
==== Front
mSystems
mSystems
msystems
mSystems
2379-5077
American Society for Microbiology 1752 N St., N.W., Washington, DC

39120143
msystems00636-24
10.1128/msystems.00636-24
msystems.00636-24
Research Article
microbial-ecologyMicrobial EcologyDemographic and zoological drivers of infectome diversity in companion cats with ascites
https://orcid.org/0000-0001-6283-4172
Sun Yankuo 1 2 3 Funding acquisition Methodology Project administration Software Writing – review and editing
https://orcid.org/0000-0002-1212-265X
Xing Jiabao 1 Conceptualization Data curation Formal analysis Investigation Methodology Software Validation Visualization Writing – original draft Writing – review and editing
Xu Sijia 1 Data curation Methodology Writing – review and editing
Li Yue 1 Data curation Resources
Zhong Jianhao 1 Data curation
https://orcid.org/0000-0002-2718-6349
Gao Han 1 Data curation
Cheng Song 4 Data curation
Dong Jun 4 Data curation
Zhang Tianyou 4 5 Data curation
https://orcid.org/0000-0002-1870-4490
Lu Gang 1 Writing – review and editing LG@scau.edu.cn

https://orcid.org/0000-0002-1915-7732
Baele Guy 6 Formal analysis Writing – review and editing guy.baele@kuleuven.be

https://orcid.org/0000-0003-1236-1850
Zhang Guihong 1 2 3 Funding acquisition Supervision Writing – review and editing guihongzh@scau.edu.cn

1 Key Laboratory of Zoonosis Prevention and Control of Guangdong Province, College of Veterinary Medicine, South China Agricultural University , Guangzhou, China
2 Maoming Branch, Guangdong Laboratory for Lingnan Modern Agriculture , Maoming, China
3 Guangdong Laboratory for Lingnan Modern Agriculture , Guangzhou, China
4 CAU Dong Jun laboratory, Guangzhou, China , Guangzhou, China
5 Guangzhou Chimelong Safari Park , Guangzhou, China
6 Department of Microbiology, Immunology and Transplantation, Rega Institute, KU Leuven , Leuven, Belgium
Editor Ishaq Suzanne Lynn The University of Maine , Orono, Maine, USA

Address correspondence to Gang Lu, LG@scau.edu.cn
Address correspondence to Guy Baele, guy.baele@kuleuven.be
Address correspondence to Guihong Zhang, guihongzh@scau.edu.cn
Yankuo Sun and Jiabao Xing contributed equally to this article. Author order was determined by alphabetical order of family name.

Guy Baele and Guihong Zhang are joint senior authors.

The authors declare no conflict of interest.

9 2024
09 8 2024
09 8 2024
9 9 e00636-2406 5 2024
21 6 2024
Copyright © 2024 Sun et al.
2024
Sun et al.
https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.

ABSTRACT

Cats (Felidae) have become an integral part of many households. However, our understanding of the full spectrum of pathogens affecting cats (referred to as the infectome) is limited, mainly due to the inadequacy of commonly used diagnostic tools in capturing the complete diversity of potential pathogens and the prevalence of pathogen co-infections. In this study, we employed a meta-transcriptomic approach to simultaneously characterize the infectome contributing to different disease syndromes and to investigate spatial, demographic, and ecological factors influencing pathogen diversity and community composition in a cohort of 27 hospitalized cats and seven stray cats. We identified 15 species of pathogens, with Candidatus Rickettsia tarasevichiae and Tritrichomonas foetus representing potential spillover risks. Importantly, although most cases of ascites hyperplasia were explained by coinfection with multiple pathogens, we identified the potential novel clinical outcomes of M. aubagnense infection among cats. We demonstrated that the increase in infectome diversity can be explained by a variety of predictors including age growth, temperature increase, and a higher proportion of females, with age growth presenting the strongest effect. Fine-scale analysis indicated that a higher diversity of infectomes were harbored in young cats rather than adult ones. Our results demonstrated that most feline diseases are better explained by the presence of virus–bacteria or virus–virus coinfection. This study serves as a timely endorsement for clinical diagnosis by vets to consider the cause of a disease based on a panel of cryptical co-infecting pathogens rather than on individual infectious agents.

IMPORTANCE

Frequent studies reported the risks of cats as an intermediate host of zoonotic pathogens (e.g., SARS-CoV-2). Cats have a physically close interaction with their owners through activities like petting, kissing, and being licked on the cheek and hands. However, there are still limited studies that systematically investigate the infectome structure of cats. In this study, we employed a meta-transcriptomics approach to characterize 15 species of pathogens in cats, with Candidatus Rickettsia tarasevichiae first characterizing infection in diseased cats. Most feline diseases were better explained by the presence of virus–bacteria or virus–virus coinfection. The increase in infectome diversity could be influenced by a variety of predictors including age growth, temperature increase, and a higher proportion of females. A higher diversity of pathogens was harbored in young cats rather than adults. Importantly, we showed the value of linking the modern influx of meta-transcriptomics with comparative ecology and demography and of utilizing it to affirm that ecological and demographic variations impact the total infectome.

KEYWORDS

infectome
meta-transcriptomics
demography
zoology
infectious diseases
cats
Key-Area Research and Development Program of Guangdong Province 2019B020211003 Xu Sijia Li Yue Zhong Jianhao Gao Han Cheng Song Dong Jun Zhang Tianyou Lu Gang Zhang Guihong MOST | National Natural Science Foundation of China (NSFC) 32102704 Baele Guy CHINA AGRUCULTURE RESEARCH SYSTEM OF MOF AND MARA CARS-35 Baele Guy Fonds voor Wetenschappelijk Onderzoek - Vlaanderen G0E1420N, G098321N Baele Guy Internal Funds KU Leuven C14/18/094 Baele Guy cover-dateSeptember 2024
==== Body
pmcINTRODUCTION

With the acceleration of modern-day life pace and a decrease in interpersonal interactions, pet ownership has become increasingly common (1). Domestic cat (Felis catus) is now considered an integral part of the family unit by many individuals. Recent research suggests a complex interplay between cats and their owners, with many people experiencing significant improvements in psychological wellbeing through the ownership and physical interaction with companion animals, such as petting, kissing, and being licked on the cheek and hands (2–8). Despite the psychological importance of cats in the family unit, there is a lack of systematic studies focusing on the risks of infectious diseases that affect cat and human health. Although it is now relatively easy to genetically characterize the feline virome and discover new virus sequences in cats, most studies provide only a limited understanding of the circulating viral genomic diversity in cats due to infrequent, nonsystematic, and spatially limited sampling of target species (9–15) . Consequently, less attention has been paid to investigating the simultaneous presence and dynamics of pathogens in cat populations. Moreover, the reasons for the presence of certain viruses in some cat species or populations at certain geographic sites but not in others remain unclear. In addition to the cat virome, the interactions between cats and humans present a potential pathway for zoonotic pathogens to spread to caregivers (16). For instance, in a study assessing SARS-CoV-2 infection status from 919 companion animals in Europe, although none of the animals tested positive by the PCR, 5.8% of cats had measurable SARS-CoV-2 neutralizing antibody titers (17). A similar study in Thailand subsequently pointed out that SARS-CoV-2 RNA was present in 19 cats, suggesting suspected episodes of human infection from cats that were initially infected through contact with infected humans (18, 19).

Recent molecular and serological investigations of cat viral infectious diseases in China from 2016 to 2019 have shown that most of cats (approximate 80%) tested positive for at least one virus (20). Within these, FHV-1 and FPV showed significant seasonal prevalence. In addition to this, genomic epidemiological investigation suggests that different genotypes of feline calicivirus (FCV) circulated regionally in Kunshan (China) (21). While viral discovery studies provide valuable insights into the evolutionary history and epidemiology of pathogens, they offer a limited understanding of the factors shaping the cat pathogen spectrum (22, 23). To gain a better understanding of microbial dynamics among cat populations, it is important to move beyond descriptive host–pathogen associations and strive for a mechanistic understanding of when and where infectious pathogens are transmitted, as well as how the entire pathogen community (infectome) is influenced by the environment and local host communities (24, 25). For instance, changes in parasite richness in wild animals have been linked to habitat loss and fragmentation, indicating that anthropogenic alterations in host species composition and population densities can directly impact parasite community compositions (26–31). Similarly, anthropogenic land-use changes have been observed to influence virus community compositions, suggesting their role as a key determinant of host viromes (32–35). These findings emphasize the importance of studying community traits, such as parasite richness, to understand and predict zoonotic risks spatially and temporally.

However, a major challenge in characterizing unbiased microbial communities is the lack of an approach to acquire all genetic information within a single specimen. Previous methodologies were often based on removing as much nucleic acids outside viral particles as possible by filtering, centrifugation, lysis, and nuclease treatment, although this seldom results in a complete depletion of host RNA (36–38). In contrast, in meta-transcriptomics, total RNA (i.e., the transcriptome) is directly extracted from untreated homogenates and used for library preparation without filtering and nuclease digestion steps. A key advantage of this over other diagnostic techniques is that it has the potential to detect, in an unbiased fashion, any pathogen that produces an RNA molecule (DNA viruses, bacteria, fungi, and eukaryotes), as well as the obvious case of RNA viruses. Hence, with analysis of appropriate tissues, meta-transcriptomics may provide a one-stop diagnostic shop. Another benefit of meta-transcriptomics is that it provides an organized way to quantify each virus present in a sample. Specifically, the percentage of reads that map to a particular virus genome is a good indication of virus abundance, especially in the context of conserved host genes. In turn, abundance level can provide important indications of disease associations, whether viruses are segmented (such that genomic components have similar or different expression levels), and help identify those viruses that are in fact derived from other eukaryotic organisms present in the sampled host, such as in undigested food or prey, gut microflora, and parasites, or simply contamination (and the greater the virus abundance, the more likely that active viral infection has occurred in the sampled host). In addition, compared to genomic nucleic acids, the transcriptome comprises compact information that is more balanced across domains of life, thereby preventing the overdominance of genetic information from large cellular organisms.

