Animal Microbiome

The evolutionary divergence between Henophidia (non-venomous) and Caenophidia (venomous) snakes has produced distinct cranial morphologies, digestive strategies, and presence of specialised venom systems in Caenophidia, yet the extent to which these long-standing diverging trajectories have shaped cloacal microbiota assembly remains poorly understood. We characterised cloacal microbiota in 70 captive snakes (52 Caenophidia, 18 Henophidia) by 16S rRNA amplicon sequencing. Beta diversity was tested by PERMANOVA, differential abundance by ANCOM-BC2, community types by Dirichlet Multinomial Mixture modelling (DMM), and microbial interactions by SparCC co-occurrence networks. Predicted functional potential (PICRUSt2) was analysed by ALDEx2 differential abundance testing and elastic net feature selection. Henophidia exhibited significantly higher bacterial richness and greater compositional variability than Caenophidia. Community composition showed clade-associated differences (PERMANOVA) and partitioned into two distinct DMM community types. The Henophidia network was 11.9-fold denser and more modular, with Burkholderiaceae as a keystone hub, whereas the Caenophidia network was sparse. Henophidia showed predicted enrichment in C1 metabolic pathways (ethylmalonyl-CoA, formaldehyde assimilation I, glycine betaine degradation I, methylaspartate cycle), aromatic compound catabolism, and nitrogen recycling, whilst Caenophidia showed enrichment in allantoin and glucuronate degradation. This multi-method analysis suggests Burkholderiaceae as a candidate keystone taxon in Henophidia and indicates that phylogenetic clade is a major contributor to cloacal microbiota structure. The lower richness in Caenophidia raises a testable hypothesis that broad-spectrum antimicrobial activity of their venom components may selectively filter susceptible microbial lineages, motivating future shotgun metagenomic studies in wild populations of snakes.

The fish skin microbiome serves as a protective barrier, influencing host health and facilitating interactions between the host and its environment. While several studies have characterised the composition and roles of the fish skin microbiome, there remains a paucity of data on how environmental variation influences these microbes in natural populations. Here, we used 16S rRNA gene sequencing to characterise the skin microbiome of wild three-spined stickleback populations and examine how environmental factors influence microbial diversity and community composition across 17 freshwater lochs on the island of North Uist, Scotland. Analysis of 239 samples revealed a set of dominant bacterial genera commonly associated with other fish species, including Janthinobacterium, Pseudomonas, Acinetobacter, and Psychrobacter, that constituted a core skin microbiota across lochs. Microbiome composition was primarily shaped by environmental variables, particularly habitat, water pH, conductivity, and metal concentrations, with pH emerging as a key driver of community structure. Host sex also influenced microbiome variation, with several taxa differing in relative abundance between males and females. Alpha-diversity was higher among stickleback fish from lochs with a neutral pH compared with those from alkaline and acidic environments. Differential abundance analyses identified 27 and 24 amplicon sequence variants (ASVs), respectfully, associated with variations in pH and host sex, including members of Psychrobacter, Sphingobacterium, Carnobacterium, Chryseobacterium, and Arthrobacter, highlighting the combined influence of environmental and host factors on microbiome composition in wild fish populations in freshwater environments.

Corals are complex holobionts, encompassing numerous prokaryotes, viruses, and protists. This associated microbial community strongly influences coral health and its resilience to global ocean warming. Corallicolid apicomplexans are widespread coral-infecting parasites, yet their impact on the coral host remains poorly understood. This knowledge gap largely stems from the low abundance of these parasites in coral tissues, which makes them difficult to isolate and access to their genetic material. Here we analyzed nearly 1,000 Pocillopora coral colonies collected from 32 islands during the Tara Pacific expedition to identify the drivers of corallicolid prevalence and abundance. Corallicolids were detected in almost all Pocillopora colonies with variable relative abundances between islands. The high abundance of specific corallicolid populations correlates with seawater temperature and levels of host protein carbonylation. We used a large collection of 297 metatranscriptomes to assemble a corallicolid transcriptome and we identified apicomplexan parasite signature genes, including the GRA9 and PV2 confirming the close phylogenetic relationship with the family of Eimeriidae. Gene expression patterns indicate that the high abundance of corallicolids correlates with a high transcription of genes encoding apical complex proteins and genes involved in the control of host immune defenses. Overall, this study provides new insights into corallicolid biology and its interaction with the coral host by combining a newly generated transcriptome with a large-scale sampling of Pocillopora corals across the Pacific Ocean.

