Article https://doi.org/10.1038/s41467-025-65266-1 Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households in Malawi Catherine N. Wilson 1,2,3,4 , Patrick Musicha 2,3,5, Mathew A. Beale 3, Yohane Diness2, Oscar Kanjerwa2, Chifundo Salifu2, Zefaniah Katuah2,6, Patricia Duncan7, John Nyangu7, Andrew Mungu7, Muonaouza Deleza8, Lawrence Banda8, Lumbani Makhaza2, Nicola Elviss9, Christopher P. Jewell10, Gina Pinchbeck1, Nicholas A. Feasey 2,5,11, Eric M. Fèvre 1,12,14 & Nicholas R. Thomson 3,13,14 Diverse salmonellae have the potential to cause disease and may be carried asymptomatically within the intestine of many vertebrate species. The relative contribution of human, animal, and environmental hosts to the transmission of Salmonella is unknown within and between households in low-income set- tings, especially where humans and animals may live in close contact and sanitary infrastructure is often inadequate. Between November 2018 and December 2019, we isolated Salmonella spp. from thirty households in urban and rural locations in Malawi, sampling at three time points from the stool of humans, animals, and their household environment. Using whole genome sequencing and fine-resolution bioinformatic and phylogenetic analyses we found evidence of sharing of Salmonella species and strains between humans, animals and the environment, both within and between households. The intricate web of interconnected salmonellae within this ecosystem under- scores the importance of adopting a multi-faceted ‘One Health’ strategy when considering control of Salmonella in low-intensity agricultural systems. It is now well established that ~75% of emerging and re-emerging pathogens affecting humans worldwide are zoonotic, with the greatest disease burden affecting poorer and more marginalised populations in low- and middle-income countries1–4. The One Health concept recognises that the health of humans, domestic and wild animals and their wider environment are closely linked and interdependent5. Over the last decade, changes to global ecosystems, demographics, socio-cultural and economic factors have reinforced the interconnection between humans, animals and their environment and amplified the need for collaborative, multi- sectoral and transdisciplinary approaches to understand and opti- mise responses to these changes. This has been accompanied by substantial fluctuations in climate and ecosystem health, which are associated with extension of the ranges of non-endemic pathogens Received: 16 November 2023 Accepted: 8 October 2025 Check for updates 1Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool, Liverpool, UK. 2Malawi Liverpool Wellcome Programme, Blantyre, Malawi. 3Wellcome Sanger Institute, Hinxton, UK. 4Department of Veterinary Medicine, University of Cambridge, Cambridge, UK. 5Department of Clinical Sciences, Liverpool School of TropicalMedicine, Liverpool, UK. 6Malawi Adventist University, Blantyre,Malawi. 7Department of AnimalHeath andProduction,Ministry of Agriculture and Food Security, Blantyre, Malawi. 8Lilongwe University of Agriculture and Natural Resources, Bunda, Lilongwe, Malawi. 9UK Health Security Agency, London, UK. 10School of Mathematical Sciences, Lancaster University, Lancaster, UK. 11The School of Medicine, University of Andrews, St, Andrews, UK. 12International Livestock Research Institute, Nairobi, Kenya. 13Faculty of Infectious and Tropical Diseases, The London School of Hygiene and Tropical Medicine, London, UK. 14These authors contributed equally: Eric M. Fèvre, Nicholas R. Thomson. e-mail: cnw25@cam.ac.uk; eric.fevre@liverpool.ac.uk Nature Communications | (2025) 16:9703 1 12 34 56 78 9 0 () :,; 12 34 56 78 9 0 () :,; http://orcid.org/0000-0002-7150-3081 http://orcid.org/0000-0002-7150-3081 http://orcid.org/0000-0002-7150-3081 http://orcid.org/0000-0002-7150-3081 http://orcid.org/0000-0002-7150-3081 http://orcid.org/0000-0001-7780-2173 http://orcid.org/0000-0001-7780-2173 http://orcid.org/0000-0001-7780-2173 http://orcid.org/0000-0001-7780-2173 http://orcid.org/0000-0001-7780-2173 http://orcid.org/0000-0002-4740-3187 http://orcid.org/0000-0002-4740-3187 http://orcid.org/0000-0002-4740-3187 http://orcid.org/0000-0002-4740-3187 http://orcid.org/0000-0002-4740-3187 http://orcid.org/0000-0003-4041-1405 http://orcid.org/0000-0003-4041-1405 http://orcid.org/0000-0003-4041-1405 http://orcid.org/0000-0003-4041-1405 http://orcid.org/0000-0003-4041-1405 http://orcid.org/0000-0001-8931-4986 http://orcid.org/0000-0001-8931-4986 http://orcid.org/0000-0001-8931-4986 http://orcid.org/0000-0001-8931-4986 http://orcid.org/0000-0001-8931-4986 http://orcid.org/0000-0002-4432-8505 http://orcid.org/0000-0002-4432-8505 http://orcid.org/0000-0002-4432-8505 http://orcid.org/0000-0002-4432-8505 http://orcid.org/0000-0002-4432-8505 http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65266-1&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65266-1&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65266-1&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65266-1&domain=pdf mailto:cnw25@cam.ac.uk mailto:eric.fevre@liverpool.ac.uk www.nature.com/naturecommunications such as dengue virus, West Nile virus and Vibrio cholerae6–8. In the future, pathogen spillover from animal to human populations and vice versa may occur more frequently, particularly in the face of increasing urbanisation9–11. In addition, antimicrobial resistance (AMR) has emerged as one of the leading public health threats of the twenty-first century12. In 2019 it was estimated that the highest proportion of the global burden of deaths owing directly to drug- resistant infections occurred in sub-Saharan Africa, a consequence of prevailing levels of poverty leading to inadequate investment in sanitation and healthcare infrastructure, a high overall burden of infectious diseases, poor regulation of antimicrobial use and lack of alternatives to effective antimicrobials13. Salmonella spp. are an idealmodel to investigate bacterialflux in a One Health context since they include several globally relevant pathogens and can also be carried asymptomatically within the intes- tine of a wide variety of vertebrate species, including animals reared for meat and egg production. Further, salmonellae can exist stably within the environment at ambient conditions for long time periods14,15. Worldwide, Salmonella spp. are estimated to cause 78.7 million human cases of gastroenteritis annually, with 59,100 deaths and 4.1 million disability adjusted life-years16. This estimate is lower on theAfrican continent, where forty-sixpercent of cases of illness caused by non-typhoidal Salmonella (NTS) are attributed to exposure through the foodborne pathway17. In sub-Saharan Africa, NTS has become one of the most common causes of invasive bacterial bloodstream infec- tion in humans over the last 40 years18–20. The most recent estimate states that invasive NTS accounts for 29.5% of cases of bloodstream infections in Africa, carrying an average case-fatality rate of 20.6%21,22. Most invasive NTS disease in Malawi can be attributed to the specific strain S. Typhiumurium ST313, with S. Enteritidis ST11 the secondmost common strain reported23–28. Alterations in the demographics and socio-economic status of populations may lead to changes in animal husbandry and man- agement practices, increase proximity to and contact with humans for both domestic and wild animals and consequently altered inci- dence of endemic zoonotic disease. Population growth and urba- nisation are fuelling a significant increase in the demand for meat production within sub-Saharan Africa29. Within this sector animals may often be reared and slaughtered in conditions with little pro- vision for biosecurity, particularly within low-intensity production systems around households30. Domestic, livestock and poultry ani- mals belonging to the household are often kept within the house- hold perimeter, and in some settings, humans and animals sleep under the same roof[31. Loosely structured waste management sys- tems for human and animal species within these environments offer the opportunity for environmental faecal contamination and con- sequent exposure of peri-domestic wildlife species such as geckos, wild birds and rodents to human and animal faecal residues within the household environment32. These factors increase the potential for transmission of Salmonella spp. and AMR determinants carried by these bacteria. To investigate sharing of Salmonella spp. within the extended household, considering humans, animals and the household environ- ment, we conducted a prospective longitudinal surveillance study of thirty households from a high-density urban and a rural setting in Malawi. We collected human and animal faecal samples as well as environmental samples from frequently contacted household sur- faces. Using whole genome sequencing we determined the relation- ships between Salmonella isolates, finding evidence of extensive sharing of Salmonella spp. between humans, animals and the envir- onment, both within and between households. The insights gained into the dynamics of Salmonella spp. dissemination will inform future considerations for enhanced biosecurity and surveillance within such settings. In addition, our results support a broader understanding of the risks and drivers of pathogen emergence across interfaces. Results Description of the participating households Between 19th November 2018 and 16th December 2019 thirty house- holds were recruited in two study sites in Malawi. Fifteen houses were recruited in Ndirande, an informal urban settlement in Blantyre, and a further fifteen in the rural area Chikwawa (Supplementary Fig. 1). Each household was visited three times, the second visit taking place ~2 months after the first visit (median interval between first and second visit 63 days, range 35–140 days), followed by a third visit roughly 6 months after the first visit (median interval between first and third visit 204 days, range 120–330 days). In total 411 stool samples were collected from 184 humans. For each household, at least one envir- onmental sample (totaln = 646,medianper household n = 7, range per household 1–19) was collected from areas of high human-human, animal-animal, or human-animal contact such as dwelling surfaces inside the household and animal pens, food preparation and water storage areas, dirty clothing, beds