This technique has been successful in revealing the entire pathogen spectrum—including viruses, bacteria, fungi, and parasites (i.e., total infectome)—in a single-infection case, making it a powerful tool for modern pathogen discovery efforts (39–44). For example, a recent meta-transcriptomics analysis of diseased pigs in China revealed that pig diseases were determined with multiple pathogen co-infection (43). The infectome of bronchoalveolar lavage fluid samples from Wuhan (China) before the emergence of SARS-CoV-2 depicted a stable core pathogen spectrum without the presence of SARS-CoV-2 (40). Applying meta-transcriptomic sequencing to small-scale cohorts offers a new opportunity to understand how host ecology and biogeography influence microbial diversity, especially when applied to host species occurring across different habitat types or environmental gradients.

In this study, we have employed meta-transcriptomic sequencing to systematically characterize the infectome communities in cats so that they can be characterized simultaneously in the context of specific syndromes in an unbiased manner (Table S1). A total of 34 individuals were sampled and sequenced from 10 different sites across China. Specifically, 27 clinical cases were sampled from pet hospitals with a detailed context of the specific syndrome, and seven stray cats were sampled around pet hospitals by recording the number of local stray cats through in-depth interviews. Using these data, we have (i) characterized the total pathogen community with transcriptional loads and genomic information of each case; (ii) assessed pathogen interactions, potential clinical manifestations, and epidemiological impacts; and (ii) tested how demographic and environmental factors (Table 1) influence the hospital‐level infectome community of clinical cases, and whether the population density of stray cats influences pathogen richness.

TABLE 1 Ecological covariates that may influence infectome diversity and community composition in hospitalized and stray cats

Hypothesized covariate	Tested response variables	Reasonable hypothesis of covariates on infectome diversityc	Reference(s)	
Age	Month	↑
With increasing age, more asymptomatic infections were established	(27, 45–48)	
Location	Latitude (longitude excluded due to over-correlation with latitude and precipitation; Fig. S1)	NA
Latitude effects encompass a number of covariates due to the influence of physiography	(49–51)	
Population density	No. of other living stray companions per hospital (5 km2 scale)	↑
With increase in quantity, greater microbial persistence within integral colonies and increased viral encounter probability from different individuals	(28, 29)	
Sex	Females	↑
Females are more susceptible to infections due to behavior and physiology (testosterone)a	(27, 52–54)	
Humidity	Mean humidity	↑
Sites with high humidity tend to be negative for pathogen retention	(55, 56)	
Precipitation	Average annual rainfall	↓
Sites with high rainfall tend to be negative for pathogen retention	(57)	
Temperature	Local temperature when sampling	↓
Too high a temperature may not be beneficial to pathogen retention and spreading, but positive for intrahost persistence	(57)	
Living condition	Stray/hospitalized	Stray cats may harbor more pathogens or with divergent abundance compared with hospitalized cats	(1, 58)	
COXI reads	COXI RPM	↓
Test if acquired host-relevant data size may take over the pathogen data size	–b	
Raw reads	QC reads	↑
Test if sequencing data size may influence the pathogen data size	–	
a Elevation in testosterone levels can attenuate inflammatory responses by activating regulatory T cells, which in turn suppress Th17-type reactions, thereby exerting a modulatory effect on the immunological landscape (59).

b “–” indicates no reference provided in this covariate.

c "↑" and "↓" indicate that we suppose these covariates positively and negatively correlate with infectome diversity, respectively. "NA" indicates that the effects of this variable are diverse and may not be as expected.

MATERIALS AND METHODS

Sample collection and metadata variables

From April 2022 through April 2023, we collected a total of 34 diseased cats from 10 cities in China, among which 27 had required hospitalization (Table S1). Seven of the 34 stray cats around the pet hospitals were also included in this cohort, aiming to understand the infection spectrum prevalent in the stray feline population proximal to hospitals, determine the diversity in the infectome profile of these stray cats, and compare it to hospitalized cats. Hospitalized cats included in our cohort were sampled at the initial diagnosis and admitted into the hospital, while those cats with secondary sampling or long-term hospitalization were not considered. For each case, we only collect a single sample that is unique to the neighborhood where the animal hospital is located to avoid a certain degree of empirical bias caused by subjective sampling. The severity of symptoms observed in these cases varied, ranging from mild cough to complication syndromes (i.e., fever, severe cough with expiratory dyspnea, dermatology diseases, diarrhea, and ascites hyperplasia) (Fig. 1A; Table S1). The definitions of syndrome criteria are as follows: diarrhea is defined as the passage of one or more semi-liquid or watery bowel movements, accompanied by weight loss and anorexia. Ascite hyperplasia is characterized by the presence of pleural fluid and progressive painless abdominal enlargement, which could progress to asthma or dyspnea, along with progressive muscle wasting on both sides of the spine and abdominal enlargement. Fever is defined as exceeding a temperature threshold of 39.5°C. Severe cough is described as continuous rattling sounds accompanied by back arching and chest contraction, possibly accompanied by a small amount of foam. Other possible surgically treated diseases caused by infectious pathogens, such as cutaneous diseases, acne, and dermatomycosis, were defined as dermatology diseases. However, cases involving surgical and internal medicine, such as fractures, nephrological diseases, and heart diseases, were not included in this study. Detailed symptom information for each case is provided in Table S1.

Fig 1 Overview of clinical cases analyzed in this study. (A) Phylogeny of cats based on the cox1 gene. The consensus phylogenetic tree was estimated using nucleotide sequences of the cat cox1 gene utilizing the maximum likelihood method, with 1000 replicates, with the phylogenetic tips colored by the sampling source (sequenced in this study: red, publicly available: blue). The sampled location was linked with the phylogenetic position based upon the precise location. The map was created using the R package mapdata (mapdata: Extra Map Databases, R package version 2.3, 2022). (B) Clinical diagnosis and treatment information of pet hospitals. We have specifically monitored a case of ascites hyperplasia whose etiology is the amino acid mutation of the spike gene (FIPV) in 1,058 and 1,060 (see Fig. S7).

A phylogenetic tree features cat evolution. Radiographs and photos feature clinical data from a case with ascites hyperplasia, potentially caused by a feline virus mutation.

For the sampling category, we referred to the following protocol: for the individuals who showed clinical symptoms of ascites hyperplasia, we collected 1 mL of ascitic fluid using 2-mL medical injectors (22 specimens included) and transported the sample to the laboratory in a dry ice environment followed by exposure to the DNA/RNA Shield (Zymo, California, USA) mixture (1:1). For individuals experiencing other syndromes such as vomiting, fever, diarrhea, and inappetence, we collected the whole blood samples using DNA/RNA Shield Blood Collection Tube (Zymo, California, USA) to prevent blood clots, which were then transported in a dry ice environment to the laboratory (12 specimens included). The samples were stored in a −80°C refrigerator before subsequent processing. Through the sampling process, we complemented relevant environmental metadata, enabling us to explore the potential impact of covariates on pathogen diversity (for details of metadata. refer to the supplemental material). The rationale behind the selection of both sampling types arises from the need to mitigate the variability of infectome abundance that arises during the sampling process due to differential abundance profiles associated with different sample types (such as rectal swabs from diarrheal and ascitic cats or nasal swabs from respiratory tract infection cats).

RNA extraction, library construction, and sequencing

Total RNA was extracted from the processed homogenate using the TRIzol reagent (Invitrovgen, Carlsbad, USA). No enrichment of microbial content was applied during sample processing as total transcriptomics enables to reveal the entire microbiome in a sample in an unbiased approach. RNA quality was assessed by using an Agilent 2100 Bioanalyzer (Agilent Technologies, California, USA), and the RNA quantity was quantified using Qubit 4.0 fluorometer (Invitrogen) before library construction. RNA solutions were stored at −80°C until use. Ribosomal RNA (human/mouse/rat) was removed using the Ribo-MagOff rRNA Depletion Kit (Vazyme, Nanjing, China) during the library construction, and all the 150-bp pair-end sequencing libraries were constructed using the VAHTS Universal V8 RNA-seq Library Prep Kit for Illumina (Vazyme). The library alignment’s quality and quantity were further assessed using Qubit 4.0 fluorometer (Invitrogen) and Agilent 2100 Bioanalyzer (Agilent Technologies), respectively, before sequencing. Finally, the qualified libraries were sequenced using the NovaSeq 6000 sequencer (Illumina, San Diego, USA).

Pathogen discovery and characterization

For pathogen discovery and characterization, adapters and low-quality reads (Q30 quality) were first removed using Trimmomatic (v0.39) (60). Pathogen discovery was subjected to a pipeline analysis. All the libraries were either mapped against the non-redundant (nr) protein database using Diamond 0.9.22 with an e cutoff of 10−5 or were assembled de novo into viral contigs using MEGAHIT v1.2.8 before comparison with the nr database (61). For virus characterization, viral contigs whose amino acid content shared <90% similarity with known viruses were characterized as potential novel virus species. Viral genomes were verified by mapping each of the near-complete viral contigs to the corresponding libraries. Final coverage information in the BAM file was inspected using UGENE v40.1(62). The incomplete viral genomes were further validated using Sanger sequencing. Viral abundance information was iteratively called using Bowtie2 and calculated using the formula viral RPM abundance = total viral reads/total non-redundant reads * 1,000,000 (i.e., reads per million of total non-redundant reads [RPM]). For bacterial pathogens, MetaPhlAn2 was run to obtain the initial bacterial taxonomic profiling (63), after which their reference genomes were downloaded from NCBI GenBank as templates to call the BAM files. All contigs were further generated using BAM files and were compared with the nr database for final taxonomic classification at the species level. For fungi and parasites, relevant background mitochondrial genomes were acquired from NCBI GenBank as references to estimate the abundance. A pathogen was considered “positive” in a sample if its abundance level was greater than 1 RPM (40). To identify potential false-positives resulting from index hopping, we used a threshold of 0.1% for pathogens present in the libraries from the same sequencing lane: any read numbers < 0.1% of the most abundant library were treated as “negative,” as previously deployed (40, 44). In addition, marker genes (groEL for Rickettsia and 5.8S ribosomal RNA for Tritrichomonas foetus) were mapped to conduct phylogenetic analysis for determination of evolutionary groups (see Phylogenetic analysis below).