A hosts diet can alter the course of parasite infection. This is especially true of trophic parasites, which a host acquires through feeding. While a large body of work attests to the role of diet in the spread of disease within-hosts, diet can also impact host density and encounter rate with parasites, both of which are expected to modify disease dynamics. When parasites are acquired through feeding, epidemics may be larger and more severe on high-quality diets if these diets support a higher density of hosts that feed more and thus ingest more parasites. Alternately, epidemics may be more severe on low-quality diets if malnourishment decreases hosts ability to resist disease. To differentiate these hypothesized effects of diet on disease, we quantified individual infections and epidemic dynamics for the natural microsporidian parasite Nematocida ironsii infecting its nematode host Caenorhabditis elegans. We measured feeding rate, parasite transmission, and host fitness across three bacterial diets that vary in quality and elicit distinct feeding behaviors in C. elegans. We found that low-quality diets reduced feeding rate, which corresponded to reduced acquisition of parasite spores. However, these diet-mediated differences in parasite acquisition did not directly map onto fitness consequences: hosts eating the poor-quality diet had similar reductions in fitness to those on higher quality diets. During epidemics, a combination of increased parasite acquisition and higher population growth rates resulted in higher parasite abundance for hosts on high-quality diets. Our work underscores the importance of considering both individual- and population-level impacts acting in concert to determine how diet affects the spread of infectious disease.

Ocean warming is altering abiotic environments and biotic interactions experienced by marine organisms, where sensitive early developmental windows occur in biologically complex seawater communities. The impact of these interactions on developmental processes and fitness in hosts is not well understood, but likely contingent on the establishment of a host-associated microbiome. Here, we hypothesize that temperature and microbial exposure during embryogenesis influence larval microbiome assembly and host morphology. Strongylocentrotus purpuratus embryos were raised in low microbial richness (LMR) or high microbial richness (HMR) seawater at ambient (14 {degrees}C) or elevated (18 {degrees}C) temperature, then collected at 2, 4, and 6 days post-fertilization (dpf) following multiple feedings. Higher microbial diversity was observed in larvae that developed in HMR seawater when compared to LMR. Differences in relative abundances of dominant microbial families between seawater and larvae suggest some degree of host selectivity in microbiome assembly. Temperature did not strongly alter microbiome composition, but both temperature and microbial condition led to differences in larval morphology by 6 dpf, potentially due to enrichment of microbes with chemoheterotrophic functions. By linking how temperature and microbial communities interact with host development, we contribute novel insights into how early-life environmental conditions impact holobiont formation and morphology. One sentence summaryEarly developmental temperature and microbial conditions shape larval microbiome establishment and morphology.