and latrine areas. All households kept at least one livestock, domestic or poultry animal and we col- lected 1023 animal stool samples from a range of livestock (cattle, sheep, goats, pigs), domestic animals (dogs, cats and guinea pigs), poultry (chicken, ducks, doves, guinea fowl, jungle fowl) and peri- domestic wildlife (rodents, geckos and wild birds). In total, 2080 individual samples were collected, 965 (46.4%) fromNdirande and 1115 (53.6%) from Chikwawa. Samples were cultured and the presence of Salmonella was screened for by PCR for ttr. PCR for ttr was positive in 233/2080 samples (11.2%) (Supplementary Fig. 2, Supplementary Table 1−3). In total 87/965 (9.0%) of samples from Ndirande and 146/1115 (13.1%) of samples from Chikwawa were positive for Salmonella spp. Distribution of Salmonella genomes Whole genome sequencing was used to confirm the presence of Salmonella spp. We performed DNA extraction and whole genome sequencing of two-three ttr positive colony picks per sample to enable us to assess multi-serovar carriage and within-host diversity. We therefore sequenced 403 isolates in total, which were taken from a total of 214 samples. After quality assessment of sequence data, 227 genomes were identified as Salmonella genomes, passing quality thresholds and were subsequently included in our analy- sis (Fig. 1a). These 227 genomes originated from 111 discrete samples. On examination, this collection of 227 genomes contained a number of ‘identical’ bacterial isolates (i.e. identical bacterial isolates aredetectedwithin one sample, collected at the samehousehold at the same timepoint). One isolate from a pair of identical isolates was removed (de-duplicated) from this collection, leaving a total of 131 individual isolates (an individual isolate is a single Salmonella isolate of each serovar or sequence type from each individual sample). These 131 individual isolates were collected from 111 samples, therefore more than one individual Salmonella isolate was identified from twenty of these samples (within-host diversity). The 111 samples originate from 81 animal stool samples (73.0%), 16 environment samples (14.4%) and 14 human stool samples (12.6%). At least one Salmonella genome was generated from 25/30 households in the study; 14/15 (93.3%) house- holds sampled in Chikwawa and 11/15 (73.3%) households sampled in Ndirande (Fig. 1a). In total 94/1115 (8.4%) of samples collected from Chikwawa and 17/965 (1.8%) of samples collected in Ndirande contained Salmo- nella. Of the 111 samples, 53/111 (47.7%) were collected from samples taken during the first visit to the household, 32/111 (28.8%) during the second visit and 26/111 (23.4%) during the third visit. Salmonella genomes were generated from 6/30 (20%) households at all three visits, 11/30 (36.7%) households at two visits, 8/30 (26.7%) house- holds at one visit. Considering human, animal and environmental samples as separate ‘compartments’, Salmonella genomes were Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 2 www.nature.com/naturecommunications generated from all three compartments at 5/30 (16.7%) households, two compartments at 9/30 (30%) households, and one compartment at 11/30 (36.7%) households. We used the real-time analytic and genomic epidemiology plat- form Pathogenwatch33 to delineate Salmonella genomes into species, subspecies and serovars. Isolates in the collection were drawn from two subspecies of Salmonella enterica, comprising 125 (55.1%) isolates of Salmonella enterica subspp. enterica (S. enterica), and 102 (44.9%) isolates of Salmonella enterica subspp. salamae (S. salamae) (Fig. 2). Among these subspecies, 56 serovars were identified within the col- lection, 26 S. enterica and 30 S. salamae (Supplementary Fig. 3). The 227 Salmonella genomes were identified from human stool (n = 25, 11.0%), animal stool (n = 171, 75.3%) and environmental samples (n = 31, 13.7%)(Fig. 2). There was no significant difference (chi squared test, ρ = 0.49) between the number of S. enterica and S. sal- amae genomes collected from each host category. Salmonella was detected within the stool of twelve different animal species (chicken, duck, dove, guinea fowl, gecko, wild bird, rodent, cockroach, cattle, pig, goat, dog). A surprising number of S. salamae genomes were detected within the study, offering a chance to investigate this little documented subspecies of Salmonella. Of the thirty-two Salmonella genomes detected in total from Ndirande, twenty-one (65.6%) were S. enterica and eleven were S. salamae (34.4%). Of the 195 genomes in total from Chikwawa, 104were S. enterica (53.3%) and 91 were S. salamae (46.7%). There was no significant difference between the number of S. enterica Fig. 1 | Number and diversity of Salmonella genomes isolated from each household. a Total number of good quality genomes isolated from human, animal hosts or the environment at each household during the study (n = 227). b Diversity of Salmonella sequence types isolated fromeachhousehold andeachhost category (human, animal or the environment) within the collection of 227 genomes. Coloured bars represent each individual sequence type of which 64 were detected within the collection of 227 genomes. CHH= household located in Chikwawa study site, NHH= household located in Ndirande study site followed by an identifying number of the household. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 3 www.nature.com/naturecommunications and S. salamae genomes detected at each study site (chi-squared test, ρ = 0.15) demonstrating that the two subspecies of Salmonella were equally represented across both sites. Inferring the distribution of plasmids, antimicrobial resistance determinants and virulence genes within the collection Multidrug resistance is an important concern for the treatment of invasive and noninvasiveNTSdisease in humans and animals34,35. Given the previously documented increase in AMR in Salmonella in this setting36, we investigated the distribution of AMR determinants, plas- mids and virulence genes.MOB-suite37 was used to identify a total of 85 plasmid replicons in 72/227 genomes. Eight different plasmid replicon types were indentified across 61 plasmids, of which IncFII was themost commonly identified (20/227, 8.8% genomes) followed by IncFIB (16/ 227, 7.0% genomes). In 12 genomes, we found two co-occurring plas- mid replicon types, and for a single genome we found three co- occurring plasmid replicons, all located on different contigs. For the remaining 24/85 (28.2%) putative plasmids, we were unable to assign a Fig. 2 | Phylogenetic tree of the 227 genomes within the collection to show the diversity of Salmonella present within each household and amongst the dif- ferent hosts sampled. The different coloured tree branches represent the two different Salmonella subspecies present in the collection. The tips of the tree are coloured according to the host species from which the genome was isolated. The household from which each Salmonella genome was isolated (identified at the top of the figure) is depicted by a filled-circle, coloured according to study site. Coloured bars linking the Salmonella genomes (filled-circles), also coloured by study site, join salmonellae identified within a single household together. Geno- typic antimicrobial resistance determinants detected displayed in heatmap on the right of the figure. Shades of blue depict the presence of an AMR determinant, white depicts no AMR determinant detected within the genome. fosA7 and fosA7.7 have been visualised together as fosA7. gyrB, as a nonsynonymous mutation, is not visualised here. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 4 www.nature.com/naturecommunications plasmid replicon type; 20 of these were detected within S. salamae isolates (Supplementary Fig. 4). We used abricate38 to infer the presence of known virulence genes, identifying 126 different virulence genes within the collection, 40 (32%) of which were uniformly present in all 227 genomes. Thirty- one virulence genes (25%) were only detected within S. enterica, of which most were predicted to encode Type 3 Secretion System pro- teins. Virulence genes were carried by 23/85 plasmids (27.1%). None of these plasmids were conjugative, 11/23 (47.8%) were mobilisable (require a helper plasmid for conjugation to occur). Importantly, given the potential for horizontal transmission of plasmids between salmo- nellae, in this collection none of the plasmids which carried virulence genes also carried AMR determinants. We found four different types of AMR determinants across 47 genomes (20.7%) (fosA7, fosA7.7, qnrB19 and gyrB19 non-synonymous mutation). fosA7 and fosA7.7 confer resistance to fosfomycin (not confirmed phenotypically here), while qnrB19 confers resistance to quinolones (and low level fluoroquinolone resistance, confirmed phenotypically for 3/4 genomes carrying qnrB19). No genomes con- tained more than one AMR determinant. The most common AMR determinant was fosA7, detected in 36/227 genomes (Supplementary Fig. 4). Both fosA7 and fosA7.7 were located on chromosomal contigs. Four qnrB19 genes were identified on plasmid contigs, all identified as plasmid rep cluster 2355, within S. Typhimurium ST19. These four isolates demonstrated phenotypic resistance to nalidixic acid. There was a single nonsynonymous SNPmutationofgyrB (S464F) detected in each of four genomes, all present on chromosomal contigs of S. Enteritidis. These four isolates carrying the gyrB mutation were asso- ciated with intermediate phenotypic resistance to nalidixic acid. Genetic diversity of salmonellae genomes To assess the genetic diversity of the isolates in our collection, we determined multi-locus sequence types (MLST) for all Salmonella genomes, finding 64 different Salmonella sequence types (STs) across the two subspecies (Fig. 1b). Of these, only 6 STs (9.4% of genomes) were present in both Chikwawa and Ndirande, indicating a lack of sharing between the two sites, whilst 50 STs (78.1% of genomes) were present only in rural Chikwawa, and 8 STs (12.5% of genomes) were present only in urban Ndirande (Fig. 1). Across both sites, the median number of STs detected within each household was three, but this was higher for Chikwawa (n = 5, range 0–15) than for Ndirande (n = 1, range 0–3). The median number of S. enterica STs collected from each household in Chikwawa was 2 (range 0–7), that of