Phylogenetic analysis

We first retrieved all the available reference sequences related to our newly identified virus species from NCBI GenBank (nucleotide database). For example, for feline coronavirus, we searched NCBI GenBank (nucleotide database) using the keywords “feline coronavirus” and downloaded the whole sequence. Duplicated sequences were eliminated in each database with a cutoff of 99% using the CD-HIT program (64). Then, we aligned the viral conserved genes or marker genes with the associated reference sequences using the progressive FFT-NS-i algorithm embedded in MAFFT v7.475. The ambiguously aligned regions were further trimmed using the trimAl algorithm and scrutinized manually (65). Afterward, maximum likelihood (ML) consensus phylogenies were estimated based on the constructed multiple sequence alignments using IQ-TREE, employing the best-matched models according to the Bayesian information criterion (BIC), as measured by ModelFinder (66). All phylogenetic consensus trees were constructed using 1,000 bootstrapped replicates, subsequently mid-point rooted and visualized using Figtree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/) and ggtree in R v4.0.2 (67).

Recombination analysis

We evaluated recombinant events based on Recombination Detection Program 4 (RDP4) using RDP4, Chimera, BootScan, 3Seq, GENECONV, MaxChi, and SiScanto. Furthermore, we have implemented the criterion of the highest acceptable P-value cutoff of 0.05 (P < 0.05), and it was considered a true recombination event, at least when three of the seven detection methods tested positive. Other parameters were carried out by default settings. The recombinant breakpoints were confirmed using SimPlot, with a sliding window of 200 bp (step: 20 bp). All breakpoints were further confirmed using RT-PCR and Sanger sequencing. Specifically, we first conducted recombination analyses of the FIPV genome and FeLV env gene. Given that all FIPV-2 strains showed recombination signals with CCoV-2, we checked the new isolates’ recombination history between FIPV-1, FIPV-1, CCoV-2, and TGEV using RDP4. For FeLV, we used FeLV-A, FeLV-B, FeLV-C, FeLV-E, FeLV-T, and enFeLV as reference groups to examine the recombination history of FeLV_53.

Microbial diversity analysis

To infer the impact of living condition aspects to the pathogenetic microbiota composition, we estimated the α-diversity index using richness, Simpson, and Shannon indices and the β-diversity index using the Bray–Curtis dissimilarity index, using the vegan package (68). Analogously, to assess if the sampling categories may shift the pathogenetic microbiome genera structures between ascites libraries and blood libraries, we also conducted the microbial community componential analysis using α-diversity indices such as richness, Simpson, and Shannon indices and the β-diversity index using the Bray–Curtis dissimilarity index. All the statistical tests of α-diversity indices for each aspect (i.e., living conditions and sampling categories) were done using a Wilcoxon test. All the β-diversity statistics were assessed by one-factor PERMANOVA with 1,000 permutations on the Bray–Curtis matrix, using the pairwise Adonis algorithm inserted into the vegan package. Permutational tests of dispersions using the function permutest.betadisper (999 permutations, pairwise) were performed to assess whether significant effects could be influenced by differences in group dispersion. Analysis of similarity (ANOSIM) was performed using the ANOSIM function implemented in the vegan package, where R > 0 suggests that the intragroup distance is less than the intergroup distance, and the groupings are effective. Statistical significance of PERMANOVA results was assumed when P < 0.05 after the application of a Bonferroni correction (Tables S2 and S3).

Statistical analyses of ecological and demographic correlates of infectome diversity

Generalized linear models with a Gaussian distribution were employed to identify demographic and environmental correlates of infectome diversity (measured by the Shannon index) (Table 1). All relevant environmental and demographic data are detailed in the supplementl material and deposited in Table S5. Prior to undertaking GLM analyses, Pearson’s correlation coefficients between these variables were calculated and visualized for subsequent model selection (5). Submodels were built, which excluded explanatory variables with a Pearson correlation coefficient r > 0.7 (Fig. S1), to aid in addressing multicollinearity, reduce overfitting, improve interpretability, and promote model simplicity. Submodels were compared with the Akaike’s information criterion corrected for a smaller sample size (AICc). Finally, we set age, COXI reads, temperature, precipitation, humidity, sex, living conditions (stray or domesticated), latitude, and raw reads as proxy predictors. Model‐averaged effect sizes and 95% confidence intervals (CIs) were calculated for each explanatory variable using the set of GLMs in which the cumulative Akaike weight summed to 0.95 using the dredge and model.avg functions of the package mumin. Effect sizes were standardized using partial standard deviation to account for multicollinearity. Relative variable importance was calculated as the sum of Akaike weights across all submodels that included each variable. Gaussian-distributed GLMs were checked for overdispersion and the normality of residuals (Table S4). GLM analyses were visualized using jtools in R. Other statistical analyses and plots were performed in R version 4.1.1 (The R Core Team, 2021).

RESULTS

Meta-transcriptomics sequencing and species identification of clinical cases

From the 34 meta-transcriptomic libraries, this study generated approximately 1.2 billion clean reads after quality control and trimming, with a median value of 27,614,731 reads per library (Table S1; Fig. S2). Among these data, we identified contigs associated with the cytochrome c oxidase subunit 1 mitochondrial gene (cox1 gene). The cox1 genes obtained from the 34 sequencing libraries exhibited a high degree of similarity to each other, with nucleotide identities ranging from 97.6% to 100%, representing multispecies co-existence, such as Felis senzu american and Felis silvestris lybica (Fig. 1A).

Overview of the total infectome among clinical cases

Pathogens were identified by comparing sequencing reads directly against nonredundant protein, bacterial genome, and universal Cox1 gene databases (Fig. 2). Identifications were further confirmed using genome mapping and qPCR (or RT-qPCR) assays. In this study, we focused on (i) known pathogens, (ii) opportunistic pathogens (defined as showing pathogenicity when the host immune system is compromised or otherwise favorable), or (iii) uncharacterized viruses related to pathogens that have the potential to cause diseases in mammals. Our analysis of clinical specimens revealed the presence of various pathogens, including bacteria, viruses, and parasites. In total, three RNA viruses, 13 bacteria, and one parasite were identified in 31 out of the 34 cases in our cohort. Among these pathogens were those known to cause feline diseases, which are feline coronavirus (FCoV), feline leukemia virus (FLV), FCV, and a wide spectrum of zoonotic pathogens such as Staphylococcus aureus (S. aureus), Acinetobacter baumannii (A. baumannii), Mycolicibacterium aubagnense (M. aubagnense), Streptococcus pneumoniae (S. pneumoniae), Pseudomonas otitidis (P. otitidis), Haemophilus influenzae (H. influenzae), Moraxella catarrhalis (M. catarrhalis), Salmonella enterica (S. enterica), Cutibacterium acnes (C. acnes), Pseudomonas aeruginosa (P. aeruginosa), Mycobacterium tuberculosis (M. tuberculosis), Staphylococcus hominis (S. hominis), Candidatus Rickettsia tarasevichiae (Candidatus R. tarasevichiae), and Tritrichomonas foetus (T. foetus). Gastrointestinal pathogens, including FCoV (feline enteric coronavirus), T. foetus, S. aureus, and S. enterica, were also identified in our specimens. The majority of the identified pathogens were associated with respiratory infections, such as S. pneumoniae, H. influenzae, M. catarrhalis, and M. tuberculosis. Notably, we found a unique divergent infection case involving a pathogen causing infequent infection in cats, namely, M. aubagnense, which was first reported as the etiology of severe peritoneal effusion in humans in 2023 but was found to be highly abundant in most of the ascites samples (12 out of 17), suggesting its potential role as a neglected factor in causing peritoneal effusion in cats, similar to what has been observed in humans (69).

Fig 2 Total infectome spectrum with abundance information characterized in this study. (A) Prevalence of pathogens identified in this cohort. The bar corresponding to the left y-axis refers to the positive cases sequenced in this study. The line corresponding to the right y-axis refers to the percentage of the positive rate, with 95% CI represented by the transparent area estimated using Poisson rate estimation. (B) The relative abundance (log10-scale RPM) of specific pathogens is denoted in this heat map. The samples (y-axis) were divided into two groups according to the sampling categories, i.e., “ascites” and ‘blood.” They were further divided on the basis of health condition. The pathogens (x-axis) were first divided into two supergroups: viruses, prokaryotes, and eukaryotes.

A graph quantifies the number of cats that tested positive for each pathogen. A heatmap depicts the relative abundance of different pathogen types in various samples, such as ascites, blood, healthy, and unhealthy.

To assess the quantification accuracy, we compared the results obtained from both real-time qPCR (cycle threshold value [CT]) and meta-transcriptomic approaches (read per million, RPM) using a linear regression model. To do this, we designed specific real-time RT-PCR probes and primers for the most frequent infected pathogen (i.e., FIPV). The analysis revealed a strong correlation between the CT values (indicating viral copies) and the read amounts for FIPV (Fig. S3). Pearson’s correlation coefficient (r) was −0.76 (P < 0.05), indicating a reliable quantification of intrahost replicate abundance using our meta-transcriptomic approach (supplemental material; Fig. S3). Given the observed variability in pathogen abundance, we set a threshold of 1 RPM for pathogen identification, which allowed us to accurately determine the presence of pathogens in the samples (40, 43, 70).