BackgroundThe Pacific oyster (Crassostrea/Magallana gigas) is increasingly recognised as a model marine invertebrate. Valued for both ecological and commercial importance, Pacific oysters are farmed widely, supporting global food security by providing a sustainable nutrient-rich source of protein. Despite the significant and recurring economic losses caused by Ostreid herpesvirus (OsHV-1) outbreaks, only a limited number of studies have examined host-pathogen interplay at single-cell resolution. The few available studies largely focus on circulating immune cells (haemocytes), thereby overlooking the complexity of host responses across different tissues and organs. ResultsWe present a detailed single-nucleus transcriptomic atlas of the whole Pacific oysters, including during OsHV-1 infection. A total of 18 distinct transcriptomic clusters were resolved, capturing major cell populations from the gill, mantle, hepatopancreas, adductor muscle, and haemocytes. Notably, three populations- gill ciliary cells, hepatopancreas cells, and an immune-enriched cluster 1- exhibited pronounced transcriptomic responses to OsHV-1 infection. Across the 6, 24, 72, and 96 hours post-infection (hpi) time course, viral transcripts were detected almost exclusively at 72 hpi, with enrichment primarily in adductor muscle cells and two immune cell populations- immature haemocytes, and hyalinocytes. ConclusionsOur findings suggest potential entry portals and tissue-specific replication sites for the OsHV-1 virus in Pacific oysters. This atlas resource provides a high-resolution cellular framework for understanding host-virus interactions and establishes a foundation for future investigations into herpesvirus pathogenesis in marine invertebrates.

Across the United States, backyard poultry (BYP) are becoming increasingly popular as a food source as well as pets. Unfortunately, they have also been a source of annual human salmonellosis outbreaks for over a decade. Previous CDC analyses suggest baby poultry are the main source of live poultry-associated outbreaks as opposed to adult birds. However, there are few data on the frequency of pathogens, such as Salmonella enterica, in baby poultry sold to the BYP market. Further, there is a lack of data on the serovars and antimicrobial resistance (AMR) rates in these baby poultry. We collected 643 soiled bedding and shipping box samples from agricultural supply stores primarily located in Vermont. S. enterica was detected in 23.5% (151/643) of samples, with the highest rates of detection in 2021-2022. Rates of S. enterica varied by species. Turkey poult bedding samples had the highest rates of S. enterica (44.4%; 8/18), while laying chick bedding samples had the lowest (19.4%; 68/350). Meat chick bedding samples had an intermediate rate, at 36% (32/89). The most common serovar detected was Salmonella Enteritidis, which represented 51.2% (64/125) of sequenced isolates. AMR genes or AMR-associated point mutations were detected in 21.6% (27/125) of samples, but only in non-Enteritidis serovars. These data indicate that baby poultry intended for the BYP market pose a substantial risk of salmonellosis to consumers.

The human duodenum harbours a complex, dynamic microbial community that is challenging to study due to inaccessibility, particularly postprandially when nutrient-rich chyme and fluctuating metabolites create unique microbial niches. We used naso-duodenal intubation to longitudinally sample duodenal luminal contents following pea-based meals of differing food structure, alongside parallel blood collection. Shotgun metagenomic sequencing, comprehensive metabolomic profiling and gut hormone measurements were combined to explore microbe-metabolite-hormone interactions. Food structure significantly affected postprandial bacterial composition, with saccharolytic oral taxa increasing after meals with intact structure. Alpha diversity was influenced by structure type (P = 0.025), with whole pea seeds promoting greater diversity than pea flour. Network analysis revealed complex interactions between the duodenal microbiome, luminal metabolites and gut hormones, with most microbial associations linked to glucose-dependent insulinotropic polypeptide (GIP) rather than glucagon-like peptide-1 (GLP-1). Metabolic profiling showed meal-dependent changes in amino acid metabolism, including shifts in D/L amino acid ratios over time consistent with microbial metabolism. The duodenal microbiome showed close phylogenetic relationships with the oral microbiome, with composition influenced by food structuring and swallowing. These findings reveal dynamic microbe-metabolite interplay in the human duodenum during digestion and its relationship to gut hormone responses.

