The Dog In Front Of You
Most owners who land on a piece like this are not, in their own minds, here for science. They are here because something is not right with their dog, or because something is right and they want it to stay that way. The itch that comes back. The stool that has not been firm in months. The stiffness on a cold morning that was not there last winter. The vet has done the sensible things, ruled out the worst, and the dog is now, in clinical terms, fine. Just not the dog they were.
I have spent almost a decade thinking about what the everyday food in a dog’s bowl is doing to that dog at a level that no clinical visit ever sees. The answer, drawn out across twenty years of canine nutrition science and brought into sharp focus by a paper that landed in January 2026, is that it is doing more than almost any of us were taught to believe.
This article is the story of two papers. The first was published when I was nowhere near a canine nutrition career. The second was published this winter, while I was preparing the second formulation review of the year for Bonza. Read separately, each is a meaningful contribution to canine nutrition science. Read together, they form a single twenty-year intellectual arc, and they answer a question I get asked almost every week. Why does what my dog eats matter as much as you say it does?
The short answer is that diet, in dogs, sits a great deal closer to gene expression and disease risk than most owners are told. The long answer is what follows.
The 2006 Manifesto. What Swanson Saw.
In May 2006, the Journal of the American Veterinary Medical Association published a paper by Kelly S. Swanson titled Nutrient-gene interactions and their role in complex diseases in dogs.¹ Swanson was, and remains, one of the most credentialled animal-nutrition academics in North America. The paper is short by modern review standards. It is also, in retrospect, a manifesto.
Swanson’s argument ran along three lines.
The first was that nutrition does not simply pass through the dog. It interacts with the dog’s genome. Specific nutrients, and the metabolites that arise from them, modulate the expression of genes that control inflammation, immune signalling, lipid handling, and a long list of other processes that determine whether a dog’s body is operating in a state of repair or in a state of low-grade chronic damage. The dog you put food into in the morning is not the same dog at the level of gene expression by the evening, depending on what the food was. This was not a new idea in human nutrition science by 2006. In canine nutrition, where the regulatory and commercial focus had long been on adequacy rather than modulation, it was a reframing.
The second was that complex chronic disease in dogs, the obesity, the dermatological presentations, the diabetes, the cancers, the inflammatory bowel conditions, the cognitive decline, was not a set of separate problems with separate causes. It was, at least in part, a set of expressions of long-running nutrient-gene interactions. Swanson’s phrase for the field that would have to study this was canine nutrigenomics. The field did not yet exist as a coherent research community in 2006. Swanson was calling it into being.
The third was that the work would have to be canine-specific. Researchers had been borrowing from human and rodent literature for decades, on the reasonable assumption that mammalian biology shares enough common ground for the borrowing to be safe. Swanson argued that this was insufficient. Dogs have their own genome, their own dietary history through co-evolution with humans, their own gut, their own breed-level genetic variation, and almost certainly their own nutrient-gene interaction profiles. Canine nutrigenomics, he argued, would have to be built as a discipline in its own right, with its own funding, its own datasets, its own tools.
It is worth pausing on what Swanson could not yet test in 2006. The technologies that would later make canine-specific nutrigenomics a working discipline, high-throughput sequencing at affordable scale, long-read sequencing, metagenome-assembled genome reconstruction, the bioinformatics infrastructure to handle hundreds of millions of reads, were either prohibitively expensive, in their infancy, or not yet invented. The PCR-based microbiome studies that were technically feasible at the time produced taxonomic surveys that captured roughly a quarter of the bacteria actually present in the canine gut. The deeper functional question, what are these bacteria doing, was largely out of reach.
Swanson, in other words, made a case in 2006 that the science of the field was not yet capable of fully answering. He argued that it should be, and that it would be. The paper closed with an explicit call for the canine nutrigenomic research community that the next twenty years would have to build.
The Twenty-Year Bridge. How The Field Caught Up.
Between 2006 and 2026 the technological gap closed. Three threads matter here for the dog in your house.
Sequencing got cheaper, deeper, and longer. The first canine genome assemblies in the early 2000s cost orders of magnitude more than they cost today. By the late 2010s, shotgun metagenomic sequencing of a faecal sample, the technique that lets researchers reconstruct not just what bacteria are present but what their genes can do, was within the reach of well-funded institutional research groups. By the mid-2020s, long-read sequencing platforms (Oxford Nanopore, PacBio HiFi) made it possible to assemble bacterial genomes from a complex mixed sample with the accuracy that taxonomic and functional cataloguing requires.