Complex evolutionary history of eukaryotic viruses

Although we did not identify any novel viral pathogens in this study, the viruses detected exhibited significant phylogenetic diversity, indicating a complex epidemiological history within companion cats. For example, phylogenetic analysis revealed that all FCoV strains in our cohort belonged to FCoV type I, sharing nucleotide similarities of 92.6% to 99.8% (Fig. 3A). Since FCoV type II is believed to have originated from recombination events between FCoV type I and canine coronavirus (CCoV), we conducted analyses to investigate the recombination history. However, no recombination signal with either FCoV type II or CCoV was detected, supporting our phylogenetic analysis findings (Fig. S4) (71). Within the spike gene, two distinct lineages (clade 1 and clade 2) were identified, and several novel subclades were formed by our strains (Fig. 3B). We also examined specific mutations in the spike protein (M1058L and S1060A), which have been associated with feline infectious peritonitis (FIP) development in FCoV. Among our study genomes, six had the M1058L mutation, while none had the S1060A mutation (Fig. S5) (72). Phylogenetic analysis of the FCV capsid region indicated regional transmission, as our sequences clustered with other enteric sequences from east China (Fig. 3C). Although the env gene of FeLV 53 clustered within the enFeLV lineage, it grouped with the recombined strains, and further recombination analysis confirmed its recombination history in the receptor-binding domain, highlighting its potential risks (73).

Fig 3 Evolutionary history of RNA viruses. Maximum-likelihood phylogenies of different RNA viruses. (A) Phylogeny of FIPV based on the complete genome. (B) Phylogeny of FIPV based on the spike gene. (C) Phylogeny of FCV based on the capsid gene. (D) Phylogeny of FeLV based on the env gene (red: hospitalization, blue: stray), with specific recombination shown in panel E. (E) Recombination analysis of FeLV_53 at the env gene.

Phylogenetic trees depict the evolutionary relationships of RNA viruses, including feline infectious peritonitis virus based on complete genomes, FIPV based on spike gene sequences, and feline calicivirus based on capsid gene sequences.

Genomic analysis of zoonotic pathogens

In addition to identifying viral pathogens causing feline diseases, we also detected several bacterial pathogens with relatively high abundance through meta-transcriptomics analysis (Table S1). Among these findings, we characterized that cats served as a potential novel host for two pathogens. First, we discovered a novel tick-borne mammalian pathogen belonging to the species Candidatus R. tarasevichiae, with the highest amino acid similarity of 93.4% with other sequences available in GenBank at the groEL gene. Phylogenetic analysis showed that the Candidatus R. tarasevichiae in our study clustered within the well-established spotted fever group but displayed significant evolutionary divergence from strains identified in other hosts, such as dogs and ticks (Fig. 4A). This finding is noteworthy as no previous study has reported the infection of Candidatus R. tarasevichiae in cats. The detection of this pathogen, particularly in cases with fever and respiratory manifestations, suggests its ability for cross-species transmission (i.e., cats may serve as a novel host for infection transmission from ticks) (74, 75). We also conducted phylogenetic analyses of T. foetus based on the small 5.8 s subunit ribosomal RNA. Typically, all the isolates cluster according to their respective infection hosts. Interestingly, the trichomonad isolates identified in our study clustered together with the human fetal T. foetus isolate (DQ243910), which was isolated from bronchoalveolar lavage samples of AIDS patients, highlighting the risk of cross-species transmission (Fig. 4B) (76).

Fig 4 Phylogenetic analyses of potentially zoonotic pathogens. (A) Maximum-likelihood phylogeny of rickettsia genera based on the groEL gene, with the green circle at the tip indicative of identification in this study. Trees were midpoint rooted, and bootstrap values > 70% from 1,000 bootstrap replicates are linked with the gradient size of the red circle at the node. The host distributions are shown with a cartoon pattern. (B) Maximum-likelihood phylogeny of Tritrichomonas genera based on 5.8S ribosomal RNA, with the green circle at the tip indicative of identification in this study. Trees were midpoint rooted, and bootstrap values > 70% from 1,000 bootstrap replicates are linked with the gradient size of the red circle at the node. The host distributions are shown with a cartoon pattern.

Two phylogenetic trees compare potential zoonotic pathogens like Rickettsia bacteria based on groEL genes and the Tritrichomonas genus based on 5.8S ribosomal RNA genes with host distributions.

Co-infection dynamic of different pathogens

To examine the differences in pathogen composition and co-infection patterns between hospitalized cats and stray cats, we compared the infectome profiles obtained from these two groups. In hospitalized cats, the majority of cases were associated with multiple pathogens (19/26), whereas fewer cases with multiple pathogens were observed in stray cats (Fig. S6). This suggests that the infectomes of diseased cats are highly complex and cannot be fully understood using a single pathogen disease model. Moreover, we found that the abundance diversity of pathogenic microbiota was significantly higher in hospitalized cats than in stray cats (Fig. 5A through C). Principal coordinates analysis (PCoA) revealed a distinction between the infectome communities of diseased cats and stray cats (permutational multivariate analysis of variance PERMANOVA], P adjusted = 0.704, R2 = 0.1906, Fig. 5D). However, when considering the sample type, our results did not show a significant difference in pathogenetic communities between blood and ascites samples (Fig. 5E and G). Additionally, there was no significant variation in total infectome richness, and the infectome communities did not cluster based on sampling types (Fig. 5H). We further examined the co-infection frequency of paired pathogens. The co-occurrence networks of diseased cats indicated that most coinfection cases involved a combination of viruses and bacteria, such as FIPV and M. aubagnense (Pearson r = 0.72) or FIPV and S. pneumoniae (Pearson r = 0.02) or FCV and M. aubagnense (Pearson r = 0.12). Furthermore, most cases presented a bacteria-associated co-infection spectrum such as M. aubagnense and S. aureus (coefficient = 0.69) or S. aureus and S. pneumoniae (coefficient = 0.69, Fig. 5I and J).

Fig 5 Comparisons of infectome diversity and composition metrics in our cohort. (A–D) Microbiota diversity characterization against different health conditions. (A) Richness index of pathogenetic microbiota of different health conditions. B. Simpson index of pathogenetic microbiota of different health conditions. (C) Shannon index of pathogenetic microbiota of different health conditions. (D) Beta diversity analysis of different health conditions. Principal co-ordinates analysis based on Bray–Curtis dissimilarities at the species level. Differences between the different health conditions were calculated based on the PERMANOVA test. Ellipses are at the 95% CI. (E–H) Microbiota diversity characterization against different sample types. (E) Richness index of pathogenetic microbiota of different sample types. (F) Simpson index of pathogenetic microbiota of different sample types. (G) Shannon index of pathogenetic microbiota of different sample types. (H) Beta diversity analysis of different sample types. Principal co-ordinate analysis based on Bray–Curtis dissimilarities at the species level. Differences between the different sample types were calculated based on the PERMANOVA test. Ellipses are at the 95% CI level. (G) Pearson’s correlation analysis of pairwise concurrent pathogens at abundance levels, with P < 0.05 shown as shading. (I) Co-infection network of important pathogens. Specifically, thickness and gradient colors (from white to red) of links of paired pathogens denoted the frequency of pairwise coinfection of two pathogens. The gradient color of nodes was proportional to the degree of prevalence rate in our study cohort. The size of each node represented the co-infection frequency with other pathogens (from white to red). Pairwise statistical tests of α-diversity indices were examined using a Wilcoxon test.

Graphs compare diversity: healthy versus unhealthy, ascites versus blood, and communities. Diagrams depict co-infecting pathogens and statistical test summaries.

Association between clinical manifestations and pathogen co-infection

The complex nature of co-infections presents challenges in establishing definitive causal relationships between pathogens and diseases. To address this, we conducted a comprehensive analysis that considered clinical symptom results in the context of a panel of relevant pathogens with indication of abundance, rather than focusing solely on individual pathogens. In cases of ascites hyperplasia, most co-infections were associated with a high abundance of both M. aubagnense and FIPV. For cases presenting fever and respiratory symptoms, there was a convergence of co-presenting pathogens to some extent, including FIPV, S. aureus, Rickettsia, A. baumannii, and S. pneumoniae, with a lower abundance of FIPV observed in respiratory models. Conversely, digestive symptoms or diarrheal models often involved combinations of three or more pathogens, many of which were present in high abundance (Fig. 6A through E). In addition, the diversity difference of individual symptoms was calculated and compared using the Shannon metric, which indicated that the diversity of pathogens within the host showing ascite symptoms was significantly lower than that showing other symptoms (Fig. 6F). The proportion of variance in microbiome composition that can be explained by individual symptoms was calculated by permutational multivariate analysis of variance using distance matrices (adonis). These symptoms explained 15.3% of infectome taxonomic composition, with the largest weight from ascite symptoms, followed by diarrhea, cough, and dermatology (Fig. 6G).

Fig 6 Association of pathogen infection and clinical manifestations in cats. (A–E) For each of the disease types, a heat map displays the prevalence and abundance of the pathogens identified. Disease types include general symptoms (i.e., ascites hyperplasia, fever, severe cough with expiratory dyspnea, diarrhea, and dermatology diseases). (F) Shannon index of pathogenetic microbiota of different clinical manifestations. (G) Bar plot illustrating each manifestation associated with the variation of the total infectome. Each manifestation was ranked by their collective Adonis R-squared value. Pairwise statistical tests of α-diversity indices were examined using the Wilcoxon test.

Heatmaps link pathogens to cat illnesses, featuring prevalence and abundance. Graphs compare pathogen diversity across illnesses and rank illnesses by the association between pathogen variety and illness.

Ecological drivers of infectome community composition

We constructed Gaussian-distributed generalized linear models (GLMs) with proxy ecological and demographic variables as predictors for diversity metrics. Submodels of GLMs were created to exclude highly correlated variables (i.e., >0.7; Fig. S1). Since multipathogen co-infection was common across the entire study cohort, we used the Shannon index as the response variable. We tested age, COXI reads, temperature, precipitation, humidity, sex, living conditions (stray or domesticated), latitude, and raw reads as proxy predictors. Variable selection on submodels revealed that COXI reads, precipitation, humidity, living conditions (stray/hospitalization), latitude, and raw reads had no significant effect on the diversity of pathogen immunity. In comparison, age, sex, and temperature showed significant effects on the diversity and abundance of the pathogen community (Fig. 7A through C), with age presenting the strongest effect (z-value = 10.58, P-value = 3.75E−09, and coefficient = 0.41). This suggests that host age may significantly shape the structure of the infectome, more so than other predictors (Fig. 7; Table S4). Interestingly, our linear regression model also rejected the null hypothesis that the population density of stray cats has no effect on pathogen infectome diversity, indicating that as the population density of stray cats increases, the diversity of the total infectome may also increase (Fig. 7D).