S. enterica in Ndir- ande was 1 (range 0–2). By contrast, the median number of S. salamae STs in the whole collection was 1 ST from each household, the median number of S. salamae isolates collected from households in Chikwawa was 2 (range 0–9) and the median number of S. salamae isolates col- lected from households in Ndirande was zero (range 0–2). Phylogenomic distribution of salmonellae We used Panaroo39 to infer a pangenome consisting of 11,964 unique genes (coding sequences), of which 3429 genes were defined as core (present in 98% of genomes), representing 28.7% of the pangenome, with the remaining 8535 (71.3%) genes forming the accessory genome. A core gene phylogeny (Fig. 2) was inferred to determine genomic relatedness of genomes collected within and between households. Pairwise SNP distances were calculated using the core gene alignment of the entire collection of 227 genomes (total alignment length 3,065,105 bps, 241,674 (7.9%) variable positions), and additionally using two species-specific core gene alignments for 125 S. enterica subsp. enterica genomes (total alignment length 3,240,911, 133,718 (4.1%) variable positions) and 102 S. enterica subsp. salamae genomes (total alignment length 3,239,471, 100,418 (3.1%) variable positions) (Supplementary Figs. 5 and 6). Across our entire collection of 227 genomes, 1841/25,651 (7.2%) sample pairs originated from within the same household and 23,810/25,651 (92.8%) sample pairs were from different households (Table 1). The median number of pairwise SNPs between any two samples amongst the whole collection was 36,847 SNPs using the core gene alignment of 227 genomes. There was a significant difference between the number of SNPs between pairs of genomes detected within the same household (median pairwise SNP distance = 34,931 SNPs), compared to the number of SNPs between pairs of genomes originating from different households (median pairwise SNPdistanceswas86,956SNPs; KruskalWallis test,p <0.001). We examined the pairwise SNP distances present within each Salmonella subspecies amongst samples originating from the same or different households (Table 1) and observed an extremely high degree of diversity within the collection as a whole. We also found a small percentage of identical genomes with 0 pairwise SNPs using the core gene alignment (Table 1). Sharing of Salmonella between hosts We used subspecies-specific (S. enterica and S. salamae) pairwise SNP distances to investigate whether genetically related Salmonella were present within two or more epidemiologically linked hosts within the study. Based on the relatively large number of close genomic rela- tionships between samples from the same household described above (Table 1) we selected a conservative approach, considering genomes separated by zero pairwise SNPs isolated from different hosts and sources to be putative ‘shared’ pairs, either by direct transmission or recent acquisition from a common source. As described earlier, the final collection of 227 genomes consist of a collection of 131 individial isolates collected from a total of 111 samples, as two-three colony picks were submitted from each sample to assess within host diversity. In order to investigate putative sharing of genomes between hosts and to minimise bias owing to within-sample repeat sampling, we dedupli- cated identical (0 SNPs) pairwise SNP distance measurements of Sal- monella genomes from the same individual sample taken at the same household at the same visit to the household (Supplementary Figs. 5 and 6). This left 20 pairs of genomes sampled from different hosts with a pairwise SNP distance of 0 SNPs (Fig. 3). Ten of these genome pairs were S. enterica and ten S. salamae. This represented 0.12% (0.07–0.2%) of the total number of S. enterica genome pairs (n = 7750 pairs) and 0.19% (0.11–0.36%) of the total number of S. salamae gen- ome pairs (n = 5151 pairs). Eleven (11/20, 55%) of these genome pairs occurred within household and 9 (9/20, 45%) occurred between households (Table 2, Supplementary Figs. 5 and 6). No putative sharing pairs were found solely within our urban Ndirande site, 16 putative sharing pairs were found within rural Chik- wawa and 4 putative sharing pairs were identified between Chikwawa and Ndirande. This is interesting as the two study sites are 50km apart and we found no epidemiological connections between either human or animal hosts between each site. In our dataset, 11/20 (55%) putative sharing events were from animal-animal host pairs. In contrast, we did not find any human- human host pairs, nor any environment-environment host pairs (Table 2, Fig. 3). Of our putative animal-animal sharing pairs, 5/11 (45%) occurred within the same animal species, whilst 6/11 (55%) occurred between animal species (Fig. 3). There were seven pairs of Salmonella genomes (7/20, 55%) which were shared between humans and either an animal or the environment (Fig. 3). Two of these were within the same household, one isolate of each pair collected at separate visits to the household. For example one Salmonella sample collected from the swab of an outside tap during the second visit to the household in Chikwawa was 0 SNPs different to Salmonella detected within the stool of a boy living at the same household, collected during the third visit to the household 6 months later. A second sharing pair involved S. Johannesburg from an adult human male collected during the first visit to the household Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 5 www.nature.com/naturecommunications and a chicken sampled at the third visit to the same household 6 months later. Both households were located within Chikwawa and were livestock and poultry-owning households in which the animals (domestic animals, livestock and peri-domestic wildlife) and humans shared the same living space within the household. The remaining five pairs were shared between human and animal samples taken at dif- ferent households. There was no sharing detected between wild birds and other hosts within the study. Within the collection of 227 genomes only four pairs of isolates with 0 pairwise SNPs in the core gene alignment shared the same AMR determinants. Three pairs of S. salamae genomes shared fosA7, a genomicAMRdeterminant which confers phenotypic resistance to the antibiotic fosfoycin. fosA7was chromosomally integrated within these S. salamae genomes. We detected one pair of S. Typhimurium ST19 genomes which shared the determinant qnrB19 (Fig. 3b) which confers phenotypic quinolone resistance and low level fluoroquinolone resis- tance. The qnrB19 AMRgenes were detected within plasmids classified by MOB-suite as replicon cluster 2355. These plasmids are non- conjugative and require a ‘helper’ plasmid possessing the necessary relaxase and Mpf genes to be mobilised. Three of these pairs were detected within the same household, one between households. Discussion In this study, we show that closely related salmonellae are shared between humans, animals and the environment both within and betweenhouseholds inMalawi, particularly in rural areas. In these rural areas, humans are often involved in low-intensity agricultural prac- tises, with the household itself a base in which both humans and ani- mals reside overnight. Thismeans that humans and animals of a variety of species often spend at least part of each day within close proximity. Salmonella has provided an excellent model to document the occur- rence of bacterial sharing around these household sites. The study setting in thisworkdiffersmarkedly from those in other areas, particularly more industrialised agricultural settings, where the interaction between humans, animals and their shared environment may be more limited40–43. Previous work has looked at strain- and resistome-sharing of Escherichia coli in sub-Saharan Africa between humans, animals and the environment in households in the urban setting of Nairobi, and demonstrated that sharing does occur between different host populations44. This study investigates sharing of Sal- monella within households located in both rural and urban environ- ments in sub-Saharan Africa. We collected samples from a diverse range of animal species that spend time within the household peri- meter in which the humans reside and we used a detailed longitudinal household sampling strategy to investigate distribution and dis- semination of Salmonella between humans, animals and the environ- ment. This sampling framework enabled us to correlate epidemiological linkageoccurring at the household levelwith genomic relatedness of strains. Within the study, a diverse collection of Salmonella genomes spanning two subspecies was detected. These isolates were carried apparently asymptomatically in the stool of a range of hosts, or detected within the environment. Importantly, whilst asymptomati- cally carried amongst humans and animals fromMalawian households within this study, many of the serovars have been previously reported to cause clinical disease in other settings25,26,45–49. The connection and relationship between carriage and disease of NTS is not yet fully clear across all serovars, in all species. Encouragingly, we also documented low rates of AMR determinant carriage, and no multidrug resistance was detected within isolates collected within these low-intensity agri- cultural settings. However, we are aware that there is generally little regulation of antimicrobial use both discretely for animals and as an additive component in animal feed in settings such as these, which may drive the spread of AMR in the future, and should bemonitored50. Ta b le 1 |D es cr ip ti o n o f th e n um b er o f g en o m e p ai rs an d p ai rw is e S N P d is ta n ce in th e to ta lc o ll ec ti o n an d w it h in an d b et w ee n h o us eh o ld p ai rs o f g en o m es (S up p le - m en ta ry Fi g s. 5 an d 6 ) W h o le co ll ec ti o n (n = 22 7) S u b .s p en te ri ca (n = 12 5 ) S u b .s p sa la m ae (n = 10 2) To ta l B et w ee n h o us eh o ld W it h in h o us eh o ld To ta l B et w ee n h o us eh o ld W it h in h o us eh o ld To ta l B et w ee n h o us eh o ld W it h in h o us eh o ld N um b er of g en om e p ai rs 25 ,6 51 23 ,8 10 18 4 1 77 50 72 0 0 55 0 51 51 4 6 8 8 4 6 3 M ed ia n S N P d is ta nc e (S N Ps ) 36 ,8 4 7 8 6 ,9 56 34 ,9 31 34 ,6 38 34 ,8 20 25 ,5 6 3 15 ,8 4 5 15 ,9 29 13 ,2 11 R an g e (S N Ps ) 0 – 8 8 ,8 8 6 0 – 8 8 ,8 8 6 0 – 8 8 ,6 6 7 0 – 38 ,1 9 8 0 – 38 ,1 9 8 0 – 37 ,9 22 0 – 3 8 ,7 12 0 – 38 ,7 12 0 – 37 ,9 6 7 C lo se ly re la te d (< or eq ua lt o 10 0 S N Ps ) 4 14 20 7 (5 0 .0 ) 20 7 (5 0 .0 ) 27 5 15 0 (5 4 .9 ) 12 3 (4 5. 