Coral reefs deliver vital services via a complex three-dimensional framework sustained by the balance between calcium carbonate production and erosion, or the net carbonate budget state. In many tropical western Atlantic reefs, ecological decline has reduced carbonate production, yielding near-neutral or negative budgets. Yet some reefs retain high coral cover and, theoretically, should also have high net positive budgets, yet often show modest carbonate accumulation. We used the remote reef of Cayo Arenas in the Campeche Bank, Gulf of Mexico, to test whether in reefs under suboptimal (variable) environmental conditions, high coral production is offset by robust bioeroder communities, producing neutral budgets. At 14 sites, we quantified carbonate producers and bioeroders to estimate gross production, bioerosion, and net budget states. Despite relatively high live coral cover, mean net carbonate budgets were approximately neutral. Crucially, this neutrality arose not from depressed biological activity (as in degraded reefs) but from an active equilibrium: vigorous carbonate production coupled with substantial bioerosion. These reefs, therefore, represent a contemporary, functional reef state in net stasis. Distinguishing active-neutral from impoverishment-neutral regimes is critical for predicting reef trajectories under environmental change and for targeting management, although near-stasis emerging from high carbonate turnover can appear functionally intact yet operate with limited buffering capacity against net carbonate loss.

BackgroundIgA is the dominant antibody in the human gut and a key regulator of host-microbe interactions. Infants begin to produce IgA at around 6 months old and receive large quantities of IgA via human milk, but technical limitations have prevented species-level characterization of IgA binding in early life. This has left basic knowledge gaps about which species are targeted by IgA in infancy, and how modifiable lifestyle factors like breastfeeding and complementary feeding impact IgA targeting. ResultsHere we adapt Metagenomic Immunoglobulin Sequencing (MIg-Seq) for low-biomass infant fecal samples and apply this optimized protocol to 32 longitudinal samples from 16 infants enrolled in the MINT trial, a four-arm randomized controlled trial comparing meat-based, dairy-based, plant-based, and reference complementary feeding patterns, with fecal sampling at 6 and 12 months (pre and post intervention). Infant IgA targeting mirrors adults at the phylum level, with both age groups showing significantly higher IgA targeting of Pseudomonadota and lower targeting of Bacteroidota relative to other phyla. During the substantial microbiome compositional shifts noted between 6 and 12 months, IgA targeting is significantly more stable than the microbiome itself. Among persistent colonizers, IgA targeting strengthens significantly from 6 to 12 months, with the most pronounced effect observed for Bifidobacterium, a finding robust across all dietary arms and feeding modes. The feeding arm to which infants were enrolled was not significantly associated with IgA binding, but several nutrient-specific associations were discovered. Animal-derived nutrients, particularly cholesterol, are strongly positively correlated with IgA targeting of Bifidobacterium longum, while plant-derived carotenoids are positively associated with IgA targeting of Flavonifractor plautii and Ruminococcus gnavus. ConclusionsThis study introduces an experimental and computational framework for species-level IgA profiling in the infant gut. The progressive strengthening of IgA targeting of Bifidobacterium and other beneficial persistent colonizers suggests a role for IgA in reinforcing beneficial microbes during infancy. The nutrient-specific dietary effects on IgA targeting reveal the immunological consequences of the complementary feeding period, and highlight a contrast between animal-versus plant-based diets. Together, these findings point to early nutritional interventions and IgA-based therapeutics as promising tools for promoting healthy immune-microbiome development.

Mycobacterium abscessus (Mab) is an opportunistic pathogen that can cause chronic, debilitating lung disease. Mab is intrinsically resistant to most antibiotics, making Mab infections challenging to manage and frequently incurable. During infection, Mab adapts to survive various stresses, including hypoxia and nutrient starvation. In vitro, these conditions drive Mab into a drug-tolerant, non-replicating state. Changes in the Mab proteome that result from entering a non-replicating state have been minimally described despite the clinical importance of this physiological state. Using Mab reference strain ATCC 19977, we collected proteomic data comparing replicating to non-replicating states using a carbon starvation (CS) model of persistence. We identified 2251 proteins overall (46% proteome coverage), and 17% of these proteins were found in only one of the two conditions. A third of identified proteins were significantly changed in abundance, indicating an extensive proteomic response to CS. The response regulator DosR and many DosRS responsive proteins were significantly more abundant under CS, suggesting that the DosRS stress response regulator plays a key role in CS-induced Mab persistence. Many aspects of cell wall biosynthesis were changed, including changes in glycolipid abundance under CS. Proteins involved in other key cellular processes such as secretion, oxidative phosphorylation, and nutrient metabolism were altered under CS. The proteomic analysis presented provides new insights and clarity into how the Mab proteome is regulated during non-replicating persistence, a key consideration for understanding Mab pathophysiology.