Fig 7 Ecological and demographic correlates of infectome diversity in our cohort. Model‐averaged relationships of demographic and ecological factors with diversity (Shannon and richness) and univariate correlations of significant factors. (A) In the model‐averaged results, effect sizes are shown for each factor across a set of GLMs with 95% CI. Factors that remained significant in the final model are shown in green. The vertical dashed line shows an effect size of 0, such that any CI overlapping the dashed line indicates a nonsignificant effect of the factor in model‐averaged results. (B and C) Shannon values are plotted for each variable (age and temperature) that was significant according to model averaging. Solid lines show GLM predictions for univariate relationships that remained significant following correction for multiple testing, with transparent shading indicating 95% CI. Points are colored according to sampling types (blue, ascites; green, blood). (D) Richness values are fit into population density of stray cats with the linear model. Solid lines show linear model predictions, with transparent shading indicating 95% CI. Points are colored according to sampling types (blue, ascites; green, blood). (E and F) Box-line plot compared (E) superinfection numbers and (F) average abundance in three groups (i.e., adult, juvenile, and kitten) as well as the P-values of the Wilcoxon test on the levels of each group. (G–J) Locally weighted regression (LOESS) modeling the predictions for univariate association between pathogen abundance and age using FIPV, M. aubagnense, S. aureus, and S. pneumoniae, with the outlier marked with an asterisk (J). Pairwise statistical tests of α-diversity indices were examined using the Wilcoxon test.

Graphs analyze how age and temperature affect cat pathogen diversity. Other panels feature diversity versus age and temperature, co-infections and abundance by age group, and pathogen abundance by age for specific bacteria and viruses.

Given the strong impact of age, we conducted additional statistical analyses to evaluate the effect of host age structure on the infectome community. First, we categorized cats into three groups based on age: kittens (0–6 months), young (6–12 months), and adults (13–72 months) and compared in vivo pathogen diversity and abundance). Higher numbers of superinfection cases were observed among young individuals, followed by adults and kittens (Fig. 7E and F). Similarly, infection abundance was highest in kittens, followed by adults and juveniles. We also examined whether specific prokaryotic or viral pathogens (e.g., FIPV, M. aubagnense, S. aureus, and S. pneumoniae) differed in their abundance or prevalence according to host age using locally weighted smoothing fitting (Fig. 7G through J). We observed that the abundance of FIPV and M. aubagnense increased significantly during the 10–15-month period (young stage), followed by a slower increase in subsequent infection periods (adult stage). The association between pathogen abundance and age was not clearly observed for S. aureus and S. pneumoniae due to the limited number of cases available (Fig. 7H through J). These findings suggest that young cats are more likely to harbor a greater number of pathogenic taxa with higher abundance than adults and kittens (Fig. 8). In summary, these results indicate that host and ecological variation can predetermine the likelihood of infectious disease-associated infections.

Fig 8 Geographical diagram to understand the ecological and demographic contribution to the clinical infectome. Two hypotheses are proposed that may explain the focal distribution of hospitalized and stray cats.

An illustration features the factors influencing the distribution of hospitalized and stray cats in two scenarios. Scenario 1 has co-infection is neglected but common. Scenario 2 features infectome diversity for age, gender, and high temperature.

DISCUSSION

Our knowledge of the infectome of companion cats is still limited, especially when it comes to understanding the ecological factors that influence pathogen infections in these animals. The diversity and factors contributing to variation in infectome communities among populations of companion cats have been largely unexplored, despite their importance in understanding the impact of environmental changes or human activities on infectious diseases.

This study is the first epidemiological cohort research based on infectome insight. Specifically, we first determined the infectome of cats (both pet and stray) in hospitals across various geographical locations in south China. This understanding will contribute to the current epidemiological baseline of domestic cat infectome, where a broader sampling schema better characterizes the infectome of cats in the south China region. Second, we examined the potential drivers of infectomic diversity among samples across different cities. In this scenario, if sampling is performed in a singular geographic location, the inherent environmental factor variances are minimal, making it challenging to fully characterize the influences on potential infectomic diversity. Lastly, we acknowledge that the epidemiology of pet cat pathogens in a single area exhibits certain similarities and cannot adequately reflect the broader epidemiology status of domestic cats in the south China region. Based on these scenarios, we sampled diseased cats from south China rather than one single fixed-point sampling.

The deployment of meta-transcriptomic sequencing enabled us to characterize the pathogen spectrum of domestic cat populations in an unbiased manner. In total, we have acquired 15 pathogen species, comprising three RNA viruses, 13 bacteria, and one parasite. Although transcriptomic analysis of total pathogens alone is unlikely to completely reveal the clinical consequences and fine mechanisms for each pairwise occurrence of pathogens, we carefully examined the potential manifestation of pathogen interaction with different symptom models. In our study, a substantial number of the cases (64.71%, 22/34) were characterized by the presence of at least two pathogens in a single clinical case. Hence, these findings did not easily fit the paradigm of “one disease, one pathogen.” Virus–bacteria co-infection was the most frequently observed in our cohort, specific to different manifestations. In virus–bacteria co-infection cases, the highest pairwise occurrence is the co-infection of FIPV and M. aubagnense. Notably, M. aubagnense was frequently documented as the potential causative etiology of respiratory secretions and joint fluid, with a rare focus on ascites hyperplasia. However, Du et al. (69) first postulated it as the causative etiology of severe peritoneal effusion in a patient in Guangxi (69). Hence, we further examined this scenario with a single-pathogen model, which showed that in cases of ascites, 70.59% (12/17) of cats were found to be infected with M. aubagnense, corroborating that the hypothesis of M. aubagnense as the potential etiology of ascite hyperplasia is not limited to humans but occurs even among cats. Generally, an impaired host immune system, damaged epithelial barrier, and excessive inflammatory response can result in complex infection scenarios and severe clinical outcomes (23). At the very least, the strong correlation between FIPV and M. aubagnense may act synergistically, resulting in changes in pathogen activity, exacerbating the clinical manifestations, or lead to longer disease duration than that of individual infections alone.

Of particular interest is the identification of Candidatus R. tarasevichiae infection in cats. Candidatus R. tarasevichiae has been previously recognized as a tick-borne pathogen primarily associated with ticks and rodents as vector hosts (77, 78). However, there have been reports of meningitis-like manifestations in fatal cases of rickettsiosis caused by Candidatus R. tarasevichiae in China (75, 79). Additionally, a fatal infection in a 4-year-old girl with typical symptoms of tick-borne rickettsiosis (fever, rash, eschar at the site of the tick bite, and myalgia) and meningeal involvement was documented, revealing a co-infection with mixed R. sibirica and Candidatus R. tarasevichiae in Russia (80). While these studies have expanded our understanding of the host range of Candidatus R. tarasevichiae, no previous study has reported cats as the potential host for Candidatus R. tarasevichiae. In our study, a cat (no. L01_19) infected with Candidatus R. tarasevichiae exhibited severe fever, cough with expiratory dyspnea, and diarrhea. Although we failed to learn the potential source of infection, this finding also expanded the potential host spectrum of Candidatus R. tarasevichiae. Another intriguing observation is the characterization of T. foetus, which showed a closer phylogenetic relationship to the human sequence. This finding suggests a mysterious connection between the two strains. Given that companion cats frequently interact with their owners, there is a particular concern regarding that cats may represent an additional reservoir host that could impact human infection rates via dilution or amplification effects (81). Therefore, it is crucial to conduct thorough and periodic investigations of infectious pathogens in companion animals and local tick populations to mitigate potential zoonotic threats.

Although our study revealed significant variation in pathogen communities among populations of the same species, our GLM analysis rejected the null hypothesis that the similarity of pathogenic microbiota communities is solely driven by spatial proximity or symptom-clustered disease models. Instead, our modeling analysis supported the hypothesis that gender, temperature, and age, rather than geographical variation or sequencing capacity, influence the diversity of the pathogen spectrum. Specifically, we observed a significant correlation between age and pathogen diversity (r = 0.41). This suggested that older cats are more likely to harbor a greater diversity or higher abundance of pathogens. However, when we grouped cases based on three major age periods (kitten, adult, and juvenile) to examine the effect of age, we found that juveniles, rather than adults, retained the highest pathogen community diversity regardless of taxa or abundance. Further analysis focusing on individual pathogens confirmed our findings that cats aged between 10 and 20 months were more likely to harbor pathogens such as FIPV and M. aubagnense. After 20 months, the pathogen burden seemed to decline drastically (Fig. 7G through J). We postulated that this age-based difference in pathogen shedding may be attributed to age-related variations in innate immunity or possibly adaptive immunity specific to these taxa. Previous studies have shown that certain populations of swans maintain long-lasting immune responses to avian influenza virus, making adult swans less susceptible to influenza-associated mortality (82). Indeed, this is not the only case revealing the effect of age against pathogen infection. Hill et al. (22) also showed that perturbations that affect population age structures of wildlife could alter pathogen transmission dynamics (83). Similarly, a study suggested that colonies with a higher proportion of juveniles consistently had more diverse fecal viral communities (24). To the best of our knowledge, our study is the first to demonstrate that age heterogeneity may influence infectome diversity. Therefore, this host-level effect may be driven by age-related demographic factors such as mobility, diet, or maternal stress or differential exposure rates to pathogens among different age classes. For example, independent young cats may have increased exposure to many pathogens due to their higher mobility. However, their immune levels may not undergo corresponding enhancement, thereby increasing the likelihood of pathogenic infections. Another host factor that significantly influences pathogen gain is gender. A number of studies have suggested the association between gender and parasites such as cestode and nematode, with only a few rare studies explicitly focusing on total infectious diseases (52–54, 84). Conversely, female cats were observed to have more diverse pathogen infections. We hypothesized this might be linked to the distinct behavior and physiology of female cats. For example, risks of maternal complications (infection, urinary retention, hematoma, or ruptured sutures) in the postpartum period increase with duration of the second stage of labor also after accounting for maternal, pregnancy, and delivery characteristics (85). However, this scenario should be interpreted cautiously, as these observations are totally human-based, with no direct clinical-based evidence reflecting on cats explicitly. In addition to host variables, we also examined ecological factors that contribute to predicting pathogen diversity. Our findings suggest that warmer environmental conditions increase pathogen transmission (and/or environmental persistence) and lead to a higher number of pathogen infections. In humans, temperature has been strongly associated with bacterial infections, directly transmitted viruses, and helminth infections on a global scale (55).