1) 14 0 56 (4 0 .0 ) 8 4 (6 0 .0 ) V er y cl os el y re la te d (< or eq ua lt o 10 S N Ps ) 17 0 23 (1 3. 5) 14 7 (8 6 .5 ) 11 0 15 (1 3. 6 ) 9 5 (8 6 .4 ) 9 1 15 (1 6 .5 ) 76 (8 3. 5) ‘Id en tic al ’ g en om e p ai rs (0 S N Ps ) 6 7 8 (1 1. 9 ) 59 (8 8 .1 ) 51 4 (7 .8 ) 4 7 (9 2. 2) 26 7 (2 6 .9 ) 19 (7 3 .1 ) W ho le g en om e co lle ct io n re su lt s ar e g ai ne d us in g th e co re g en e al ig nm en to f2 27 g en om es .S .e nt er ic a re su lt s ar e g ai ne d fr om a co re g en e al ig nm en to f1 25 S .e nt er ic a g en om es on ly ,a nd S .s al am ae re su lt s ar e g ai ne d fr om a co re g en e al ig nm en to f1 0 2 S .s al am ae g en om es on ly .I n b ra ck et s ar e sh ow n th e p er ce nt ag e of th e w ho le g en om e co lle ct io n, S .e nt er ic a an d S .s al am ae g en om es w hi ch ar e ei th er cl os el y re la te d ,v er y cl os el y re la te d or ‘id en tic al ’ w hi ch w er e g en er at ed fr om sa m p le s co lle ct ed fr om w ith in th e sa m e ho us eh ol d or fr om tw o sa m p le s co lle ct ed fr om d iff er en t ho us eh ol d s. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 6 www.nature.com/naturecommunications Consideration should be given to the location within households in which sharing of Salmonella occurs in Malawi. Animals pass faeces freely in the compound and frequently share the same environment in which children play and food is prepared. Risk factors for the carriage of Salmonella within human stool in Malawi have been found to be linked to animal ownership and husbandry factors and so it is impor- tant that improved environmental hygiene within the household should be addressed in the context of sharing of Salmonella31. Fig. 3 | Sharing pairs of Salmonella detected within the study. a A composite map to show the nature of sharing pairs of Salmonella detected within the study within a One Health perspective. Human silhouette indicates genomes initially collected from humans. Animal silhouettes represent Salmonella initially collected from animals. The tree silhouette represents salmonellae collected fromwithin the environment of each household perimeter. Each silhouette represents the host category or animal species as a whole. Number of lines between silhouettes indi- cates the number of sharing events noted within the collection. The colour of the line between silhouettes denotes whether the sharing event occurred between isolates of the same household or different households. Sections of the circle correspond to human, animal, environment pairs as labelled by the central graphic or human-environment, human-animal, animal-environment between the corre- sponding sections. Illustration from NIAID NIH BioArt Source89–96 (Lizard Outline: Bioart.Niaid.Nih.Gov/Bioart/302; Sunflower: Bioart.Niaid.Nih.Gov/Bioart/620; Goat: Bioart.Niaid.Nih.Gov/Bioart/636; Duck Silhouette: Bioart.Niaid.Nih.Gov/ Bioart/135; Domestic Chicken: Bioart.Niaid.Nih.Gov/Bioart/131; Domestic Dog: Bioart.Niaid.Nih.Gov/Bioart/594; Lab Mouse: Bioart.Niaid.Nih.Gov/Bioart/279; Uni- sex Icon: Bioart.Niaid.Nih.Gov/Bioart/13)bNetworkproducedusing iGraph to show the occurrence of pairs of Salmonella with a SNP distance from the core genome alignment of 0 SNPs. The colour of the node denotes host species (green = animal, yellow=human, purple = animal/environment). The household number from which the isolate was sampled is shown in text in the centre of each node. CHH= Chikwawa, NHH=Ndirande followed by an identifying number of the household. A red perimeter of the node indicates that the isolate carries one antimicrobial resistant determinant. In all cases the AMR determinant is shared between the pair. Colour ofmark on upper right quadrant of circle denotes visit number. Source data are provided as a Source data file. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 7 www.nature.com/naturecommunications Within this collection of diverse salmonellae we only very rarely detected strains of Salmonellawhich have been previously found to be associated with invasive NTS disease in Malawi25,27. We detected one S. Typhimurium ST313 isolate collected from the stool of a dog within one household at one timepoint, and isolates of S. Enteritidis ST11 have been detected within stool of dogs and the environment of two households. Further investigation of the relationship of these poten- tially invasive strains with previously published isolates demonstrates that the ST313 ismore closely related to S. TyphimuriumST313 Lineage 3, which has been found to have emerged inMalawi in 2016, and the S. Enteritidis ST11 is closely related to those isolates of the outlier cluster which have been responsible for significant human disease in other settings (Supplementary Figs. 7 and 8)25,27. Within this study we have detected strong epidemiological links between Salmonella of rele- vance to human disease, but not to pathovars strongly associated with iNTS in Africa. This is predominantly consistent with previous findings, however, the potential for foodborne transmission cannot be dis- counted, as S. Enteritidis ST11 of the global epidemic clade has been previously isolated from samples collected within the livestock and poultry meat pathway in Tanzania51–53. This may be a consequence of a small sample size, but does also lead to the question of what are the reservoirs and key transmission routes of these pathovars? Despite evidence of of Salmonella circulation within and between households across the two study sites, we did not document social connection between any of the households within the study. We have, however, sampled extremely sparsely; peri-domestic wildlife or free- roamingdomestic or livestock animalsmaypassbetween and amongst households within a single study site, as described elsewhere, but equally we lacked the resolution to confirm an epidemiological link54. We suggest that Salmonella has been resident for a long period of time across both study sites, and recent spread of these particular genomes which are present within both study sites, has occurred. Considering themutation rate of S. Typhimurium (6.7 SNPs per genome per year) it is reasonable to assume that this recent spread has occurred over the last 10 years55. The total number of genome pairs detected within this study is low and therefore it is not possible to further quantify the sharing which has occurred. However, given the small number genomes (n = 227) overall, and in Ndirande (n = 33) specifically, amongst which sharing was investigated, it may have been expected that sharing could have been missed entirely. There- fore, investigation of a larger study population is warranted, in order to further quantify the sharing of Salmonella within and between households, study sites and hosts. Themost common typeof sharingwasbetween chicken anddogs. The connection of Salmonella sharing between dog and poultry is interesting. In Chikwawa dogs and poultry are often free-roaming around households during the daytime, and poultry are penned at night. As omnivores and scavengers, dogs, ducks and chickens have access to food and faeces which may be shared around the ground of the household, providing plenty of opportunity for faecal-oral trans- mission to occur. It may be that the presence of these animals within a household acts as an ecological driver for the sharing of Salmonella between hosts. A range of environmental health practises have been shown to be important to reduce the sharing of bacteria and AMR determinants31,56,57. Implementation of more stringent biosecurity procedures specifically as part of animal husbandry practises, includ- ing regular removal of animal faeces from around the household complex and improved hand hygiene are also profoundly important in the endeavour to reduce the potential spread of these bacteria within households in Malawi, and should be implemented alongside interventions. The majority of the world’s population live in developing econo- mies and small, low-intensity farms produce ~35% of the world’s food58,59. We show that sharing of identical salmonellae between humans, animals and the environment is possible and in fact likely, demonstrating the importance of considering all aspects of hygiene and biosecurity precautions within households when developing strategies to limit the movement, carriage and sharing of salmonellae and other gastrointestinal pathogens. The findings of this study have important implications for public health, livestock keeping and animal husbandry policy and practice in low-income, low-intensity farming settings and should be used to shape efforts to draft effective, durable evidence-based policies to safeguard human health and ensure sus- tainable livestock systems in these settings. Methods Study site Between November 2018-December 2019 a longitudinal prospective study recruited 30 households from two study sites in Malawi: Ndir- ande, Blantyre District and Chikwawa District. Within each geographic area polygons were created using QGIS software to create areas for inclusion61. Fifteen households were selected at random in the two study areas using R software version 2022.12.0 + 353 to generate Table 2 | Description of the pairwise SNP distance of pairs of salmonellae of 0 SNP of the within-Salmonella subspecies core gene distance (deduplicated samples only) Category salamae (n = 10) enterica (n = 10) Total (n = 20) Total number of pairs within collection Proportion of 0 SNP pairs within collec- tion % (95% CI) Sharing within or between household Within household 4 7 11 2026 0.54 (0.3–0.97) Between households 6 3 9 23,776 0.04 (0.02–0.07) Sharing between hosts Animal-Animal 4 7 11 7357 0.15 (0.08–0.27) Human-human 0 0 0 146 0 Environment-Environment 0 0 0 227 0 Animal-Environment 2 0 2 2616 0.08 Human-Environment 0 1 1 383 0.26 Human-animal 4 2 6 2172 0.28 (0.13–0.60) Sharing between study sites Chikwawa 9 7 16 9348 0.17 (0.11–0.28) Ndirande 0 0 0 286 0 Chikwawa-Ndirande 1 3 4 3267 0.12 Confidence intervals displayed in brackets where there are more than five pairs within each category. