Your Gut Bacteria Reprogram Based on Their Neighbors · Thaena Inc.
New research reveals that gut bacteria aren't static machines running the same program every day — they completely reorganize their molecular toolkit based on who they're living next to. Which means cataloging species names in your gut may tell you almost nothing about what your microbiome is actually doing.
Lit Review Friday · Learn Something with Thaena · Published 2026 · Reading time: ~12 minutes
🎧 Listen to the full episode: Your Gut Bacteria Reprogram Based on Their Neighbors
Available on Spotify (Apple Podcasts, and wherever you listen coming soon)
The Finding That Shatters the Static Model
In 2026, Stephan Kamrad and colleagues at the Francis Crick Institute published a landmark paper in Nature Ecology and Evolution with a deceptively simple premise: what happens inside a gut bacterium when it meets a neighbor?
The answer turned out to be far more radical than anyone expected.
Across 104 pairwise co-cultures of 15 gut bacteria, the researchers found that roughly half of each bacterium's active protein machinery changed significantly depending on which neighbor it was paired with. Same species. Same genome. Completely different molecular toolkit.
- ~50% of each bacterium's quantified proteome changed in at least one co-culture condition
- 38% of those responsive proteins changed in only a single pairing — a tailored, not generic, response
- 15 species spanning 4 phyla, including commensals, two pathogens (C. difficile, K. aerogenes), and two probiotics (L. acidophilus, L. gasseri)
- 104 pairwise co-cultures, each with proteomics and metabolomics measured at stationary phase
- Top predictors of how much a bacterium changed: pH shift and relative abundance — not phylogenetic distance
If half of the active machinery in a gut bacterium is being swapped out based purely on neighbor presence, our static models of the microbiome are not just incomplete. They are fundamentally wrong.
What They Did: Bacterial Speed Dating at the Molecular Level
To understand this paper's significance, you have to appreciate what made its methodology a genuine leap forward. Most microbiome studies are essentially ecological surveys — they tell you who is present. Kamrad's team wanted something far more ambitious: a real-time map of what changes inside the cells when bacteria encounter each other.
They selected 15 human gut bacteria representing the four major phyla. The Bacteroidetes are gram-negative powerhouses with massive genomes dedicated to breaking down complex polysaccharides — the gut's primary carbohydrate degraders. The Firmicutes specialize in taking the intermediate products from Bacteroidetes and fermenting them further into short-chain fatty acids like butyrate. The Actinobacteria, particularly Bifidobacterium species, are key for early-life colonization and immune communication. And the Proteobacteria are the facultative anaerobes — low abundance in a healthy gut, but capable of blooming rapidly when inflammation lets oxygen creep in.
Critically, they didn't just use friendly commensals. They introduced two pathogens and two probiotics into the mix, creating what amounts to a highly pressurized ecological arena.
Then came the combinatorial design: every possible two-species pairing, 104 combinations total, grown in parallel. At stationary phase — when nutrients are scarce and bacteria are forced to negotiate with neighbors rather than just sprint for growth — they harvested both the proteome (what proteins are actively being made) and the metabolome (what small molecules are being secreted into the surrounding liquid).
The Technology That Made It Possible
Getting high-resolution proteomics across 104 varied conditions was historically impossible. The team used a mass spectrometry technique called dia-PASEF (data-independent acquisition, parallel accumulation serial fragmentation). Older mass spectrometers essentially focus on the most abundant proteins and ignore quieter signals — which means they miss the regulatory proteins, the sensor kinases, the transcription factors that are actually making the decisions inside the cell. dia-PASEF separates ions by shape and size before fragmentation, cycling through and capturing literally everything. The result is an unfiltered snapshot of the entire active proteome, including the low-abundance machinery that previous methods missed entirely.
Why stationary phase? During rapid growth, bacteria are essentially just sprinting to build biomass — the interesting ecological dynamics are suppressed. Stationary phase is when nutrients run low and waste accumulates, forcing bacteria to actively manage shared resources and respond to their neighbors. It's the difference between watching people in a buffet line versus watching them negotiate the last available food in a room they've been locked in for a week.
What They Found: The Proteome Is Not Fixed
The 50% Finding
The headline number is staggering: roughly half of each bacterium's quantified proteome changed significantly in at least one co-culture. We typically assume a bacterium's core metabolic pathways are stable — that Bacteroides thetaiotaomicron, for instance, is always out there doing its textbook job of foraging for complex carbohydrates. This paper shows that assumption is wrong. The same species, with the same genome, unpacks a completely different molecular toolkit depending on who is sitting next to it.