Our study has several limitations. (i) While sequencing to high depth allowed us to recover the total microbiome and detect how ecological and host conditions shaped pathogen prevalence and abundance, our characterization was nevertheless dominated by bacteria rather than viruses. As mentioned before, this might be the nature of differences in ethology, living conditions, or human intervention in divergent animal models. (ii) Although we did test the ecological function and evolutionary details of variation in bacteria, pathogenic viruses, and parasite in this study, it is possible that a more generalized microbiota deputation such as total virome (including prokaryotic and eukaryotic) may equally hold ecological importance. Although this is not the major focus of this study, follow‐up analyses which include environmental viruses and bacteriophages not limited to those linked with health or disease in mammals might be better in deciphering the impact of ecological variation regarding the virome rather than pathogenic virus, examining associations between variation in viral diversity and host fitness or testing interactions between viral communities and other components of the host microbiome. (iii) The abundance and community composition of pathogens detected in domestic cats were slightly influenced by specific host behaviors such as vaccination. Future research could focus on selecting disease models based on clinical information, including vaccination status. (iv) Our stringent statistical tests and correction for multiple comparisons may have resulted in low(er) power and potential false-negative results. It is possible that other microbial taxa are differentially abundant or prevalent with age, in addition to those reported here. (v) Comparison of the healthy and diseased group would provide more specific infectomic dynamics such as contribution of each pathogen to specific symptoms. But due to the sparseness of ascitic fluid in healthy cats, sampling would be hardly feasible. Note that we failed to include ascitic fluid in our study, the healthy group from blood would lack reference and comparison. Future research sampling of swabs from the respiratory and digestive tract of diseased and healthy cats would be informative to understand pathogen contribution to a specific disease.

Despite these limitations, our study showed the capacity of meta-transcriptomics to reveal host in vivo transcription to discover the total pathogen array in a single assay. This approach allowed us to accurately assess pathogen abundance in diseased models and explore their genomic evolutionary details, enabling the identification of potential pathogens and the evaluation of their threat to human health. Additionally, we presented a comprehensive framework that combines meta-transcriptomics with experimental approaches to investigate the effects of environmental and host variables on the spread of infectious diseases. Specifically, we investigated the total infectome associated with diseased cats in China, revealing a diverse array of viruses, bacteria, and eukaryotic pathogens. While most of these pathogens are well-known cat pathogens, we identified a divergent species within the genus Rickettsia associated with self-limited disease in cats, as well as potential novel clinical outcomes associated with M. aubagnense in cats. In addition, our findings revealed that most of the cat diseases examined were better explained by the presence of co-infection with multiple pathogens rather than infection with a single pathogen. We concluded that altered demographic structure and ecological variables may have a cascading effect on the diverse dynamic of the infectome. More importantly, we showed the value of linking the modern influx of meta-transcriptomics with comparative ecology and demography and utilizing it to affirm that the ecological and demographic variation impacts not only the single-pathogen model but also the total infectome.

ACKNOWLEDGMENTS

This study is supported by the Key-Area Research and Development Program of Guangdong Province [grant number 2019B020211003], the National Natural Science Foundation of China [grant number 32102704], and China Agriculture Research System of MOF and MARA. G.B. acknowledges support from the Research Foundation-Flanders (“Fonds voor Wetenschappelijk Onderzoek-Vlaanderen,” G0E1420N, G098321N), from the Internal Funds KU Leuven (Grant No. C14/18/094), from the European Union Horizon 2023 RIA project LEAPS (grant agreement no. 101094685), and from the DURABLE EU4Health project 02/2023-01/2027, which is co-funded by the European Union (call EU4H-2021-PJ4) under Grant Agreement No. 101102733.

J.X. designed the study. Y.S., S.C., J.D., H.G., and T.Z. collected the samples as well as the metadata. Y.L. and S.X. prepared the libraries and performed the sequencing. J.X. performed all data analyses and the visualization. J.X. wrote the manuscript. G.B. revised the manuscript. G.L., G.Z., Y.S., and G.B. provided supervisory support. All authors contributed to the review and editing of the manuscript.

DATA AVAILABILITY

All shotgun sequencing files have been deposited in the China National GeneBank DataBase (CNGBdb) under the accession number CNP0004735. All sequencing files were deposited in CNGBdb under the sequencing accession numbers N_AAAECY010000000 to N_AAAEDH010000000 (https://db.cngb.org/). Any remaining data can be acquired in the supplemental files or by request to the corresponding authors.

ETHICS APPROVAL

Our sampling procedures were approved by the Animal Ethics Committee of South China Agricultural University and conducted under the guidance of the South China Agricultural University Institutional Animal Care and Use Committee (SCAU-AEC-2022A010).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/msystems.00636-24.

10.1128/msystems.00636-24.SuF1 Supplemental material msystems.00636-24-s0001.docx

Supplemental methods; Fig. S1 to S7.

10.1128/msystems.00636-24.SuF2 Supplemental tables msystems.00636-24-s0002.xlsx

Tables S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
==== Refs
REFERENCES