The relationship between Campylobacter levels in broiler caeca and on carcass skin is central to quantitative microbial risk assessment along the poultry production chain, underpinning modelling of intervention impacts, including EFSA assessments of the public health impact of control measures. However, this relationship is typically inferred from monitoring data generated under sampling designs that do not preserve pairing between specimens and may involve pooling. In this study, we used a simulation framework to evaluate whether commonly used sampling strategies allow reliable recovery of the caecal-skin relationship. A simulated broiler population was generated, assigning caecal and skin loads to individual birds based on a specified linear relationship. Sampling was conducted under paired and unpaired designs, with and without pooling, reflecting approaches used in surveillance programmes and in policy-oriented models. Regression models were fitted to sampled data across 1,000 simulations for a range of assumed slopes. Under paired sampling, estimated slopes closely matched the true relationship across most scenarios. In contrast, unpaired sampling consistently failed to recover the association, with estimated slopes centred around zero regardless of the true slope. These findings were robust to variation in within-flock prevalence, residual error, and intercept. The results show that sampling design fundamentally affects identifiability of relationships between stages of the production chain. This has implications for interpretation of parameters derived from monitoring data and used in quantitative Campylobacter risk assessments informing policy. Parameters derived from unpaired and pooled monitoring data should therefore be interpreted with caution when used to support risk assessment and decision-making. Campylobacter; broiler chickens; sampling strategy; unpaired sampling; carcass contamination; quantitative microbial risk assessment; simulation.

Industrialization has been identified as the single biggest factor driving global microbiome diversity. While many studies examining gut microbiomes attribute these shifts to dietary increases in fat and reductions in protein, oral microbiome responses to industrialization remains debated. The oral microbiome is more resilient due to long-standing coevolution with host tissues and biofilm stability. However, limited geographic and historical representation has constrained our understanding of how these transitions unfolded globally in the oral microbiomes. Here, we investigate oral microbiome variation in Batwa rainforest hunter-gatherers and neighboring Bakiga subsistence farmers from southwestern Uganda, comparing them with publicly available data from Tanzanian, Venezuelan, and industrialized populations from North America, Europe, and Australia. Using 16S rRNA gene sequencing, we characterized salivary microbiota and evaluated differences in local and global diversity, composition, and differential abundance. Ugandan populations contained significant compositional differences but similar levels of diversity, suggesting that shared environments and dietary overlap may shape microbial assemblages despite distinct cultural histories. Globally, strong continental and industrialization effects were observed in the oral microbiome, with all industrial populations clustering separately from people living in other locations. African populations also clustered separately from non-African groups. Oral microbiome diversity was highest in Ugandan individuals and lowest in industrialized populations, mirroring patterns previously observed in the gut microbiome. Together, these findings demonstrate that both geography and subsistence strategy structure global oral microbiome variation. They also clarify the position that oral microbial communities record biocultural transitions and highlight the need to better understand the industrial mechanisms that shape microbial diversity in the oral cavity.

Understanding risks of biological invasions associated with ballast water (BW) requires full understanding of the biodiversity transported in ballast tanks. Here we characterize the remarkable level of diversity that can be carried in the BW of a single vessel. To maximize our ability to capture BW diversity we: 1) utilized DNA-based methods to describe biodiversity, including both native and non-native taxa; 2) exploited multiple primer sets targeting multiple genomic loci with different expectations for taxonomic coverage; 3) assessed multiple tanks on a single vessel to capture different communities present in different tanks; and 4) sampled those tanks with far higher-than-usual replication both to improve representation of the diversity present and to enable statistical estimation of total richness. Using this approach, we found extraordinarily high diversity associated with a single vessel. Across all loci, we estimate a total of 272 taxa that can be assigned species names; looking more broadly at unnamed molecular operational taxonomic units, our estimates are between 425 and 742 individual taxa, depending on the locus. We confirm that only a fraction of this diversity would be captured with typical sampling efforts. We found that different loci capture different snapshots of biodiversity and that our ability to detect taxa of interest (e.g., non-native species) depends on sampling effort and genomic locus. Our results expand upon previous studies describing highly diverse BW communities and add to a growing literature that demonstrates the value of molecular methods for characterizing those communities and assessing the associated risk of non-native species introduction.