The microbiome reframe arrived. By the early 2010s, human microbiome research had reframed the gut from a passive food-processing tube into a metabolic organ in its own right. The intestinal microbiota was, increasingly, understood as a regulatory partner with the host, producing short-chain fatty acids, modulating bile acid pools, training immune development, and signalling to the brain through the vagus nerve and through circulating metabolites. The eight gut-organ axes Bonza now uses as our scientific spine, the Gut-Skin axis, the Gut-Brain axis, the Gut-Joint axis, the Gut-Liver axis, the Gut-Immune axis, the Gut-Heart axis, the Gut-Longevity axis, and the Gut-Oral axis, are the structured expression of the microbiome reframe applied to dogs. Each will get its own pillar article in this Health Hub.
Canine-specific microbiome work matured. Early canine microbiome studies relied heavily on human reference databases for taxonomic assignment, which is roughly equivalent to using a French dictionary to translate Welsh. Recognised gaps drove specific projects, including foundational work by Pilla and Suchodolski on canine microbial communities and dysbiosis, and by Jergens, Heilmann, and others on chronic enteropathy and microbial signatures.²,³ The gradual accumulation of canine-specific data showed two things. The shares broad-stroke similarities with the human and other mammalian microbiomes. The species-level composition does not. Akkermansia muciniphila, for example, the species that has driven a substantial fraction of human gut-health commercial activity, is largely absent from the canine gut and behaves as a contextual species when occasionally present, not as a foundational community member.⁴
By 2025, then, the canine nutrigenomic field that Swanson called for in 2006 had its tools, its theoretical framework, and its growing dataset. What it lacked was a single, comprehensive, canine-specific reference that mapped the at functional resolution across enough dogs to be representative. That gap is what the second paper closed.
The 2026 Catalogue. What Castillo-Fernandez And The Waltham Team Found.
In January 2026, Microbiome published Waltham catalogue for the canine gut microbiome: a complete taxonomic and functional catalogue of the canine gut microbiome through novel metagenomic based genome discovery by Juan Castillo-Fernandez, Rachel Gilroy, and colleagues at the Waltham Petcare Science Institute, the Mars Petcare research group based in Leicestershire.⁵
The shape of the work is straightforward. Faecal samples were collected from 107 healthy dogs across the United States and Europe, drawn from both kennel and home-living populations, providing 501 samples in total. The samples were processed with both long-read and short-read metagenomic sequencing. The reads were assembled into 5,753 metagenome-assembled genomes, which were then dereplicated to 1,031 distinct strains, of which 982 were new to canine science. From this set the team identified 240 core species that account for approximately 83 per cent of the canine gut microbiome by abundance in an independent validation dataset. Among the new findings were 89 novel species and 10 novel bacterial genera previously unknown anywhere.
The technical framing matters less than what the catalogue lets us say.
Before this paper, when researchers tried to map a canine faecal sample against the available reference databases, they could typically classify roughly 25 per cent of the reads. Three out of four bacterial sequences in a healthy dog’s gut belonged to organisms that had not been catalogued. The Waltham catalogue lifts the mapping rate to as high as 95 per cent in their validation work. We have, for the first time, a near-complete picture of who is in the .
The functional findings are more important than the taxonomic ones for nutritionists. The catalogue reports an average of 71 carbohydrate-active enzymes (CAZymes) per bacterial species in the canine gut. CAZymes are the molecular tools bacteria use to degrade dietary fibre and other carbohydrates into the metabolites that influence the dog’s biology. The number is large, and it is large because the canine microbiome is configured for substrate degradation at scale. Specifically, the catalogue documents that approximately 75 per cent of the catalogued species carry chitin-degrading capability, 36 per cent xylan-degrading capability, 36 per cent cellulose-degrading capability, and 22 per cent starch-degrading capability. Most of these are plant-cell-wall components or insect-derived structural carbohydrates. A microbiome with this profile is a microbiome built to ferment plant substrates and the hard-fibre components of insect biomass, not a microbiome built around animal protein degradation.
Two further findings sit in the same direction. The catalogue documents widespread bacterial capacity for short-chain fatty acid production, in particular butyrate and propionate, which are the metabolites most consistently associated with intestinal-barrier integrity, regulatory T-cell development, and the broader inflammatory tone of the host. The catalogue also documents widespread bacterial capacity for the biosynthesis of essential amino acids, lysine prominent among them, and for bile acid metabolism. The microbiome is producing nutrients the host cannot synthesise on its own, including amino acids that have historically been used as the strongest argument against plant-based feeding of dogs.
The catalogue also confirms what canine-specific microbiome work had been suggesting. The canine gut microbiome is genuinely distinct from the human. It is not a smaller version of the human microbiome and the species that dominate human gut-health commercial conversations are mostly not the species that dominate the canine gut. Akkermansia muciniphila is, again, essentially absent. The two most abundant of the catalogue’s newly discovered canine species were previously unknown to science. Strain specificity in canine probiotics is not, as the marketing might suggest, a refinement. It is a biological necessity. The strain has to fit the gut.