And it wasn't a generic panic response. When you hit a bacterium with bleach, it upregulates a standard stress response — universal, non-specific. What Kamrad's team found was nothing like that. 38% of the responsive proteins changed in only a single co-culture pairing. B. uniformis deployed a completely different arsenal when it met E. coli versus when it met C. difficile. These are bespoke molecular conversations, not alarm sirens.
pH Is the Master Switch
If phylogenetic distance — how closely related two species are — doesn't predict the proteome response, what does? The data pointed to two primary drivers: the relative abundance of the partner species, and shifts in pH.
pH was the stronger predictor. When a heavy fermenter like a Bacteroidetes species metabolizes sugars, it excretes organic acids and the pH of the local environment drops sharply. That physical shift alters the shape of sensory proteins on neighboring bacteria's membranes, which cascades down to the DNA and triggers wholesale proteome overhaul. The most responsive biological processes in the data were carbohydrate transport systems (the PTS system and ABC transporters) and carbon metabolism — exactly what you'd expect when a neighbor changes the pH and depletes a carbon source, forcing rapid synthesis of new transporters to scavenge alternative food.
Even orthologous proteins — genes that are directly related across species and perform the same function — showed divergent expression profiles in response to the same partner. The tool is identical, but the sensory circuitry deciding when to use the tool has been completely rewired. You genuinely cannot extrapolate behavior based on a bacterium's family tree.
Emergent Metabolism: Chemistry That Requires a Community
The metabolomics data revealed something even more striking than the proteome shifts: whole categories of metabolites that simply didn't exist when any bacterium was grown alone, but appeared in specific co-culture pairings. This is emergent metabolism — the whole is chemically distinct from the sum of its parts.
GABA: The Silent Neurotransmitter Factory
GABA (gamma-aminobutyric acid) is the primary inhibitory neurotransmitter in the human brain and a key signaling molecule in the enteric nervous system of the gut. The researchers found GABA emerging specifically in co-cultures involving Bacteroides uniformis. B. uniformis has the genes to produce GABA, but when it's alone, that pathway is silent. It apparently needs the physical and chemical friction of a specific neighbor to trigger the proteome shift that turns the GABA factory on.
Indole-Acetic Acid: The Pathogen That Strengthens Your Gut Barrier
Indole-acetic acid (IAA) is a tryptophan metabolite and a known ligand for the aryl hydrocarbon receptor (AhR) in the human immune system — crucial for maintaining intestinal barrier integrity and modulating immune tone. IAA spiked specifically in co-cultures containing C. difficile.
That is an ecological paradox worth sitting with. The notorious pathogen responsible for antibiotic-resistant infections is being chemically coerced by its commensal neighbors into producing a molecule that actively strengthens the host's gut barrier. The commensals are, in effect, pacifying the pathogen.
The Polyamine Escape Room
The most thoroughly mapped example of emergent metabolism was the appearance of the polyamines agmatine and putrescine — and the mechanism reads like a chemical escape room.
- E. coli and a Bacteroidetes neighbor are co-cultured. The Bacteroidetes ferments sugars and pumps out organic acids. pH drops sharply.
- E. coli is now facing lethal acid stress. The influx of protons threatens to collapse its internal chemistry.
- To survive, E. coli alters its proteome and upregulates arginine decarboxylase — an enzyme that converts arginine into agmatine. That specific reaction consumes a proton, chemically neutralizing the acid threat.
- Agmatine spills into the environment as a byproduct of E. coli's survival mechanism. The human host absorbs it — and benefits, because agmatine has documented neuroprotective and anti-inflammatory properties.
We benefit from their microscopic warfare. Neither bacterium was "trying" to help us. The emergent chemistry is a byproduct of two organisms solving their own survival problems in the same space. This is the community context the static model completely misses.
The C. difficile Paradox: Context Determines Everything
The finding that will likely generate the most citations involves C. difficile — and it completely overturns one of the central assumptions in gut microbiology.
The standard model holds that a healthy, diverse microbiome acts as colonization resistance: the good bacteria starve out or fight off pathogens like C. diff. But when paired with several Bacteroidetes species — the commensals we typically consider beneficial — C. difficile grew significantly better than it did alone.