Fecal microbiota transplantation (FMT) is an effective therapy for recurrent Clostridioides difficile infection and is increasingly explored for other dysbiosis-related disorders. However, its implementation as a regulated therapeutic strategy still requires robust donor screening, biosafety frameworks, and standardized processing workflows. Here, we describe the establishment of the first fecal microbiota biobank in the south of Brazil and evaluate the incorporation of metagenomic sequencing as a complementary layer of donor safety assessment. A structured donor selection pipeline based on international guidelines was implemented, integrating clinical screening, biochemical and serological testing, and microbiological analyses. Of 100 screened candidates, only four donors met all eligibility criteria and were included in the biobank, highlighting the stringency of the selection process. Shotgun metagenomic sequencing revealed a diverse resistome across all donors, including a shared core set of resistance-related genes alongside marked interindividual variability. Dominant antibiotic resistance genes included tetracycline-associated determinants, as well as ermF, CfxA-type {beta}-lactamases, and aminoglycoside-modifying enzymes, each linked to specific gut taxa. Notably, the relatively high abundance of tetW and ermF in Bacteroides fragilis suggests that this dominant commensal species may act as a reservoir for tetracycline and multidrug resistance determinants within the intestinal microbiota. Rather than serving as exclusion criteria, such determinants highlight the importance of integrating functional genomic profiling into donor characterization. Overall, this study provides a framework for microbiota biobank implementation and supports the use of metagenomics as a complementary strategy to improve biosafety and functional assessment in FMT.

Near-shore coral reefs in southern Guam (Mariana Islands) experience severe sedimentation, in particular during the wet season when rainfall and erosion are high. We sampled fragments of the reef-forming coral Porites lobata from opposite ends of a sedimentation gradient in Fouha Bay, southern Guam, during dry and wet seasons. Using DNA metabarcoding, we characterized the diversity and composition of P. lobata-associated Symbiodiniaceae and bacterial microbiome communities. As in many species of Porites, Symbiodiniaceae communities of P. lobata were dominated by variants of Cladocopium C15 with sites showing differences in Symbiodiniaceae communities attributable to variation in these Cladocopium C15 variants. Bacterial microbiomes of P. lobata were dominated by Endozoicomonadaceae, a family of putative coral bacterial endosymbionts involved in nutrient cycling. Site and seasonal differences in bacterial diversity and community composition were apparent. In close proximity to the mouth of the river draining into Fouha Bay, bacterial diversity was highest during the wet season when sedimentation is generally severe. Microbiome reorganization in response to sedimentation may explain this result, but we also found overrepresentation of bacteria associated with terrestrial origin close to the river mouth and/or during the wet season. Together these patterns highlight that coral Symbiodiniaceae and bacterial communities are both spatially and temporally structured in this disturbed system. IMPORTANCEThis study provides a time series dataset of coral-associated microorganisms, including dinoflagellate algae and bacteria, from a tropical bay impacted by sedimentation that results from upstream erosion of disturbed soils. Characterizing temporal patterns of coral-associated microbes provides insights into the dynamic nature of these communities. While microbiome variability across sites and seasons may be a result of acclimatization to different environmental conditions, we identified bacterial groups of putative terrestrial origin in sampled coral microbiomes that may have been exported from eroded soils to the near-shore reef. Considering that disturbed soils act as hotspots for the proliferation of potentially harmful substances, such as antimicrobial resistance genes, understanding microbial community connections at the marine-freshwater-terrestrial interface is an important step toward evaluating environmental impacts across connected ecosystems from ridge to reef.