A final observation, easy to miss at first read. All 89 novel species characterised in the catalogue are commensal organisms. None are pathobionts. The functional tilt of the resident canine microbiome, in healthy dogs, is constructive rather than destructive. The implication is that the question for nutrition is not how to displace bad bacteria but how to feed and protect the good ones we already share the household with.
The most consequential nutrigenomic lever available to dog owners is dietary composition itself. Plant-based diets, properly formulated, drive the fibre fermentation, short-chain fatty acid production, and microbial metabolite signalling that constitute the diet-genome interface. For the full peer-reviewed evidence on plant-based canine nutrition and its measurable effects on microbiome function and downstream health outcomes, see the Bonza evidence review on plant-based dog food research.
The Marriage. How The Catalogue Validates The Manifesto.
This is where reading the two papers together does work that reading either alone does not.
Swanson’s 2006 manifesto argued that diet modulates gene expression in dogs through nutrient-gene interactions, and that the resulting modulation is causally important to complex chronic disease. He could not yet test the proposition at scale because the gut microbiome, the largest single biological surface across which diet meets host genome, was substantially uncharacterised in dogs. The bacteria that produce the metabolites that signal to the host’s genes were, in 2006, opaque. Swanson knew the conversation was happening. He could not, with the tools available, listen in.
Castillo-Fernandez 2026 is the listening device. The catalogue does not test specific nutrigenomic hypotheses, and the authors are properly cautious about not overreaching. What the catalogue does is provide the comprehensive, canine-specific reference that any future nutrigenomic study now starts from. We can now ask, of a given dog and a given diet, which species are present, what enzymatic tools do they carry, what metabolites do they produce, and how does the dog’s gene expression respond. Twenty years on from Swanson’s call, the field has the tools to do the work.
Read together, the two papers also validate four propositions Bonza had already locked into our brand foundation before either paper was the active reference text in our chat.
The first is that nutrition is the single biggest daily input for canine health. The 2006 paper said it; the 2026 catalogue describes the mechanism by which it is true. The dog cannot fully process its own diet without microbial assistance because the canine genome encodes few of the enzymes that release calories and metabolites from complex carbohydrates. The microbiome does that work, and the microbiome is shaped, at every meal, by the diet. Diet is the daily lever.
The second is that fibre diversity is the active mechanism, not fibre quantity. Different bacterial species in the catalogue carry different CAZyme families and degrade different substrates. A diet that supplies one or two fibre sources at high concentration feeds the species that can use those fibres and starves the rest. A diverse microbiome is the product of diverse substrates. We arrived at this proposition through the Bonza Fibre Review and a long reading of Le Bon and colleagues among others; the catalogue puts an explicit functional resolution under it.
The third is that the Biotics Triad, our framework for prebiotics that feed, probiotics that produce, and postbiotics that deliver, reflects actual microbiome biology and not just a marketing structure. The catalogue’s widespread documentation of CAZyme-rich species, of short-chain fatty acid production, and of bile acid metabolism is, in three sentences, the Triad. Prebiotic substrates are degraded by probiotic-class organisms into postbiotic outputs, and the catalogue describes all three legs of that cycle in the species it characterises. Calsporin® (Bacillus Velezensis DSM 15544), EFSA-authorised), TruPet™ and Lactobacillus helveticus HA-122 are the named branded actives we use to make the Triad operative in our formulations; the catalogue is the population-scale picture of why such actives matter.
The fourth is that plant-based, in dogs, is not an ethical statement we layer onto the food. It is the formulation outcome of building a food for the gut. The canine microbiome is configured for plant-substrate metabolism. Its bacteria synthesise the essential amino acids that the conventional argument against plant-based feeding claims dogs cannot get without animal protein. Building a food that maximises fibre diversity, that avoids the most common dietary disruptors, and that supplies the substrates the canine gut is built to ferment, brings you to a plant-based formulation by way of mechanism, not by way of premise. The catalogue is the most direct empirical defence of that reasoning chain produced to date.
What This Means For The Food In Your Dog’s Bowl
I want to be careful here. A Health Hub article is a place to think out loud about the science, not to write product copy. The reasonable expectation, having read this far, is for me to tell you what to do about it for the dog you have.
Three things follow.
The first is that the food you choose, every day, twice or three times a day, across the dog’s life, is the dominant input to a regulatory system that touches every organ in the dog’s body. This is not a marketing claim. It is the implication of putting Swanson 2006 next to Castillo-Fernandez 2026 and looking honestly at the picture. Episodic interventions, the targeted supplement for the flare, the prescription course for the bout, are necessary in their place. They are not the system. The bowl is the system.