1 Wells DL. 2019. The state of research on human–animal relations: implications for human health. Anthrozoös 32 :169–181. doi:10.1080/08927936.2019.1569902
2 Beetz A, Uvnäs-Moberg K, Julius H, Kotrschal K. 2012. Psychosocial and psychophysiological effects of human-animal interactions: the possible role of oxytocin. Front Psychol 3 :234. doi:10.3389/fpsyg.2012.00234 22866043
3 Morales-Jinez A, López-Rincón FJ, Ugarte-Esquivel A, Andrade-Valles I, Rodríguez-Mejía LE, Hernández-Torres JL. 2018. Allostatic load and canine companionship: a comparative study using biomarkers in older adults. Rev Lat Am Enfermagem 26 :e3071. doi:10.1590/1518-8345.2755.3071 30462785
4 Wells DL. 2007. Domestic dogs and human health: an overview. Br J Health Psychol 12 :145–156. doi:10.1348/135910706X103284 17288671
5 HaagAF, Fitzgerald JR, Penadés JR. 2019. Staphylococcus aureus in animals. Microbiol Spectr 7 . doi:10.1128/microbiolspec.GPP3-0060-2019
6 Pocar P, Grieco V, Aidos L, Borromeo V. 2023. Endocrine-disrupting chemicals and their effects in pet dogs and cats: an overview. Animals (Basel) 13 :378. doi:10.3390/ani13030378 36766267
7 Ahmed F, Cappai MG, Morrone S, Cavallo L, Berlinguer F, Dessì G, Tamponi C, Scala A, Varcasia A. 2021. Raw meat based diet (RMBD) for household pets as potential door opener to parasitic load of domestic and urban environment. Revival of understated zoonotic hazards? A review. One Health 13 :100327. doi:10.1016/j.onehlt.2021.100327 34584928
8 Seyedmousavi S, Guillot J, Arné P, de Hoog GS, Mouton JW, Melchers WJG, Verweij PE. 2015. Aspergillus and aspergilloses in wild and domestic animals: a global health concern with parallels to human disease. Med Myco 53 :765–797. doi:10.1093/mmy/myv067
9 Di Martino B, Di Profio F, Melegari I, Marsilio F. 2019. Feline virome-a review of novel enteric viruses detected in cats. Viruses 11 :908. doi:10.3390/v11100908 31575055
10 Van Brussel K, Wang X, Shi M, Carrai M, Feng S, Li J, Holmes EC, Beatty JA, Barrs VR. 2022. The enteric virome of cats with feline panleukopenia differs in abundance and diversity from healthy cats. Transbound Emerg Dis 69 :e2952–e2966. doi:10.1111/tbed.14646 35765950
11 Li Y, Gordon E, Idle A, Altan E, Seguin MA, Estrada M, Deng X, Delwart E. 2020. Virome of a feline outbreak of diarrhea and vomiting includes bocaviruses and a novel chapparvovirus. Viruses 12 :506. doi:10.3390/v12050506 32375386
12 Chu S, Wylie TN, Wylie KM, Johnson GC, Skidmore ZL, Fleer M, Griffith OL, Bryan JN. 2020. A virome sequencing approach to feline oral squamous cell carcinoma to evaluate viral causative factors. Vet Microbiol 240 :108491. doi:10.1016/j.vetmic.2019.108491 31902496
13 Ng TFF, Mesquita JR, Nascimento MSJ, Kondov NO, Wong W, Reuter G, Knowles NJ, Vega E, Esona MD, Deng X, Vinjé J, Delwart E. 2014. Feline fecal virome reveals novel and prevalent enteric viruses. Vet Microbiol 171 :102–111. doi:10.1016/j.vetmic.2014.04.005 24793097
14 Di Profio F, Sarchese V, Palombieri A, Fruci P, Massirio I, Martella V, Fulvio M, Di Martino B. 2022. Feline chaphamaparvovirus in cats with enteritis and upper respiratory tract disease. Transbound Emerg Dis 69 :660–668. doi:10.1111/tbed.14032 33559350
15 Zhang W, Li L, Deng X, Kapusinszky B, Pesavento PA, Delwart E. 2014. Faecal virome of cats in an animal shelter. J Gen Virol 95 :2553–2564. doi:10.1099/vir.0.069674-0 25078300
16 Overgaauw PAM, Vinke CM, Hagen M van, Lipman LJA. 2020. A one health perspective on the human-companion animal relationship with emphasis on zoonotic aspects. Int J Environ Res Public Health 17 :3789. doi:10.3390/ijerph17113789 32471058
17 Patterson EI, Elia G, Grassi A, Giordano A, Desario C, Medardo M, Smith SL, Anderson ER, Prince T, Patterson GT, Lorusso E, Lucente MS, Lanave G, Lauzi S, Bonfanti U, Stranieri A, Martella V, Solari Basano F, Barrs VR, Radford AD, Agrimi U, Hughes GL, Paltrinieri S, Decaro N. 2020. Evidence of exposure to SARS-CoV-2 in cats and dogs from households in Italy. Nat Commun 11 :6231. doi:10.1038/s41467-020-20097-0 33277505
18 Sila T, Sunghan J, Laochareonsuk W, Surasombatpattana S, Kongkamol C, Ingviya T, Siripaitoon P, Kositpantawong N, Kanchanasuwan S, Hortiwakul T, Charernmak B, Nwabor OF, Silpapojakul K, Chusri S. 2022. Suspected cat-to-human transmission of SARS-CoV-2, Thailand, July-September 2021. Emerg Infect Dis 28 :1485–1488. doi:10.3201/eid2807.212605 35666777
19 Piewbang C, Poonsin P, Lohavicharn P, Wardhani SW, Dankaona W, Puenpa J, Poovorawan Y, Techangamsuwan S. 2022. SARS-Cov-2 transmission from human to pet and suspected transmission from pet to human, Thailand. J Clin Microbiol 60 :e0105822. doi:10.1128/jcm.01058-22 36314788
20 Liu C, Liu Y, Qian P, Cao Y, Wang J, Sun C, Huang B, Cui N, Huo N, Wu H, Wang L, Xi X, Tian K. 2020. Molecular and serological investigation of cat viral infectious diseases in China from 2016 to 2019. Transbound Emerg Dis 67 :2329–2335. doi:10.1111/tbed.13667 32511839
21 Kim S, Cheng Y, Fang Z, Liu X, Zhongqi Q, Weidong Y, Yilmaz A, Yilmaz H, Umar S. 2024. Molecular epidemiology and phylogenetic analysis of feline calicivirus in Kunshan, China. Virol J 21 :50. doi:10.1186/s12985-024-02319-9 38414028
22 Li SL, Acosta AL, Hill SC, Brady OJ, de Almeida MAB, Cardoso J da C, Hamlet A, Mucci LF, Telles de Deus J, Iani FCM, Alexander NS, Wint GRW, Pybus OG, Kraemer MUG, Faria NR, Messina JP. 2022. Mapping environmental suitability of Haemagogus and Sabethes spp. mosquitoes to understand sylvatic transmission risk of yellow fever virus in Brazil. PLoS Negl Trop Dis 16 :e0010019. doi:10.1371/journal.pntd.0010019 34995277
23 Susi H, Barrès B, Vale PF, Laine A-L. 2015. Co-infection alters population dynamics of infectious disease. Nat Commun 6 :5975. doi:10.1038/ncomms6975 25569306
24 Bergner LM, Orton RJ, Benavides JA, Becker DJ, Tello C, Biek R, Streicker DG. 2020. Demographic and environmental drivers of metagenomic viral diversity in vampire bats. Mol Ecol 29 :26–39. doi:10.1111/mec.15250 31561274
25 Fearon ML, Tibbetts EA. 2021. Pollinator community species richness dilutes prevalence of multiple viruses within multiple host species. Ecology 102 :e03305. doi:10.1002/ecy.3305 33571384
26 Mbora DNM, McPeek MA. 2009. Host density and human activities mediate increased parasite prevalence and richness in primates threatened by habitat loss and fragmentation. J Anim Ecol 78 :210–218. doi:10.1111/j.1365-2656.2008.01481.x 19120603
27 Lo CM, Morand S, Galzin R. 1998. Parasite diversity/host age and size relationship in three coral-reef fishes from French Polynesia. Int J Parasitol 28 :1695–1708. doi:10.1016/s0020-7519(98)00140-4 9846606
28 Griffiths ME, Broos A, Bergner LM, Meza DK, Suarez NM, da Silva Filipe A, Tello C, Becker DJ, Streicker DG. 2022. Longitudinal deep sequencing informs vector selection and future deployment strategies for transmissible vaccines. PLoS Biol 20 :e3001580. doi:10.1371/journal.pbio.3001580 35439242
29 Meza DK, Mollentze N, Broos A, Tello C, Valderrama W, Recuenco S, Carrera JE, Shiva C, Falcon N, Viana M, Streicker DG. 2022. Ecological determinants of rabies virus dynamics in vampire bats and spillover to livestock. Proc Biol Sci 289 :20220860. doi:10.1098/rspb.2022.0860 36069012
30 Jacquot M, Wallace MA, Streicker DG, Biek R. 2022. Geographic range overlap rather than phylogenetic distance explains rabies virus transmission among closely related bat species. Viruses 14 :2399. doi:10.3390/v14112399 36366496
31 Rojas-Sereno ZE, Streicker DG, Medina-Rodríguez AT, Benavides JA. 2022. Drivers of spatial expansions of vampire bat rabies in Colombia. Viruses 14 :2318. doi:10.3390/v14112318 36366416
32 Campbell SJ, Ashley W, Gil-Fernandez M, Newsome TM, Di Giallonardo F, Ortiz-Baez AS, Mahar JE, Towerton AL, Gillings M, Holmes EC, Carthey AJR, Geoghegan JL. 2020. Red fox viromes in urban and rural landscapes. Virus Evol 6 :veaa065. doi:10.1093/ve/veaa065 33365150
33 Myer MH, Johnston JM. 2019. Spatiotemporal Bayesian modeling of West Nile virus: identifying risk of infection in mosquitoes with local-scale predictors. Sci Total Environ 650 :2818–2829. doi:10.1016/j.scitotenv.2018.09.397 30373059
34 Susi H, Laine A-L. 2021. Agricultural land use disrupts biodiversity mediation of virus infections in wild plant populations. New Phytol 230 :2447–2458. doi:10.1111/nph.17156 33341977
35 Hermanns K, Marklewitz M, Zirkel F, Kopp A, Kramer-Schadt S, Junglen S. 2023. Mosquito community composition shapes virus prevalence patterns along anthropogenic disturbance gradients. Elife 12 :e66550. doi:10.7554/eLife.66550 37702388
36 Firth C, Lipkin WI. 2013. The genomics of emerging pathogens. Annu Rev Genomics Hum Genet 14 :281–300. doi:10.1146/annurev-genom-091212-153446 24003855
37 Mokili JL, Rohwer F, Dutilh BE. 2012. Metagenomics and future perspectives in virus discovery. Curr Opin Virol 2 :63–77. doi:10.1016/j.coviro.2011.12.004 22440968
38 Harvey E, Holmes EC. 2022. Diversity and evolution of the animal virome. Nat Rev Microbiol 20 :321–334. doi:10.1038/s41579-021-00665-x 34983966
39 Li C-X, Shi M, Tian J-H, Lin X-D, Kang Y-J, Chen L-J, Qin X-C, Xu J, Holmes EC, Zhang Y-Z. 2015. Unprecedented genomic diversity of RNA viruses in arthropods reveals the ancestry of negative-sense RNA viruses. eLife 4 :e05378. doi:10.7554/eLife.05378 25633976
40 Huang X, Wu W, Tian X, Hou X, Cui X, Xiao Y, Jiao Q, Zhou P, Liu L, Shi W, Chen L, Sun Y, Yang Y, Chen J, Zhang G, Liu J, Holmes EC, Cai X, An T, Shi M. 2022. A total infectome approach to understand the etiology of infectious disease in pigs. Microbiome 10 :73. doi:10.1186/s40168-022-01265-4 35538563