The mechanism traces to starch. The experiment used a medium called mGAM, which is rich in complex starches. C. difficile lacks the enzymatic scissors to break down complex starches itself — it's effectively starving in a room full of locked food. But B. thetaiotaomicron has massive starch utilization systems on its outer membrane. It hacks the starches into smaller simple sugars and doesn't consume all of them itself. The liberated sugars diffuse outward, and C. difficile pulls up a chair and feeds on the leftovers.
To prove this genetically, the team created a knockout mutant: the exact same B. thetaiotaomicron, but missing a single gene — susG, which codes for the specific alpha-amylase that initiates starch breakdown. When they paired this mutant with C. difficile, the growth promotion vanished completely. One protein. One enzyme. The difference between a pathogen that starves and one that thrives.
When the researchers swapped to a standard lab medium (BHI) without complex starches, the relationship between B. thetaiotaomicron and C. difficile completely inverted. Without starch to share, they competed for the same amino acids — and B. theta heavily inhibited C. diff.
In one nutritional environment, they are partners. In another, mortal enemies. You cannot categorize these interactions as inherently good or bad. Microbial interaction is an emergent property of the nutritional landscape.
The Metabolite That Actually Controls C. diff
The Kamrad data raises a deeper question: if C. diff's behavior is shaped by the metabolic environment its neighbors create, what specific molecular signals are doing the controlling? A 2018 paper by McDonald et al. in Gastroenterology answered that with unusual precision.
The researchers zeroed in on valerate — a short-chain fatty acid produced by a healthy, diverse gut community through fermentation. The finding was direct: when valerate is present at physiological concentrations, C. difficile growth is suppressed. When valerate disappears — which is exactly what happens when broad-spectrum antibiotics disrupt the community that produces it — C. diff blooms. And when fecal microbiota transplantation restores the community, valerate comes back, and C. diff goes back down.
This is not a colonization resistance story, where good bacteria physically crowd out bad ones. It is a postbiotic story. One chemical signal, produced by the community as a metabolic byproduct, is continuously suppressing a dangerous pathogen. The community doesn't fight C. diff directly. It maintains a chemical environment in which C. diff cannot thrive. Remove the chemistry and the pathogen has the space it needs.
When You Remove All the Bacteria — and It Still Works
The most direct clinical test of this idea came from Ott et al. in 2017, also published in Gastroenterology. The team had five patients with severe recurrent C. difficile infections — patients who had failed multiple rounds of antibiotics. Standard FMT was either unavailable or contraindicated for some due to immunosuppression. So the researchers tried something different.
They took healthy donor stool and ran it through a sterile 0.2-micron filter — removing every living bacterium. What remained was a light brown liquid containing metabolites, proteins, antimicrobial compounds, bacteriophage DNA, and the other non-living products of a healthy community. No viable organisms. Just the chemistry a functioning microbiome had made.
That sterile filtrate was administered to all five patients via nasojejunal tube. All five resolved. Most became symptom-free within two to four days. Several remained symptom-free for over two years of follow-up.
The therapeutic signal in fecal microbiota transplantation does not require living bacteria. Sterile filtrate — the metabolites and non-living products of a healthy community — was sufficient to resolve recurrent C. difficile infection in every patient treated.
The authors concluded: "This finding indicates that bacterial components, metabolites, or bacteriophages mediate many of the effects of FMT."
Read together, McDonald 2018 and Ott 2017 reframe the entire logic of microbiome therapeutics. If a single community-produced metabolite can directly suppress a dangerous pathogen, and if the active ingredient in FMT is the chemistry the community makes rather than the bacteria themselves, then the central therapeutic target was never the microbial roster. It was always the metabolic output of a healthy ecosystem. These two papers are among the foundational evidence behind Thaena's postbiotic thesis — and the Kamrad 2026 paper now gives us a mechanistic explanation for why that output is so complex, so environment-dependent, and so difficult to replicate with individual strains.
What It Means: The Roster Doesn't Predict the Game
This paper's implications reach far beyond its specific findings. They force a reckoning with how we've been thinking about the microbiome for two decades.
The central dogma of molecular biology describes the basic pipeline of life: DNA is transcribed into RNA, which is translated into proteins (enzymes), and those enzymes produce metabolites. We have spent enormous resources on metagenomic sequencing — mapping the DNA of the gut microbiome — because the assumption was that knowing who's there would tell us what they do.