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 8 www.nature.com/naturecommunications random GPS coordinates using a spatial inhibitory design with close pairs62. Households in each locationmet the inclusion criteria of being located within the study sites, all human household members were able to give informed consent or assent to take part in the study themselves, and the Head of the Household was able to provide informed consent to sample animals and the environment within the household. Households were excluded if a household member or representative was unable to provide informed consent, household members spoke neither Chichewa or English and if the household was located outside the boundary of the study sites. Sampling was carried out at three time points in all households. Identical sampling proce- dures were used at each time point to collect samples from humans, animals and the environment. Sample collection Questionnaires detailing household composition, socioeconomic data, animal ownership, husbandry, contact of members with animals, health seeking behaviour of humans were administered at each household using an electronic case report form on a Samsung© tablet device using Open Data Kit Collect version 1.1863. At each visit to a household, faecal samples were collected from all consenting human participants and stool samples, rectal or cloacal swabswere taken from a representative number of each species of animal and/or birds pre- sent in the household. Environmental samples were collected using 3M® swabs from areas of suspected high human-human, animal- human, animal-animal contact. Method of faecal sample collection The method of faecal sample collection is explained in detail below. Human samples. Fieldworkers left sterile faecal sample containers for each human study participant in the household on Day 1 of sampling. These were clearly labelled to identify which container should be used for each participant. Participants were also given nitrile gloves, bio- degradable bowls and sample containers with spoons to facilitate collection of the sample. The bowls could be disposed of in a pit latrine or collected by the fieldworkers for hygienic disposal at the same time as stool sample collection. Families were given sealable opaque ‘free- zer bags’ or similar to store samples whilst waiting for collection. Samples were collected on Day 2. Domestic animal and livestock samples (dogs, cats, sheep, goats, cattle, pigs.). On Day 2 faeces (2–20 g) were collected directly from the rectum of animals should they be available at the household and placed into appropriately labelled individual sample containers. Appropriate Personal Protective Equipment including wellington boots, a boiler suit and protective gloves were used for sample collection. Animals were appropriately restrained by trained personnel during sample collection. Collection per rectum was con- ducted either manually or using a sterile swab, depending on the size of the animal. Poultry (chickens, ducks, geese, guinea fowl, turkeys, jungle fowl, pigeons and doves.). Domestic poultry were sampled on themorning of Day 2 prior to release from the overnight housing. Poultry were appropriately restrained and a single cloacal swab was obtained from each bird sampled by either an Assistant Veterinary Officer or the author. Peri-domestic wildlife samples Rodents. Household members were supplied with a pair of nitrile gloves and an appropriately labelled sterile faecal sample container with which to do this. The faecal sample container was then collected at the same time as the human stool samples on the second day of sampling. Geckos. Household members were taught how to recognise and col- lect gecko excreta on Day 1 of sampling at each household and appropriate sterile faecal sample containers and nitrile gloveswere left with the household in order that the gecko faeces be collected over- night prior to collectionof thepot onDay2. If householdmembershad not collected gecko faeces, the field team would perform the sample collection on Day 2. Wild birds. Again, following a pilot study to trial the efficacy of using bird nets to collect samples, the preferred method for wild bird faecal sample collection was found to involve placing appropriate clean tar- paulins underneath roosts located within the household perimeter on the morning of Day 2, upon first arrival at the household. Any faeces deposited on the tarpaulins by wild birds were collected and pooled into an appropriately labelled sterile sample container at the end of sampling on Day 2. Environmental samples Environmental samples were collected individually using sterile 3M® swabs. Each 3M® swab contains a sterile sponge swab and 10ml sterile buffered peptone water. Each sample was taken in a sterile manner, repeatedly rubbing a fresh 3M® swab over an area of the object to be sampled of up to 20 × 20 cm for 30 s. The soiled swab was then replaced directly into the original sterile 3M® swab packet, the plastic handle broken off and the packet securely fastened. The nature of location or object fromwhich the swab was taken was recorded on the outside of each packet. Where possible, environmental swabs were taken from certain consistent areas at each household and at each visit. These areas which included the door or curtain to the latrine, around the edge of the latrine, cooking areas, front door to house, the inside of water carriers, bootsocks on the floor outside the house, bootsocks on the floor inside the house, a dirty chitenje (ladies’ skirt-wrap), sur- face of bed. Bootsock samples were collected using plastic overshoes. Whilst wearing clean nitrile gloves, two clean plastic overshoes were placed onto one foot of the fieldworker or author. The fieldworker or author walked around the area of interest for roughly 1min. Once the boot- sock sample hadbeen collected,whilst wearing cleannitrile gloves, the outer plastic shoe cover was removed and placed into a sterile self- sealing, appropriately labelled, plastic bag, sealed and placed into the cool box for transportation to the laboratory. The total number of household environmental samples taken per visit was normally up to one third of the total samples taken, depen- dent on the total number of animal species present and total number of humans providing samples. Samples were stored at 4 °C with ice packs in a cool box until arrival at the College of Medicine (now KUHES) laboratories. Labora- tory processing commenced within 4 h of sample collection. Follow-up sampling Samples were taken from each of the fifteen randomly selected households at three time points: • TP0 • TP1 = TP0 + 2 months • TP2 = TP0+ 6 months Aside from the initial consenting steps, identical human and ani- mal sampling strategies were used at each time point. Microbiological testing Salmonellae were isolated and identified by selective culture using enrichment steps using buffered peptone water and Rappaport Vas- siliadis for 24 h each respectively and a loop of bacterial solution from each were streaked out onto each of CASE and XLD selective agar Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 9 www.nature.com/naturecommunications plates prior to incubation at 37 °C (Supplementary Fig. 2). From the XLD and CASE agar culture plates up to 5 colonies per sample were randomly selected and underwent O and Vi antigen testing to confirm the presence of salmonellae, and the absence of typhoidal-Salmonella. Once isolated, aliquots of pure bacterial growth were stored at −80 °C in individualmicrobanks ormodifiedmicrobank tubes.Up to five picks of suspected Salmonella isolates from each sample were stored. Salmonella ttr qPCR Quantitative PCR using bacteria extracted using the boilate method was carried out of all stored colonies of suspected Salmonella at the Malawi Liverpool Wellcome Programme64. Positive qPCR confirmation of each isolate was denoted by the presence of the tetrathionate reductase (ttr) gene65. This gene is involved in the respiration of Sal- monella and is constitutively expressed in all salmonellae. The assays were run on a QuantStudio 7500 PCR machine. All qPCR assays in this study were run for 40 cycles. The highest acceptable cycle threshold is 35, but maintaining the cycle threshold at 40 allows for the identifi- cation of late amplification of the DNA (without it being due to chi- meras). A standard curve (serial amplification of amplification target for which the concentration is known) was included in each run to allow estimation of the Salmonella bacterial load (copy numbers). Primers, Master mix and probe The reaction mix is detailed in Supplementary Table 1. The primers were chosen to detect presence of the tetrathionate reductase (ttr) gene which is constitutively present in all Salmonella spp.65. The pri- mers used were ttr−4 (AGCTCAGACCAAAAGTGACCATC) and ttr−6 (CTCACCAGGAGATTACAACATGG) (Supplementary Table 2),madeup and supplied by Sigma66. Primers, probe and Master mix were stored at −20 °C. Controls Three control samples were used in the PCR reaction. A positive control (S. Typhimurium NCTC), a negative boilate control (sterile distilled water only) and a negative PCR control (Master Mix omitted, other constituents were present). Results were analysed by reviewing positive and negative controls, adjusting the cycle threshold for detection above any background noise and reviewing the standard curve. The cycle threshold was set at the beginning of the exponential curve in the linear graph, and the middle of the linear phase of the log graph. For the standard curve a correlation coefficient (R2) of >0.9, amplification effi- ciency of >80% and a minimum of 5 points within the assay linear range was considered adequate. Cycle threshold vales of the standard curve were also checked against typical and expected values. For the run to be accepted all negative controls had to be below the threshold with no amplification and positive controls had to demonstrate a cycle threshold value <35 and a sigmoid curve. Analysis was performed by the laboratory technicians, reviewed and approved by the author. Assays were repeated when samples failed quality control. Reaction procedure Following preparation of the reaction mix the procedure was as follows. 1. 22.5 µl of Master Mix, primer, probe and nuclease free water solution (Supplementary Table 1) were loaded into eachwell to be used of a new 96-well fast optical plate. 2. 2.5 µl of samplewere added to the appropriatewells, including the controls. 