Bile acids modulate the intestinal microbiota and serve as key signaling molecules in host physiology. Bile acid dysregulation has been implicated in nutritional and inflammatory diseases; however, data on the pool of bile acids present in stunted children or children suffering of environmental enteric dysfunction (EED) is limited, particularly in the upper intestinal compartment where disease phenotypes are most relevant. In this study, we performed a targeted metabolomics approach on 75 bile acids and their derivatives, including gastric and duodenal aspirates and fecal samples from almost 1000 children from two Sub-Saharan cities. We found that levels of secondary bile acids, especially lithocholic acid, are significantly lower in the feces of stunted and EED children, while ursocholic acid and its derivatives are significantly higher. Levels of primary and sulfated bile acids are also increased in the feces of children with EED. Microbiota sequencing revealed that high lithocholic acid levels are positively associated with butyrate-producing bacteria, while negatively associated with oral taxa like Streptococcus and Veillonella. In vitro tests on a panel of reference strains showed that oral bacteria bioaccumulate and are inhibited by a variety of bile acids, while lithocholic and chenodeoxycholic acids modulate the growth of several butyrate-producing bacteria. This effect was even stronger with tauro- or glycol-conjugated bile acids. Exposing stool-derived in vitro communities from children in Afribiota to these bile acids confirmed their positive impact on butyrate producers and negative effect on overgrowing oral taxa. Our findings suggest that secondary bile acids, reduced in stunting and EED, modulate the growth of butyrate-producing bacteria while suppressing harmful oral taxa, highlighting their potential as tools to modulate microbiota composition.

Microbiome-based machine learning classifiers show increasing promise for disease identification across gastrointestinal, metabolic, and immune-mediated conditions. Inflammatory bowel disease (IBD), a chronic immune-mediated disorder associated with disruption of the gut microbiome, has been a particularly successful application area. However, while many predictive models achieve high performance within individual datasets, their ability to generalize across independent populations and geographic contexts remains unclear. Here, we tested whether model class and training dataset composition influence model generalizability across geographically diverse evaluation studies. We compiled seven publicly available shotgun metagenomic studies spanning five geographic regions, comprising 697 individuals with IBD or healthy controls. We trained 246,986 model configurations across seven model classes and five distinct training dataset combinations and evaluated top-performing models on independent studies from the USA, Ireland, Germany, Israel and China Extreme gradient boosting and random forest models showed the highest and most consistent performance across training datasets, a ranking that was maintained on independent evaluation studies. However, models trained on geographically diverse datasets did not outperform those trained on USA-only datasets. Instead, model performance was strongly dependent on the evaluation study itself, with consistent differences in achievable accuracy across studies. Despite most models achieving similar AUC scores, there was limited overlap in the key microbial species identified. Furthermore, even for the small set of disease predictive microbes shared between models, the direction of enrichment between IBD or healthy subjects often varied in opposing directions across study populations. These findings suggest that study-specific factors constrain generalization and may help explain the lack of consistent microbiome-based biomarkers for IBD. ImportanceMachine learning models based on the human gut microbiome are increasingly proposed as diagnostic tools for inflammatory bowel disease, but our findings suggest that identifying reliable microbiome biomarkers poses a challenge. Models trained on different datasets often selected different species as important predictors, even when diagnostic performance was similar, indicating that disease-associated microbes may depend strongly on the patient populations studied. Even species repeatedly selected across training datasets frequently showed inconsistent associations with disease, helping explain low agreement across microbiome studies. Importantly, models performed well across new patient groups independent of the geographic diversity present in the training datasets. By identifying microbial species repeatedly selected across datasets, model types, and evaluation studies, we identified a smaller group of more consistent biomarkers, including enrichment of Klebsiella pneumoniae and Erysipelatoclostridium ramosum and depletion of Lachnospiraceae and Alistipes species, which may represent stronger candidates for transferable microbiome markers.