This paper shows the assumption is broken. The DNA is the potential. The community context determines what actually gets realized. The proteome — which enzymes are actually being built and deployed — is community-dependent. And the metabolites downstream, the molecules that actually communicate with your immune system, your brain, and your liver, are determined by community context, not by species identity.
16S sequencing and metagenomics give you a roster. This paper shows the roster doesn't predict the game.
The Redox Hypothesis: Going Beyond What the Paper Says
pH was the top predictor of proteome response in the Kamrad data. But pH is a proxy for something deeper: the gut's redox environment, its oxidation-reduction potential.
A healthy colon is highly reduced — it maintains a very low, negative redox potential. Strict anaerobes require that low redox state to function. Their enzymes are highly sensitive to oxidative damage. When Bacteroidetes species ferment and drop the pH, they're also generating reducing equivalents and maintaining that low redox state. When the redox potential rises — from inflammation, a poor diet, or antibiotic disruption — it creates oxidative stress throughout the microbial community. Bacteria that were peacefully producing butyrate and GABA and indoles are now spending their proteomic budget on defensive survival pathways instead.
Here is the hypothesis this paper prompts but does not state: if you change the redox environment, you may change what your microbiome does without changing which species are present at all. The bacteria are the same. The community composition looks identical on a sequencing readout. But the chemical environment they're swimming in has shifted — and so has everything they're producing.
This reframes dysbiosis not as a compositional problem (bad species taking over) but as a thermodynamic one. The right bacteria are still there. But they've been forced into a defensive configuration by an environment that's demanding it. And interventions that only address composition — by adding new species or removing old ones — may be solving the wrong problem.
Why Probiotics Fail (and What That Means)
The probiotic industry is built on the premise that you can take a specific beneficial strain, grow it in a vat, and drop it into a human gut to confer its benefits. The Kamrad paper explains, with mechanistic clarity, why this so often doesn't work.
If you drop a probiotic strain into a dysbiotic gut — one with elevated redox potential, abnormal pH, and a disrupted community — that strain immediately encounters a hostile chemical context. Based on what this paper shows, it will undergo a massive proteomic reorganization just to survive. It will unpack a completely different toolkit than the one it expressed in the controlled, optimized lab conditions where its benefits were characterized. You're delivering a specific machine, but the environment is reprogramming its software the moment it arrives.
Beyond that, some of the most health-relevant metabolites we identified above — GABA, polyamines, indole-acetic acid — require a community context to even appear. A single strain, alone, cannot produce them. They are emergent properties of multi-species interaction. A probiotic capsule with one or three species cannot replicate the chemical richness of a functional microbial ecosystem.
The Postbiotic Argument: Fix the Playing Field, Not the Roster
If individual strains can't replicate the community's chemical output, and if the community's chemical output is what actually matters for host health, then the logical intervention shifts: deliver the finished metabolic products of a healthy, diverse, functioning community directly.
This is the core thesis behind postbiotics. Rather than trying to seed the gut with new bacteria and hope they behave appropriately in a potentially hostile environment, you deliver what healthy bacteria actually make — the SCFAs, bile acids, indoles, antioxidant molecules, and signaling compounds that collectively represent the chemical output of a thriving ecosystem.
Frame this through the redox lens and something elegant emerges. Bile acids, short-chain fatty acids, and indoles in a full-spectrum postbiotic collectively act as antioxidants — they absorb oxidative stress and support a reduced gut environment. And a reduced gut environment, as this paper shows, is the condition under which beneficial bacteria produce their full repertoire of beneficial chemistry. You're not overwriting the roster. You're fixing the playing field so the players you already have can do their jobs.
A full-spectrum postbiotic derived from a healthy human donor microbiome captures two things that no single-strain approach can:
- The emergent chemistry — metabolites like GABA, polyamines, and indoles that only appear when a full community is functioning. These are already present as finished products, not dependent on community context to be produced.
- The redox-supporting environment — bile acids, SCFAs, antioxidant molecules that collectively recreate the reduced chemical conditions under which your existing microbiome produces its full beneficial output.
We believe this is a plausible mechanism. It's a thesis, not a proven clinical outcome. The human trial data needed to validate it is part of what Thaena is working toward.
The Honest Limitations
The Kamrad paper is a landmark, but it's important to be clear about what it can and cannot tell us.
These were pairwise co-cultures in a tube — not the full complexity of the human gut, where hundreds of species interact simultaneously. The emergent properties of a three-way or ten-way interaction might look entirely different from any pairwise combination. If B. theta is sharing starch with C. difficile, but a third aggressive scavenger species is present, it may outcompete C. diff for those sugars before any sharing happens. Non-linear dynamics in full communities are genuinely unpredictable from pairwise data.