3. Optical seals were applied to seal the plate. 4. The plate was spun for 5 s in plate centrifuge. 5. Plate cycled at the following temperatures for each reaction, for 40 cycles: a. Denaturation 30 s 95 °C b. Annealing 30 s 60 °C c. Extension 10 s 72 °C The ramping up and down of temperature was set to 1.6 °C per second. Outcome and DNA extraction techniques Isolates which were positive for the ttr gene were deemed to be Sal- monella (previous work had confirmed that these were not typhoidal- salmonellae). These Salmonella isolates were stored and DNA extrac- tion subsequently carried out using Qiagen DNA extraction kits (Qia- gen DNA Mini kit). The DNA of two-three frozen colonies per sample was quantified as required by the guidelines of the Wellcome Sanger Institute using a Qubit© (Thermo Fisher Scientific, MA, USA). The extracted DNA was then sent for whole genome sequencing at the Wellcome Sanger Institute, UK. Whole genome sequencing and bioinformatics Genomic sequencing techniques. Samples of whole bacterial isolates forwhole genomesequencingwere submitted to theWellcomeSanger Institute, Hinxton. Half of the DNA from each sample remained in storage at −80 °C at MLW, and half were transferred via sterile pipette into a 0.3ml sterile FluidX 2D Sequencing Tube (FluidX Ltd, UK). Genomic sequencing libraries were prepared using the NEBNext Ultra II (New EnglandBiolabs,Massachusetts, USA),multiplexed at 384 unique dual indexed barcode combinations, and sequenced on Illu- mina HiSeq X10 to generate 150 bp paired end reads. Post sequencing quality control showed a mean insert size of 180 bp and a mean frag- ment size of 450 bp. The median depth of coverage was 74.4. FastQC (version 0.11.9) and multiQC (version 0.11.8) were used to assess per base sequence quality, quality scores per sequence, per base sequence content, per base GC content, per sequence GC content, per base N content, contig length distribution and sequence duplication levels67 (for cut-off values see Supplementary Table 4). Read quality control was undertaken using Kraken (cut-off pro- portion reads <70% abundance Salmonella, Kraken version 1.1.1)68. CheckM (cut-offs used; contamination > 20% or completeness <90% removed, CheckM version 1.1.2) was run to assess contamination, strain heterogeneity and completeness of the genomes69. Assembly Statistics (genome length > 7Mbp or contigs > 500 removed, Assem- bly Statistics version 1.0.1) was run to analyse the total genome length and number of contigs70. The Quality Assessment Tool for Genome Assemblies (QUAST)(cut-offs used contigs >500, N50 <20kbp or total base pairs <4Mbpor >5.8Mbp,QUASTversion5.0.2)wasused to assess the number of contigs, N50 and total length of the genome71 (Sup- plementary Table 4). Following the completion of quality control procedures, genomes were submitted to Pathogenwatch, which uses SISTR to assess the species, serovar and ST of the bacteria present33,72. One of the genomes which passed quality control procedures was incompletely assembled in the WSI pipeline (lane ID 34747_4#7), therefore SPADES (version 3.14) was used to assemble the genome and Prokka (version 1.14.5)was used for genome annotation73,74. In total 227 good quality whole genome sequences were identified which passed the stipulated thresholds. Core-genome phylogeny and SNP analysis. A core and pangenome analysis was performed using Panaroo (version 1.3.3)39. A gene was considered core if it was present in 100% of the genomes at a match identity threshold of 98%75. A core genome sequence alignment was generated using Panaroo by concatenating the alignments of the core genes. Single nucleotide polymorphic (SNP) site alignment was Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 10 www.nature.com/naturecommunications generated from the core genome alignment using SNP-sites (version 2.5.1)76. IQtree version 2.2.0) was run on the resulting core SNP- alignment to construct amaximum likelihood tree using the core gene SNP alignment of all 227 isolates77. Reliability of inferred branch par- titions was assessed with 100 bootstrap replicates. The tree was visualised using ITOL (version 5) and ggtree (version 3.2)78–80. Identification of AMR determinants, virulence factors and plasmid typing. AMRFinderPlus (version 3.10) was used to detect chromosomal mutations encoding for AMR, acquired AMR genes (ARGs) and heavy metal resistance genes81,82. Those ARGs with an identity of 95% and a coverage of 95% were taken forward for further analysis. Determining appropriate thresholds for epidemiological analy- sis of putative bacterial sharing During the laboratory culture work, up to five picks of Salmonellawere isolated from each positive sample to capture multi-serovar carriage and within-host diversity. To refine the collection to include solely genetically distinct Salmonella isolates from each host, pairwise SNP distance measurement using the core genome alignment was used to detect the SNP distance between any Salmonella isolates originating from the same individual. Pairwise SNP distances were calculated using ‘pairwise difference count’ and snp-dists in order to measure the average number of SNP differences between strains within each sub-clade83,84. A SNP distance of 0 SNPs was used as a cut-off to define putative sharing of salmo- nellae, a ‘sharing pair’44. Epidemiological analysis of ‘sharing-pairs’. Systematic considera- tion of each sharing pair alongside the metadata was carried out and epidemiological links between sharing-pairs were established. Household-level sharing was defined as a sharing-pair of which both genomes within the pair originated from samples collected from dif- ferent hosts within the same household. A between-household shar- ing-pair pairs was defined as a pair in which each of the isolates were collected fromdifferent households. IGraph (version 1.3.5)was used to visualise a network of the sharing pairs85,86. Statistical methods Analysis was conducted using R version 2022.12.0 + 35387. Missing data were rare and unless otherwise specified missing variables were man- aged by exclusion from analysis. Ethics The study complies with all relevant ethical regulations. Ethical approval for this study was obtained from the University of Liver- pool Veterinary Research Ethics Committee (Reference number VREC686) and the College of Medicine Research Ethics Committee (COMREC), Malawi (Reference Number P.02/18/2368). Informed, written consent was obtained from all household heads and indivi- dual household members (or their representatives) prior to their entry into the study following discussion of the study protocol, risks and benefits, financial and confidentiality considerations and details of methods to obtain more information. Should the prospective study participant be illiterate, the study was explained verbally and the consent form was read to the participant by the study team, witnessed by an additional neighbour who was not a member of the household. If the participant agreed to enter the study, the witness signed and dated the form and the witness documented their con- sent with an inked thumbprint. Parents or guardians were invited to consent for their children/wards of less than 18 years to join the study. Written assentwas sought for children between the ages of 8–18 in accordance with WHO guidelines60. For children younger than 8 years, the parent or guardian of the childwas asked to provide full written consent. This research does not result in stigmatistion, incrimination, discrimination or otherwise personal risk to the par- ticipants. All human data has been anonymised. Benefit sharing measures have been discussed and agreed with the local host insti- tution in Malawi, and at least one aliquot of all sample materials, including extracted DNA, remain in country. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Data availability The data supporting the findings of this study are available in this article, the Source data file (Source Data- Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households in Malawi) and have been deposited in the Zenodo data base under at the following https://doi.org/10.5281/zenodo.17191987. Raw sequencing reads for all novel sequences are deposited at the European Nucleotide Archive (ENA) under project (PRJEB32657). All accession numbers (both novel and previously published) used in this project are listed in the Source Data file (‘Source Data- Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households inMalawi’) alongwith allmetadata used for analysis in Figs. 1, 2, 3 and Supplementary Figs. 3, 4, 5 and 6. Previously published contextual metadata used in Supplementary Fig. 5 and Supplementary Fig. 6 are displayed in Source Data file (‘Source Data— Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households in Malawi’). Publically available sequence data was downloaded from one of the following sources: GenBank (https://www.ncbi.nlm.nih.gov/genbank/), Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra), European NucleotideArchive (https://www.ebi.ac.uk/ena) or Enterobase (https:// enterobase.warwick.ac.uk). Source data are provided with this paper. Code availability TheR code for the current study is publically availableonGitHub at the following https://doi.org/10.5281/zenodo.171918788. References 1. Blancou, J., Chomel, B. B., Belotto, A. &Meslin, F. X. Emerging or re- emerging bacterial zoonoses: factors of emergence, surveillance and control. Vet. Res. 36, 507–522 (2005). 2. Jones, K. E. et al. Global trends in emerging infectious diseases. Nature 451, 990–993 (2008). 3. Shaheen,M. N. F. The concept of one health applied to the problem of zoonotic diseases. Rev. Med. Virol. 32, e2326 (2022). 4. Nandi, A. & Allen, L. J. S. Probability of a zoonotic spillover with seasonal variation. Infect. Dis. Model. 6, 514–531 (2021). 5. One Health High-Level Expert Panel (OHHLEP) One health: a new definition for a sustainable and healthy future. PLOS Pathog. 18, e1010537 (2022). 6. Nava, A., Shimabukuro, J. S., Chmura, A. A. & Luz, S. L. B. The impact of global environmental changes on infectious disease emergence with a focus on risks for Brazil. ILAR J. 58, 393–400 (2017). 7. LaDeau, S. L., Calder, C. A., Doran, P. J. &Marra, P. P. West Nile virus impacts in American crow populations are associated with human land use and climate. Ecol. Res. 26, 909 (2011). 