The second is that fibre diversity, in the food itself, is closer to the active mechanism than fibre quantity is. A dog food that is high in a single fibre source is doing less for the microbiome than a food that supplies a structured panel of fibres, oligosaccharides, beta-glucans, plant-cell-wall components, and resistant starch fractions. We will return to this in detail in the dedicated fibre diversity pillar article. For now, the practical move is to read ingredient panels with an eye for diversity rather than for headline fibre percentage.
The third is that probiotics worth including are the ones that have been clinically researched in the species they are meant to colonise, at the strain level, with named identifiers. The catalogue’s strain-specificity finding is the explicit version of what we have been saying about Calsporin® for years. A probiotic strain has to be the right strain for the dog’s gut. Generic Lactobacillus or Bifidobacterium claims, with no strain-level identifier, are claims about a category, not about the bacterium that is supposed to do work in your dog. Bonza names the strain on every product where one is included, including the EFSA authorisation status, because the catalogue is the kind of paper that makes that level of specificity necessary.
If I were writing a checklist for a friend, it would not be three commands. It would be three questions. Does the food I am buying supply diverse fibre sources, or one or two at concentration? Are the named probiotic strains I am paying for actually identified, or just genus-and-species labelled? And does the brand I am buying from understand the gut microbiome as the central system of canine health, or as one feature among many?
Those are the questions Bonza was built to answer in the affirmative. I will not pretend to neutrality on that point.
Why This Is Why Bonza Exists
I came into canine nutrition the long way. I added the Diploma in Canine Nutrigenetics, with Distinction, after the Diploma in Canine Nutrition because the more I learned about how diet shapes the dog, the more I needed to understand how diet shapes the dog at the level of gene expression. The discipline I trained in for that second qualification is the discipline Swanson called into being in 2006. The catalogue Castillo-Fernandez and the Waltham team published in 2026 is the field’s coming of age, in the same year I am formulating Bonza’s next round of products against it. The intellectual lineage runs cleanly from Swanson’s manifesto through the credential I trained in to the food we make.
That is the reason Bonza is a canine gut health company expressed through nutrition, rather than a plant-based dog food brand that talks about gut health. It is the reason every Bonza product is anchored to the Biotics Triad and to one or more of the eight gut-organ axes. It is the reason the Founder credential line on this page reads as it does. It is the reason we will not pretend, ever, that a dog’s diet is a peripheral concern of canine medicine.
The dog in front of you is, at the level of cellular signalling, what they have been fed. The science took twenty years to be in a position to say that with confidence. We have spent that time arriving at the food we now make.
One Gut. Whole Dog. Fibre Diversity Over Fibre Quantity. Plant-Based By Consequence, Not By Trend.
This article is for informational purposes only and does not constitute veterinary advice. Always consult a qualified veterinarian before making changes to your dog’s diet or supplement regimen.
Glendon Lloyd, Dip.CN, Dip.CNG. Founder, Bonza.
References
- Swanson KS. Nutrient-gene interactions and their role in complex diseases in dogs. Journal of the American Veterinary Medical Association. 2006;228(10):1513-1520. doi:10.2460/javma.228.10.1513. PMID: 16677119.
- Pilla R, Suchodolski JS. The Gut Microbiome of Dogs and Cats, and the Influence of Diet. Veterinary Clinics of North America: Small Animal Practice. 2021;51(3):605-621. doi:10.1016/j.cvsm.2021.01.002. PMID: 33653538.
- Jergens AE, Heilmann RM. Canine chronic enteropathy: Current state-of-the-art and emerging concepts. Frontiers in Veterinary Science. 2022;9:923013. doi:10.3389/fvets.2022.923013. PMID: 36213409. PMCID: PMC9534534.
- Garcia-Mazcorro JF, Minamoto Y, Kawas JR, Suchodolski JS, de Vos WM. Akkermansia and Microbial Degradation of Mucus in Cats and Dogs: Implications to the Growing Worldwide Epidemic of Pet Obesity. Veterinary Sciences. 2020;7(2):44. doi:10.3390/vetsci7020044. PMID: 32326394. PMCID: PMC7355976.
- Castillo-Fernandez J, Gilroy R, Jones RB, Honaker RW, Whittle MJ, Watson P, Amos GCA. Waltham catalogue for the canine gut microbiome: a complete taxonomic and functional catalogue of the canine gut microbiome through novel metagenomic based genome discovery. Microbiome. 2026;14(1):25. doi:10.1186/s40168-025-02265-w. PMID: 41547860. PMCID: PMC12811905.