Four German shepherds in Dhaka produced fecal samples with no trace of any bacterium on the researchers' zoonotic watchlist. Four golden retrievers, sampled the same way in the same city, produced five. That gap, small and drawn from a handful of animals, is the kind of detail that makes the first microbial census of Bangladeshi pets worth reading carefully rather than quickly.
A team led by researchers at Sher-e-Bangla Agricultural University collected feces from 24 apparently healthy household pets in Dhaka City between July 2024 and June 2025: 12 dogs and 12 cats, four animals from each of six breed groups. Owners consented, and samples were caught immediately after defecation, sealed in sterile tubes and frozen. The team then extracted bacterial DNA and sequenced a stretch of the 16S ribosomal RNA gene, a genetic barcode that differs just enough between bacterial lineages to sort them into groups.
What came back was 2.27 million usable sequence reads, which sorted into 1,148 distinct genetic variants spread across 20 phyla and 258 genera. Five phyla accounted for more than 99 percent of everything sequenced. Firmicutes dominated in both animals, at roughly 71 percent of reads in dogs and 67 percent in cats. Proteobacteria, Actinobacteria, Bacteroidetes and Fusobacteria filled out the rest. These are the same broad groups other teams have found in dogs and cats elsewhere, which is reassuring: the Dhaka pets are not microbial outliers.
Cats carried slightly more diverse gut communities than dogs by all three measures the team used. But the difference was not statistically significant, and the authors say plainly why: 24 animals is a small sample, and gut microbiomes vary enormously from one individual to the next even within a breed. When the team plotted the samples by overall similarity, dogs and cats did not separate into clusters. Neither did breeds. Individual variation, they conclude, shapes these communities more than species or pedigree does.
That framing runs through the whole paper. Breeds differed in how much of each bacterium they carried, not in which bacteria they carried. Bacillus was most abundant in local-breed dogs; Lysinibacillus and Brevibacillus in German shepherds; Bifidobacterium and Blautia in Persian cats. Only seven genetic variants were shared across all six breed groups, while golden retrievers alone accounted for 311 unique ones.
The zoonotic question, handled cautiously
The part most readers will care about is the smallest part of the data. The team checked their sequences against bacteria previously linked to human infection and found ten such taxa, all at low read counts. Enterococcus cecorum, an emerging poultry pathogen that occasionally causes bone and abdominal infections in people, showed 133 reads in golden retrievers and 4 in local-breed dogs. Schaalia canis, associated with bite-wound infections, appeared 58 times in retrievers and 4 times in Persian cats. Campylobacter helveticus, which can cause gastroenteritis in humans, turned up only in Persian cats, with 41 reads. Sutterella wadsworthensis appeared only in domestic shorthairs. Set against the millions of reads from dominant gut bacteria, these are faint signals.
The authors are careful about what that means, and they repeat the caveat in the abstract, the methods, the discussion and the conclusion. Sequencing a gene fragment tells you a matching stretch of DNA was present. It does not tell you the bacterium was alive, that it was capable of causing disease, or that it could reach a human and infect them. Some of the taxa they found are normally mouth-dwellers, not gut residents; the team suggests these may simply be passing through after grooming, swallowed rather than resident.
Why it matters
Dhaka is one of the densest cities on earth, and pet ownership there is growing. Yet until now nobody had built even a baseline picture of what lives in the guts of the dogs and cats sharing those households. Earlier Bangladeshi work looked for specific culprits, finding Staphylococcus aureus, E. coli and Salmonella in cats. This study instead asks the open question: what is actually in there?
A baseline is useful precisely because it is boring. If a future outbreak traces back to household animals, or if antibiotic resistance starts spreading through urban pets, researchers now have a reference point for what a typical Dhaka pet gut looked like in 2024 and 2025. Without that, any later finding floats free of context.
The limits are real and the authors list them: modest sample size, one city, no information on the animals' diets, ages or veterinary history, no culture work to confirm anything was alive, and no negative controls to rule out contamination. Taken together, those constraints mean this paper is a starting point rather than a verdict. The authors' own recommendation is modest and sensible: routine surveillance of pet microbes, and better public awareness of how diseases move between animals and people.