BackgroundWhether the lung microbiome represents a stable microbial colonization or a transient ecosystem shaped by continuous microbial turnover and controlled by host immunity remains unresolved. The murine lung microbiome largely consists of species from the former Lactobacillus genus with Ligilactobacillus murinus as a dominant species, bacterial genera such as Streptococcus, Staphylococcus, Mammaliicoccus, Enterococcus and other less frequently detected bacteria. Here, we directly addressed the question of persistence and host interaction of a dominant murine lung commensal in vivo and focused on the host immune response towards lung commensal bacteria. ResultsWe developed a transformation strategy for stable genomic integration of a green fluorescent protein (GFP)-encoding gene to track the fate of a lung bacterium. Following intranasal administration of GFP-labeled L. murinus in mice, bacteria were readily detected in the lungs at early time points but declined rapidly and became undetectable after 72 hours, as determined by quantification of viable bacteria and qPCR. Flow cytometry and fluorescence imaging revealed efficient uptake of GFP-labeled bacteria by lung phagocytes. These findings indicate that even dominant members of the murine pulmonary microbiota normally detected at low abundances are transiently present in the lungs without causing infection. We further analyzed the effects of moderate and high bacterial concentrations. While moderate bacterial loads were efficiently controlled without clinical effects, high concentrations induced severe lethargy, indicating a threshold-dependent host response. Finally, we demonstrated that pulmonary commensals such as L. murinus, Staphylococcus xylosus, and Mammaliicoccus sciuri, as well as conidia of the opportunistic lung pathogen Aspergillus fumigatus, are phagocytosed at comparable rates in macrophage assays. ConclusionsOur data demonstrate that even lung-adapted bacterial species fail to establish stable colonization and are instead subject to rapid immune-mediated elimination contributing to the maintenance of a low microbial burden in the lungs. While this homeostatic balance supports health, elevated bacterial loads trigger immune activation and, at high levels, lead to health deterioration. Together, these results support a model of a highly dynamic and transient lung microbiome, maintained by continual microbial immigration rather than long-term colonization. Accounting for the lung microbiome dynamics is essential for understanding host-microbiota interactions and respiratory health.

Colibacillosis, caused by Avian Pathogenic Escherichia coli (APEC), result in substantial economic losses in global poultry production. The emergence of multidrug-resistant (MDR) APEC poses zoonotic risks through horizontal transfer of antimicrobial resistance (AMR) genes. Bacteriophage therapy emerges as a safe alternative to antibiotherapy; however, comprehensive characterization of phages targeting MDR-APEC from diverse geographical regions remains limited. We isolated five lytic bacteriophages from poultry fecal samples collected from five Indian states and characterized them through morphological analysis, physiological stability testing, whole-genome sequencing, and in vivo efficacy assessment. Host range was determined against APEC isolates, and therapeutic potential was validated in Galleria mellonella infection model. All five phages showed Myovirus-like morphology and stability across physiologically relevant temperatures (up to 55-70{degrees}C) and pH conditions (3-11). Their genome size ranges from 170 to 356 kb, belonging to three distinct genera; Dhakavirus, Gaprivervirus, and Asteriusvirus. Genomic analysis confirmed absence of antimicrobial resistance, virulence, toxin, or lysogeny genes. 51 APEC strains were isolated, of which 23 (45.1%) were MDR. Individual phages lysed 37-51% of tested APEC and 17-39% of MDR strains. Three Escherichia phages (fBSZT1, fUAMT1, fPKPT2) significantly improved larval survival to 60-80% at MOI 10 in G. mellonella infection models compared to untreated controls. This study establishes a well-characterized phage bank targeting MDR-APEC strains, providing foundation for developing phage-based interventions to reduce antibiotic dependency and mitigate AMR transmission risks under One Health framework.