8. Kruger, S. E., Lorah, P. A. & Okamoto, K. W. Mapping climate change’s impact on cholera infection risk in Bangladesh. PLOS Glob. Public Health 2, e0000711 (2022). 9. Hassell, J. M., Begon, M.,Ward,M. J. & Fèvre, E. M. Urbanization and disease emergence: dynamics at the wildlife–livestock–human interface. Trends Ecol. Evol. 32, 55–67 (2017). 10. Carlson, C. J. et al. Climate change increases cross-species viral transmission risk. Nature 607, 555–562 (2022). Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 11 https://doi.org/10.5281/zenodo.17191987 https://www.ncbi.nlm.nih.gov/genbank/ https://www.ncbi.nlm.nih.gov/sra https://www.ebi.ac.uk/ena https://enterobase.warwick.ac.uk/ https://enterobase.warwick.ac.uk/ https://doi.org/10.5281/zenodo.1719187 www.nature.com/naturecommunications 11. Mora, C. et al. Over half of known human pathogenic diseases can be aggravated by climate change. Nat. Clim. Change 12, 869–875 (2022). 12. Murray, C. J. et al. Global burden of bacterial antimicrobial resis- tance in 2019: a systematic analysis. Lancet 399, 629–655 (2022). 13. Kariuki, S., Kering, K., Wairimu, C., Onsare, R. & Mbae, C. Anti- microbial resistance rates and surveillance in sub-Saharan Africa: where are we now?. Infect. Drug Resist. 15, 3589–3609 (2022). 14. Fatica, M. K. & Schneider, K. R. Salmonella and produce: survival in the plant environment and implications in food safety. Virulence 2, 573–579 (2011). 15. Liu, H., Whitehouse, C. A. & Li, B. Presence and persistence of Sal- monella in water: the impact onmicrobial quality of water and food safety. Front. Public Health 6, 159 (2018). 16. Majowicz, S. E. et al. The global burden of nontyphoidal Salmonella gastroenteritis. Clin. Infect. Dis. Publ. Infect. Dis. Soc. Am. 50, 882–889 (2010). 17. Hald, T. et al. World Health Organization estimates of the relative contributions of food to the burden of disease due to selected foodborne hazards: a structured expert elicitation. PLoS ONE 11, e0145839 (2016). 18. Reddy, E. A., Shaw, A. V. & Crump, J. A. Community-acquired bloodstream infections in Africa: a systematic review and meta- analysis. Lancet Infect. Dis. 10, 417–432 (2010). 19. Feasey, N. A. et al. Three epidemics of invasive multidrug-resistant salmonella bloodstream infection in Blantyre, Malawi, 1998–2014. Clin. Infect. Dis. Publ. Infect. Dis. Soc. Am. 61, S363–S371 (2015). 20. Uche, I. V., MacLennan, C. A. & Saul, A. A systematic review of the incidence, risk factors and case fatality rates of invasive non- typhoidal Salmonella (iNTS) disease in Africa (1966 to 2014). PLoS Negl. Trop. Dis. 11, e0005118 (2017). 21. Marchello, C. S., Dale, A. P., Pisharody, S., Rubach,M. P. &Crump, J. A. A systematic review and meta-analysis of the prevalence of community-onset bloodstream infections among hospitalized patients in Africa and Asia. Antimicrob. Agents Chemother. 64, https://doi.org/10.1128/aac.01974-19 (2019). 22. Balasubramanian, R. et al. The global burden and epidemiology of invasive non-typhoidal Salmonella infections. Hum. Vaccines Immunother. 15, 1421–1426 (2018). 23. Park, S. E. et al. The genomic epidemiology of multi-drug resistant invasive non-typhoidal Salmonella in selected sub-Saharan African countries. BMJ Glob. Health 6, e005659 (2021). 24. Akullian, A. et al. Multi-drug resistant non-typhoidal Salmonella associatedwith invasive disease in western Kenya. PLoSNegl. Trop. Dis. 12, e0006156 (2018). 25. Feasey, N. A. et al. Distinct Salmonella Enteritidis lineages asso- ciated with enterocolitis in high-income settings and invasive dis- ease in low-income settings. Nat. Genet. 48, 1211–1217 (2016). 26. Kingsley, R. A. et al. Epidemic multiple drug resistant Salmonella Typhimurium causing invasive disease in sub-Saharan Africa have a distinct genotype. Genome Res. 19, 2279–2287 (2009). 27. Pulford, C. V. et al. Stepwise evolution of Salmonella Typhimurium ST313 causing bloodstream infection in Africa. Nat. Microbiol. 6, 327–338 (2021). 28. Van Puyvelde, S. et al. An African Salmonella Typhimurium ST313 sublineage with extensive drug-resistance and signatures of host adaptation. Nat. Commun. 10, 4280 (2019). 29. LEI International Policy et al. Dynamics of Food Systems in Sub- Saharan Africa: Implications for Consumption Patterns and Farmers’ Position in Food Supply Chains. https://research.wur.nl/en/ publications/973b48c2-2933-4e13-825e-3815fe9be428; https:// doi.org/10.18174/417176 (2017). 30. Nyokabi, S. et al. Informal value chain actors’ knowledge and per- ceptions about zoonotic diseases and biosecurity in Kenya and the importance for food safety and public health. Trop. Anim. Health Prod. 50, 509–518 (2018). 31. Cocker, D. et al. Investigating One Health risks for human coloni- sation with extended spectrum β-lactamase-producing Escherichia coli and Klebsiella pneumoniae in Malawian households: a long- itudinal cohort study. Lancet Microbe S2666524723000629 https://doi.org/10.1016/S2666-5247(23)00062-9 (2023). 32. Hassell, J. M. et al. Clinically relevant antimicrobial resistance at the wildlife–livestock–human interface in Nairobi: an epidemiological study. Lancet Planet. Health 3, e259–e269 (2019). 33. Pathogenwatch | A Global Platform for Genomic Surveillance. https://pathogen.watch/. 34. Kariuki, S., Gordon, M. A., Feasey, N. & Parry, C. M. Antimicrobial resistance and management of invasive Salmonella disease. Vac- cine 33, C21–C29 (2015). 35. Parisi, A. et al. The role of animals as a source of antimicrobial resistant nontyphoidal Salmonella causing invasive and non- invasive human disease in Vietnam. Infect. Genet. Evol. 85, 104534 (2020). 36. Musicha, P. et al. Trends in antimicrobial resistance in bloodstream infection isolates at a large urban hospital in Malawi (1998–2016): a surveillance study. Lancet Infect. Dis. 17, 1042–1052 (2017). 37. Robertson, J. & Nash, J. H. E. MOB-suite: software tools for clus- tering, reconstruction and typing of plasmids from draft assem- blies. Microb. Genomics 4, e000206 (2018). 38. Seemann, T. ABRicate. Github https://github.com/tseemann/ abricate. (2021). 39. Tonkin-Hill, G. et al. Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol. 21, 180 (2020). 40. Thorpe, H. A. et al. A large-scale genomic snapshot of Klebsiella spp. isolates inNorthern Italy reveals limited transmissionbetweenclinical and non-clinical settings. Nat. Microbiol. 7, 2054–2067 (2022). 41. Day, M. J. et al. Extended-spectrum β-lactamase-producing Escherichia coli in human-derived and foodchain-derived samples from England, Wales, and Scotland: an epidemiological surveil- lance and typing study. Lancet Infect. Dis. 19, 1325–1335 (2019). 42. Ludden, C. et al. One health genomic surveillance of Escherichia coli demonstrates distinct lineages andmobile genetic elements in isolates from humans versus livestock.mBio 10, https://doi.org/10. 1128/mbio.02693-18 (2019). 43. Gouliouris, T. et al. Genomic surveillance of enterococcus faecium reveals limited sharing of strains and resistance genes between livestock and humans in the United Kingdom. mBio 9, e01780- 18 (2018). 44. Muloi, D. M. et al. Population genomics of Escherichia coli in livestock-keeping households across a rapidly developing urban landscape. Nat. Microbiol. 7, 581–589 (2022). 45. Fukushima, K., Yanagisawa, N., Sekiya, N. & Izumiya, H. Bacteremia caused by Salmonella Poona in a healthy adult in Tokyo, Japan. Intern. Med. 59, 289–292 (2020). 46. Lang, B. Y., Varman, M., Reindel, R. & Hasley, B. P. Salmonella Gaminara osteomyelitis and septic arthritis in an infant with expo- sure to bearded dragon. Infect. Dis. Clin. Pract. 15, 348–350 (2007). 47. Makendi, C. et al. A phylogenetic and phenotypic analysis of Sal- monella enterica serovar Weltevreden, an emerging agent of diar- rheal disease in tropical regions. PLoS Negl. Trop. Dis. 10, e0004446 (2016). 48. Beyene, G. et al. Multidrug resistant Salmonella Concord is a major cause of salmonellosis in children in Ethiopia. J. Infect. Dev. Ctries. 5, 023–033 (2011). 49. Nutrition, C. for F. S. & A. Outbreak Investigation of Salmonella Oranienburg: Whole, Fresh Onions (October 2021). FDA (2022). 50. Mankhomwa, J. et al. A qualitative study of antibiotic use practices in intensive small-scale farming in urban and peri-urban Blantyre, Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 12 https://doi.org/10.1128/aac.01974-19 https://research.wur.nl/en/publications/973b48c2-2933-4e13-825e-3815fe9be428 https://research.wur.nl/en/publications/973b48c2-2933-4e13-825e-3815fe9be428 https://doi.org/10.18174/417176 https://doi.org/10.18174/417176 https://doi.org/10.1016/S2666-5247(23)00062-9 https://pathogen.watch/ https://github.com/tseemann/abricate https://github.com/tseemann/abricate https://doi.org/10.1128/mbio.02693-18 https://doi.org/10.1128/mbio.02693-18 www.nature.com/naturecommunications Malawi: implications for antimicrobial resistance. Front. Vet. Sci. 9, 876513 (2022). 51. Post, A. S. et al. Supporting evidence for a human reservoir of invasive non-Typhoidal Salmonella from household samples in Burkina Faso. PLoS Negl. Trop. Dis. 13, e0007782 (2019). 52. Koolman, L. et al. Case-control investigation of invasive Salmonella disease in Malawi reveals no evidence of environmental or animal transmission of invasive strains, and supports human to human transmission. PLoS Negl. Trop. Dis. 16, e0010982 (2022). 53. Crump, J. A. et al. Investigating the meat pathway as a source of human nontyphoidal Salmonella bloodstream infections and diar- rhea in East Africa. Clin. Infect. Dis. 73, e1570–e1578 (2021). 54. Hassell, J. M. et al. Epidemiological connectivity between humans and animals across an urban landscape. Proc. Natl Acad. Sci. USA 120, e2218860120 (2023). 55. Octavia, S., Wang, Q., Tanaka, M. M., Sintchenko, V. & Lan, R. Genomic variability of serial human isolates of Salmonella enterica serovar Typhimurium associated with prolonged carriage. J. Clin. Microbiol. 53, 3507–3514 (2015). 56. Chidziwisano, K., Slekiene, J., Kumwenda, S., Mosler, H.-J. & Morse, T. Toward complementary food hygiene practices among child caregivers in rural Malawi. Am. J. Trop. Med. Hyg. 101, 294–303 (2019). 57. Musoke, D. et al. The role of Environmental Health in preventing antimicrobial resistance in low- and middle-income countries. Environ. Health Prev. Med. 26, 100 (2021). 