The C. difficile growth results were entirely medium-dependent, as the researchers demonstrated themselves. mGAM is starch-rich; BHI is not. The gut lumen is neither, and its nutritional environment changes with every meal. Translating these findings to in vivo conditions requires caution.
The proteomics snapshot was taken at stationary phase only. The gut is a continuous flow environment where nutrients pulse in and populations shift constantly. A mid-log phase snapshot might tell a different story.
Redox potential was not directly measured. pH was used as a proxy, and while the two are related, direct measurement of oxidation-reduction potential would be needed to fully validate the thermodynamic hypothesis the paper suggests.
And perhaps most humbling: a large fraction of the proteins undergoing the biggest regulatory shifts were completely unknown to science — no known gene function, no characterization, no idea what they do. Hundreds of highly conserved proteins, essential for bacterial communication, and we are entirely blind to their function. Deciphering that "genetic dark matter" is arguably the most urgent directive for microbiome research right now.
The Bottom Line
The question this paper leaves us with is more important than any specific finding: if microbial behavior is entirely shaped by the chemical environment bacteria swim in, then perhaps our ultimate goal isn't to seed our guts with new bacteria at all.
The more pressing question is what signals we are sending to the bacteria we already have. Are the highly processed diets, chronic stress, and environmental exposures of modern life inadvertently triggering our normal, peaceful microbes to unpack a hostile toolkit — not because the bad bacteria won, but because the environment is demanding a defensive response?
The bugs aren't the problem. The environment we're providing might be.
- ~50% of a gut bacterium's active proteome changes depending on its neighbors — the same genome, radically different molecular behavior
- These responses are highly partner-specific, not generic stress reactions — bacteria have tailored molecular conversations
- pH shift and relative abundance predict proteome response better than species identity or phylogeny
- Emergent metabolites — GABA, indole-acetic acid, polyamines — only appear in co-culture; no single species produces them alone
- C. difficile growth behavior completely reversed depending on medium — commensals can feed pathogens in the right nutritional context
- The roster doesn't predict the game: metagenomic sequencing tells you who's present, not what they're doing
- Dysbiosis may be a thermodynamic collapse — existing good bacteria forced into defensive mode — not just a compositional takeover
- Probiotics that enter a dysbiotic gut may immediately reprogram into unhelpful configurations; single strains cannot produce community-emergent metabolites
- Full-spectrum postbiotics may address both the missing emergent chemistry and the redox environment needed for resident bacteria to function
Think of your gut not as a list of species to be managed, but as a dynamic, interconnected conversation. The words being spoken — the metabolites being made — depend entirely on who's in the room and the chemistry they're swimming in.
Stay curious. Take care of your ecosystem.
References
- Kamrad S, Aulakh SK, Mozzachiodi S, et al. Interspecies interactions drive bacterial proteome reorganization and emergent metabolism. Nat Ecol Evol. 2026. https://doi.org/10.1038/s41559-026-03030-4
- McDonald JAK, Mullish BH, Pechlivanis A, et al. Inhibiting Growth of Clostridioides difficile by Restoring Valerate, Produced by the Intestinal Microbiota. Gastroenterology. 2018;155(5):1495–1507. https://doi.org/10.1053/j.gastro.2018.07.014
- Ott SJ, Waetzig GH, Rehman A, et al. Efficacy of Sterile Fecal Filtrate Transfer for Treating Patients With Clostridium difficile Infection. Gastroenterology. 2017;152(4):799–811. https://doi.org/10.1053/j.gastro.2016.11.010
- Rooks MG, Garrett WS. Gut microbiota, metabolites and host immunity. Nat Rev Immunol. 2016;16(6):341–352.
- Visconti A, Le Roy CI, Rosa F, et al. Interplay between the human gut microbiome and host metabolism. Nat Commun. 2019;10(1):4505.
- Roager HM, Licht TR. Microbial tryptophan catabolites in health and disease. Nat Commun. 2018;9(1):3294.
- Pepke ML, Hansen SB, Limborg MT. Unraveling host regulation of gut microbiota through the epigenome-microbiome axis. Trends Microbiol. 2024;32(12):1229–1240.
This post accompanies the Lit Review Friday episode of Learn Something with Thaena.