM, Liu Z, Olm MR, Martin M, Sprockett DD, Ghadermazi P, Trumble BC, Kaplan H, Stieglitz J, Rodriguez DE, Relman DA, Sonnenburg ED, Gurven M, Good BH, Sonnenburg JL (2026). Prehistoric global migration of vanishing gut microbes with humans. Nature, published online October 7, 2026. — doi:10.1038/s41586-026-11106-1 · PubMed PMID: 42844483
2. Schlebusch CM, Vukovikj M (2026). Gut microbes that trace ancient human journeys. Nature (News & Views), October 7, 2026. — doi:10.1038/d41586-026-02969-5
3. Carter MM, Olm MR, Merrill BD, Dahan D, Tripathi S, Spencer SP, et al. (2023). Ultra-deep sequencing of Hadza hunter-gatherers recovers vanishing gut microbes. Cell 186(14):3111-3124.e13 (source of the 730 and 277 species figures). — doi:10.1016/j.cell.2023.05.046 · PubMed PMID: 37348505

Coverage: Stanford Medicine news release as republished by News-Medical, October 7, 2026 (Ancient gut microbes migrated with humans across the globe); NPR, October 7, 2026, as republished by KPBS (Gut microbes followed migrating humans across continents, new study suggests).

Back to Table of Contents

Key Research Papers

Background papers in the order they appear in the article. Each entry was checked against PubMed and Crossref before it was written down.

1. Linz B, Balloux F, Moodley Y, Manica A, Liu H, Roumagnac P, et al. (2007). An African origin for the intimate association between humans and Helicobacter pylori. Nature 445(7130):915-918. — PubMed PMID: 17287725
2. Moeller AH, Caro-Quintero A, Mjungu D, Georgiev AV, Lonsdorf EV, Muller MN, et al. (2016). Cospeciation of gut microbiota with hominids. Science 353(6297):380-382. — PubMed PMID: 27463672
3. Sanders JG, Sprockett DD, Li Y, Mjungu D, Lonsdorf EV, Ndjango JN, et al. (2023). Widespread extinctions of co-diversified primate gut bacterial symbionts from humans. Nature Microbiology 8(6):1039-1050. — PubMed PMID: 37169918
4. Suzuki TA, Fitzstevens JL, Schmidt VT, Enav H, Huus KE, Mbong Ngwese M, et al. (2022). Codiversification of gut microbiota with humans. Science 377(6612):1328-1332. — PubMed PMID: 36108023
5. Good BH (2026). Limited codiversification of the gut microbiota within humans. mBio 17(8):e0372725. — PubMed PMID: 42405797
6. Wibowo MC, Yang Z, Borry M, Hübner A, Huang KD, Tierney BT, et al. (2021). Reconstruction of ancient microbial genomes from the human gut. Nature 594(7862):234-239. — PubMed PMID: 33981035
7. Smits SA, Leach J, Sonnenburg ED, Gonzalez CG, Lichtman JS, Reid G, et al. (2017). Seasonal cycling in the gut microbiome of the Hadza hunter-gatherers of Tanzania. Science 357(6353):802-806. — PubMed PMID: 28839072
8. Sprockett DD, Martin M, Costello EK, Burns AR, Holmes SP, Gurven MD, Relman DA (2020). Microbiota assembly, structure, and dynamics among Tsimane horticulturalists of the Bolivian Amazon. Nature Communications 11(1):3772. — PubMed PMID: 32728114
9. Sonnenburg JL, Sonnenburg ED (2019). Vulnerability of the industrialized microbiota. Science 366(6464):eaaw9255. — PubMed PMID: 31649168
10. Sonnenburg ED, Smits SA, Tikhonov M, Higginbottom SK, Wingreen NS, Sonnenburg JL (2016). Diet-induced extinctions in the gut microbiota compound over generations. Nature 529(7585):212-215. — PubMed PMID: 26762459
11. Wastyk HC, Fragiadakis GK, Perelman D, Dahan D, Merrill BD, Yu FB, et al. (2021). Gut-microbiota-targeted diets modulate human immune status. Cell 184(16):4137-4153.e14. — PubMed PMID: 34256014

PubMed Topic Searches

1. PubMed: The Hadza gut microbiome
2. PubMed: The Tsimane microbiota
3. PubMed: Gut microbiota codiversification with humans
4. PubMed: Loss of gut microbial diversity with industrialisation
5. PubMed: Fermented food trials and gut microbial diversity

Back to Table of Contents

Connections

• News Archive
• Gut Health — the site’s hub for digestion and the gut
• Your Microbiome — what the gut’s bacteria are and what they do
• The Gut Barrier and Your Microbiome — animated explainer
• CAG-170: A Hidden Microbe Found in Healthy Guts — an earlier news report on a little-known gut bacterium
• Fermented Foods — how fermentation works and the research on it
• Fermented Foods for Gut Microbiome Diversity — the 2021 diet trial in detail
• Dietary Fiber — the plant material gut bacteria ferment
• Prebiotics — fibres that feed gut bacteria
• Probiotics — live bacteria in foods and supplements
• Helicobacter Pylori — the stomach bacterium whose genes trace human migration
• Escherichia Coli — a species the study found widely shared

Back to Table of Contents