58. Now 8 billion and counting: Where the world’s population has grown most and why that matters |. UNCTAD https://unctad.org/ data-visualization/now-8-billion-and-counting-where-worlds- population-has-grown-most-and-why (2022). 59. FAO—News Article: Small family farmers produce a third of the world’s food. https://www.fao.org/news/story/en/item/1395127/ icode/. 60. Paediatrics, R. C., Committee, C. H. & HULL, P. S. Guidelines for the ethical conduct of medical research involving children. Arch. Dis. Child. 82, 177–182 (2000). 61. Welcome to the QGIS project! https://www.qgis.org/en/site/. 62. Chipeta, M., Terlouw, D., Phiri, K. & Diggle, P. Inhibitory geostatis- tical designs for spatial prediction taking account of uncertain covariance structure. Environmetrics 28, e2425 (2017). 63. Hartung, C. et al. Open data kit: tools to build information services for developing regions. in Proceedings of the 4th ACM/IEEE Inter- national Conference on Information and Communication Technolo- gies and Development 1–12. https://doi.org/10.1145/2369220. 2369236 (Association for Computing Machinery, 2010). 64. De Medici, D. et al. Evaluation of DNA extraction methods for use in combination with SYBRGreen I real-time PCR To detect Salmonella enterica serotype enteritidis in poultry. Appl. Environ.Microbiol.69, 3456–3461 (2003). 65. Hensel, M., Hinsley, A. P., Nikolaus, T., Sawers, G. & Berks, B. C. The genetic basis of tetrathionate respiration in Salmonella typhimur- ium. Mol. Microbiol. 32, 275–287 (1999). 66. Custom DNA Oligos. https://www.sigmaaldrich.com/GB/en/ custom-pdp/061ac088-00b9-47c0-8a48-faab9ca7f281. 67. Babraham Bioinformatics—FastQC A Quality Control tool for High Throughput Sequence Data. https://www.bioinformatics. babraham.ac.uk/projects/fastqc/. 68. Kraken: ultrafast metagenomic sequence classification using exact alignments | Genome Biology | Full Text. https://genomebiology. biomedcentral.com/articles/10.1186/gb-2014-15-3-r46. 69. Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P. & Tyson, G. W. CheckM: assessing the quality of microbial genomes recov- ered from isolates, single cells, andmetagenomes.GenomeRes25, 1043–1055 (2015). 70. assembly-stats. Pathogen Informatics, Wellcome Sanger Institute. Github https://github.com/sanger-pathogens/assembly-stats. (2022). 71. Gurevich, A., Saveliev, V., Vyahhi, N. & Tesler, G. QUAST: quality assessment tool for genome assemblies. Bioinforma. Oxf. Engl. 29, 1072–1075 (2013). 72. Yoshida, C. E. et al. The Salmonella In Silico Typing Resource (SISTR): an open web-accessible tool for rapidly typing and sub- typing draft Salmonella genome assemblies. PLoS ONE. 11, e0147101 (2016). 73. Bankevich, A. et al. SPAdes: a newgenome assembly algorithmand its applications to single-cell sequencing. J. Comput. Biol. 19, 455–477 (2012). 74. Seemann, T. Prokka: rapid prokaryotic genome annotation. Bioin- formatics 30, 2068–2069 (2014). 75. Page, A. J. et al. Roary: rapid large-scale prokaryote pan genome analysis. Bioinformatics 31, 3691–3693 (2015). 76. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments | Microbiology Society. https://www. microbiologyresearch.org/content/journal/mgen/10.1099/mgen. 0.000056. 77. Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020). 78. Hadfield, J. et al. Phandango: an interactive viewer for bacterial population genomics. Bioinformatics 34, 292–293 (2018). 79. Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation.Nucleic Acids Res 49, W293–W296 (2021). 80. Yu, G. Using ggtree to visualize data on tree-like structures. Curr. Protoc. Bioinforma. 69, e96 (2020). 81. AMRFinderPlus—Pathogen Detection—NCBI. https://www.ncbi. nlm.nih.gov/pathogens/antimicrobial-resistance/AMRFinder/. 82. Feldgarden, M. et al. AMRFinderPlus and the Reference Gene Cat- alog facilitate examination of the genomic links among anti- microbial resistance, stress response, and virulence. Sci. Rep. 11, 12728 (2021). 83. Harris, S. Pairwise Difference Count. https://github.com/ simonrharris/pairwise_difference_count. (2020). 84. Seemann, T. snp-dists. https://github.com/tseemann/snp-dists (2021). 85. igraph—Network analysis software. https://igraph.org/. 86. Yu, D., Zhao, Y., Yin, C., Liang, F. & Chen, W. A network analysis of the association between intergroup contact and intergroup rela- tions. Psychol. Res. Behav. Manag. 15, 51–69 (2022). 87. R Core Team. R: a language and environment for statistical com- puting. https://www.r-project.org/ (2021). 88. Wilson, C. Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households in Malawi. (2025). 89. NIAID Visual & Medical Arts. (10/7/2024). Lizard Outline. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/302. 90. NIAID Visual & Medical Arts. (3/27/2025). Sunflower. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/620. 91. NIAID Visual & Medical Arts. (4/24/2025). Goat. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/636. 92. NIAID Visual &Medical Arts. (10/7/2024). Duck Silhouette. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/135. 93. NIAID Visual & Medical Arts. (10/7/2024). Domestic Chicken. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/131. 94. NIAID Visual & Medical Arts. (3/12/2025). Domestic Dog. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/594. 95. NIAID Visual & Medical Arts. (10/7/2024). Lab Mouse. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/279. Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 13 https://unctad.org/data-visualization/now-8-billion-and-counting-where-worlds-population-has-grown-most-and-why https://unctad.org/data-visualization/now-8-billion-and-counting-where-worlds-population-has-grown-most-and-why https://unctad.org/data-visualization/now-8-billion-and-counting-where-worlds-population-has-grown-most-and-why https://www.fao.org/news/story/en/item/1395127/icode/ https://www.fao.org/news/story/en/item/1395127/icode/ https://www.qgis.org/en/site/ https://doi.org/10.1145/2369220.2369236 https://doi.org/10.1145/2369220.2369236 https://www.sigmaaldrich.com/GB/en/custom-pdp/061ac088-00b9-47c0-8a48-faab9ca7f281 https://www.sigmaaldrich.com/GB/en/custom-pdp/061ac088-00b9-47c0-8a48-faab9ca7f281 https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ https://genomebiology.biomedcentral.com/articles/10.1186/gb-2014-15-3-r46 https://genomebiology.biomedcentral.com/articles/10.1186/gb-2014-15-3-r46 https://github.com/sanger-pathogens/assembly-stats https://www.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.000056 https://www.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.000056 https://www.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.000056 https://www.ncbi.nlm.nih.gov/pathogens/antimicrobial-resistance/AMRFinder/ https://www.ncbi.nlm.nih.gov/pathogens/antimicrobial-resistance/AMRFinder/ https://github.com/simonrharris/pairwise_difference_count https://github.com/simonrharris/pairwise_difference_count https://github.com/tseemann/snp-dists https://igraph.org/ https://www.r-project.org/ www.nature.com/naturecommunications 96. NIAID Visual & Medical Arts. (10/7/2024). Unisex Icon. NIAID NIH BIOART Source. Bioart.Niaid.Nih.Gov/Bioart/13. Acknowledgements We would like to thank the households recruited to this study for their participation in the sample collection. This work has been sup- ported by a Wellcome Trust Clinical Fellowship award to Catherine N. Wilson, grant number 203919/Z/16/Z. M.A.B. and N.R.T. were supported by Wellcome Funding to the Sanger Institute (#206194). N.R.T., M.A.B., P.M. and C.N.W. were also supported by Wellcome Funding to the Sanger Institute (220540/Z/20/A) during the course of this work. Author contributions C.N.W., E.M.F., N.A.F. and G.P. were involved in the conceptualisation and designing the overall study. C.N.W., J.N., A.M., M.D., L.B. and P.D. performed the fieldwork and sample collection. N.E. and C.J. assisted with study and laboratory protocol design. C.N.W. and L.M. designed the ODK data capture database and questionnaire and L.M. pro- grammed this. C.N.W., Y.D., O.K., Z.K. and C.S. performed the microbiological and molecular laboratory work for this study. C.N.W. performed the bioinformatic analysis with the assistance of P.M., M.A.B., under the overall supervision of N.R.T. E.M.F., N.A.F. and G.P. supervised the entire study. C.N.W. wrote the paper and all of the authors commented on the paper draft. E.M.F. and N.R.T. are the co- senior authors of this paper. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-025-65266-1. Correspondence and requests for materials should be addressed to Catherine N. Wilson or Eric M. Fèvre. Peer review information Nature Communications thanks Martyn Kirk, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jur- isdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2025, modified publication 2026 Article https://doi.org/10.1038/s41467-025-65266-1 Nature Communications | (2025) 16:9703 14 https://doi.org/10.1038/s41467-025-65266-1 http://www.nature.com/reprints http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ www.nature.com/naturecommunications Circulation of Salmonella spp. between humans, animals and the environment in animal-owning households in Malawi Results Description of the participating households Distribution of Salmonella genomes Inferring the distribution of plasmids, antimicrobial resistance determinants and virulence genes within the collection Genetic diversity of salmonellae genomes Phylogenomic distribution of salmonellae Sharing of Salmonella between hosts Discussion Methods Study site Sample collection Method of faecal sample collection Human samples Domestic animal and livestock samples (dogs, cats, sheep, goats, cattle, pigs.) Poultry (chickens, ducks, geese, guinea fowl, turkeys, jungle fowl, pigeons and doves.) Peri-domestic wildlife samples Rodents Geckos Wild birds Environmental samples Follow-up sampling Microbiological testing Salmonella ttr qPCR Primers, Master mix and probe Controls Reaction procedure Outcome and DNA extraction techniques Whole genome sequencing and bioinformatics Genomic sequencing techniques Core-genome phylogeny and SNP analysis Identification of AMR determinants, virulence factors and plasmid typing Determining appropriate thresholds for epidemiological analysis of putative bacterial sharing Epidemiological analysis of ‘sharing-pairs’ Statistical methods Ethics Reporting summary Data availability Code availability References Acknowledgements Author contributions Competing interests Additional information