41 Shi M, Lin X-D, Chen X, Tian J-H, Chen L-J, Li K, Wang W, Eden J-S, Shen J-J, Liu L, Holmes EC, Zhang Y-Z. 2018. The evolutionary history of vertebrate RNA viruses. Nature 556 :197–202. doi:10.1038/s41586-018-0012-7 29618816
42 Shi M, Lin X-D, Tian J-H, Chen L-J, Chen X, Li C-X, Qin X-C, Li J, Cao J-P, Eden J-S, Buchmann J, Wang W, Xu J, Holmes EC, Zhang Y-Z. 2016. Redefining the invertebrate RNA virosphere. Nature 540 :539–543. doi:10.1038/nature20167 27880757
43 Shi M, Zhao S, Yu B, Wu W-C, Hu Y, Tian J-H, Yin W, Ni F, Hu H-L, Geng S, Tan L, Peng Y, Song Z-G, Wang W, Chen Y-M, Holmes EC, Zhang Y-Z. 2022. Total infectome characterization of respiratory infections in pre-COVID-19. PLoS Pathog 18 :e1010259. doi:10.1371/journal.ppat.1010259 35176118
44 Sun Y, Xing J, Li Y, Xu S, Wang Z, Lin J, Cao X, Zheng Z, Gong L, Shi M, Zhang G. 2023. Total infectome of etiology and epidemiology unravels a complex infection landscape in Chinese acute diarrheal pig herds. Res Square. doi:10.21203/rs.3.rs-2623397/v1
45 Nunn CL, Altizer S, Jones KE, Sechrest W. 2003. Comparative tests of parasite species richness in primates. Am Nat 162 :597–614. doi:10.1086/378721 14618538
46 Cross PC, Lloyd‐Smith JO, Johnson PLF, Getz WM. 2005. Duelling timescales of host movement and disease recovery determine invasion of disease in structured populations. Ecol Lett 8 :587–595. doi:10.1111/j.1461-0248.2005.00760.x
47 Benavides JA, Huchard E, Pettorelli N, King AJ, Brown ME, Archer CE, Appleton CC, Raymond M, Cowlishaw G. 2012. From parasite encounter to infection: multiple-scale drivers of parasite richness in a wild social primate population. Am J Phys Anthropol 147 :52–63. doi:10.1002/ajpa.21627 21989507
48 Poirotte C, Basset D, Willaume E, Makaba F, Kappeler PM, Charpentier MJE. 2016. Environmental and individual determinants of parasite richness across seasons in a free-ranging population of Mandrills (Mandrillus sphinx). Am J Phys Anthropol 159 :442–456. doi:10.1002/ajpa.22888 26515669
49 Anthony SJ, Islam A, Johnson C, Navarrete-Macias I, Liang E, Jain K, Hitchens PL, Che X, Soloyvov A, Hicks AL, Ojeda-Flores R, Zambrana-Torrelio C, Ulrich W, Rostal MK, Petrosov A, Garcia J, Haider N, Wolfe N, Goldstein T, Morse SS, Rahman M, Epstein JH, Mazet JK, Daszak P, Lipkin WI. 2015. Non-random patterns in viral diversity. Nat Commun 6 :8147. doi:10.1038/ncomms9147 26391192
50 Chow C-E, Suttle CA. 2015. Biogeography of viruses in the sea. Annu Rev Virol 2 :41–66. doi:10.1146/annurev-virology-031413-085540 26958906
51 Bergner LM, Orton RJ, da Silva Filipe A, Shaw AE, Becker DJ, Tello C, Biek R, Streicker DG. 2019. Using noninvasive metagenomics to characterize viral communities from wildlife. Mol Ecol Resour 19 :128–143. doi:10.1111/1755-0998.12946 30240114
52 Poulin R. 1996. Helminth growth in vertebrate hosts: does host sex matter? Int J Parasitol 26 :1311–1315. doi:10.1016/s0020-7519(96)00108-7 9024877
53 Zuk M, McKean KA. 1996. Sex differences in parasite infections: patterns and processes. Int J Parasitol 26 :1009–1023. doi:10.1016/S0020-7519(96)80001-4 8982783
54 Negro SS, Caudron AK, Dubois M, Delahaut P, Gemmell NJ. 2010. Correlation between male social status, testosterone levels, and parasitism in a dimorphic polygynous mammal. PLoS One 5 :e12507. doi:10.1371/journal.pone.0012507 20856933
55 Guernier V, Hochberg ME, Guégan J-F. 2004. Ecology drives the worldwide distribution of human diseases. PLoS Biol 2 :e141. doi:10.1371/journal.pbio.0020141 15208708
56 Dunn RR, Davies TJ, Harris NC, Gavin MC. 2010. Global drivers of human pathogen richness and prevalence. Proc Biol Sci 277 :2587–2595. doi:10.1098/rspb.2010.0340 20392728
57 Raghwani J, Faust CL, François S, Nguyen D, Marsh K, Raulo A, Hill SC, Parag KV, Simmonds P, Knowles SCL, Pybus OG. 2023. Seasonal dynamics of the wild rodent faecal virome. Mol Ecol 32 :4763–4776. doi:10.1111/mec.16778 36367339
58 Sharma V, Sharma R, Aulakh RS, Singh BB. 2023. Prevalence of Brucella species in stray cattle, dogs and cats: a systematic review. Prev Vet Med 219 :106017. doi:10.1016/j.prevetmed.2023.106017 37669604
59 Buendía-González FO, Legorreta-Herrera M. 2022. The similarities and differences between the effects of testosterone and DHEA on the innate and adaptive immune response. Biomolecules 12 :1768. doi:10.3390/biom12121768 36551196
60 Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30 :2114–2120. doi:10.1093/bioinformatics/btu170 24695404
61 Li D, Liu C-M, Luo R, Sadakane K, Lam T-W. 2015. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics 31 :1674–1676. doi:10.1093/bioinformatics/btv033 25609793
62 Okonechnikov K, Golosova O, Fursov M, the UGENE team. 2012. Unipro UGENE: a unified bioinformatics toolkit. Bioinformatics 28 :1166–1167. doi:10.1093/bioinformatics/bts091 22368248
63 Truong DT, Franzosa EA, Tickle TL, Scholz M, Weingart G, Pasolli E, Tett A, Huttenhower C, Segata N. 2015. MetaPhlAn2 for enhanced metagenomic taxonomic profiling. Nat Methods 12 :902–903. doi:10.1038/nmeth.3589 26418763
64 Li W, Godzik A. 2006. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22 :1658–1659. doi:10.1093/bioinformatics/btl158 16731699
65 Capella-Gutiérrez S, Silla-Martínez JM, Gabaldón T. 2009. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25 :1972–1973. doi:10.1093/bioinformatics/btp348 19505945
66 Nguyen L-T, Schmidt HA, von Haeseler A, Minh BQ. 2015. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol 32 :268–274. doi:10.1093/molbev/msu300 25371430
67 Yu G, Lam T-Y, Zhu H, Guan Y. 2018. Two methods for mapping and visualizing associated data on phylogeny using ggtree. Mol Biol Evol 35 :3041–3043. doi:10.1093/molbev/msy194 30351396
68 Oksanen J, Kindt R, Legendre P, O’Hara B, Stevens H, Oksanen M. 2010. vegan: community ecology package. The vegan package
69 Du Y, Huang J, Nong Y, Zhong R, Zhu R, Song W, Huang H, Tan Q, Xu J, Xiao X, Jia J. 2023. A case of ascetic fluid Mycobacterium aubagnense infection in a patient with severe peritoneal effusion. BMC Infect Dis 23 :84. doi:10.1186/s12879-023-08041-1 36750816
70 Shi W, Shi M, Que T-C, Cui X-M, Ye R-Z, Xia L-Y, Hou X, Zheng J-J, Jia N, Xie X, et al. . 2022. Trafficked Malayan pangolins contain viral pathogens of humans. Nat Microbiol 7 :1259–1269. doi:10.1038/s41564-022-01181-1 35918420
71 Tekes G, Thiel HJ. 2016. Feline coronaviruses: pathogenesis of feline infectious peritonitis. Adv Virus Res 96 :193–218. doi:10.1016/bs.aivir.2016.08.002 27712624
72 Chang H-W, Egberink HF, Halpin R, Spiro DJ, Rottier PJM. 2012. Spike protein fusion peptide and feline coronavirus virulence. Emerg Infect Dis 18 :1089–1095. doi:10.3201/eid1807.120143 22709821
73 Cano-Ortiz L, Tochetto C, Roehe PM, Franco AC, Junqueira DM. 2022. Could phylogenetic analysis be used for feline leukemia virus (FeLV) classification? Viruses 14 :249. doi:10.3390/v14020249 35215842
74 Hiraoka H, Shimada Y, Sakata Y, Watanabe M, Itamoto K, Okuda M, Inokuma H. 2005. Detection of rickettsial DNA in ixodid ticks recovered from dogs and cats in Japan. J Vet Med Sci 67 :1217–1222. doi:10.1292/jvms.67.1217 16397379
75 Jia N, Jiang J-F, Huo Q-B, Jiang B-G, Cao W-C. 2013. Rickettsia sibirica subspecies sibirica BJ-90 as a cause of human disease. N Engl J Med 369 :1176–1178. doi:10.1056/NEJMc1303625 24047079
76 Duboucher C, Caby S, Dufernez F, Chabé M, Gantois N, Delgado-Viscogliosi P, Billy C, Barré E, Torabi E, Capron M, Pierce RJ, Dei-Cas E, Viscogliosi E. 2006. Molecular identification of Tritrichomonas foetus-like organisms as coinfecting agents of human Pneumocystis pneumonia. J Clin Microbiol 44 :1165–1168. doi:10.1128/JCM.44.3.1165-1168.2006 16517921
77 Yuan T-T, Ma L, Jiang B-G, Fu W-M, Sun Y, Jia N, Jiang J-F. 2020. First confirmed infection of Candidatus Rickettsia tarasevichiae in rodents collected from northeastern China. Vector Borne Zoonotic Dis 20 :88–92. doi:10.1089/vbz.2019.2443 31453762
78 Eremeeva ME, Oliveira A, Robinson JB, Ribakova N, Tokarevich NK, Dasch GA. 2006. Prevalence of bacterial agents in Ixodes persulcatus ticks from the Vologda Province of Russia. Ann N Y Acad Sci 1078 :291–298. doi:10.1196/annals.1374.054 17114724
79 Liu W, Li H, Lu Q-B, Cui N, Yang Z-D, Hu J-G, Fan Y-D, Guo C-T, Li X-K, Wang Y-W, Liu K, Zhang X-A, Yuan L, Zhao P-Y, Qin S-L, Cao W-C. 2016. Candidatus Rickettsia tarasevichiae Infection in eastern central China: a case series. Ann Intern Med 164 :641–648. doi:10.7326/M15-2572 27019406
80 Rudakov N, Samoylenko I, Shtrek S, Igolkina Y, Rar V, Zhirakovskaia E, Tkachev S, Kostrykina T, Blokhina I, Lentz P, Tikunova N. 2019. A fatal case of tick-borne rickettsiosis caused by mixed Rickettsia sibirica subsp. sibirica and "Candidatus Rickettsia tarasevichiae" infection in Russia. Ticks Tick Borne Dis 10 :101278. doi:10.1016/j.ttbdis.2019.101278 31477530
81 Luis AD, Kuenzi AJ, Mills JN. 2018. Species diversity concurrently dilutes and amplifies transmission in a zoonotic host-pathogen system through competing mechanisms. Proc Natl Acad Sci U S A 115 :7979–7984. doi:10.1073/pnas.1807106115 30012590
82 Hill SC, Hansen R, Watson S, Coward V, Russell C, Cooper J, Essen S, Everest H, Parag KV, Fiddaman S, Reid S, Lewis N, Brookes SM, Smith AL, Sheldon B, Perrins CM, Brown IH, Pybus OG. 2019. Comparative micro-epidemiology of pathogenic avian influenza virus outbreaks in a wild bird population. Philos Trans R Soc Lond B Biol Sci 374 :20180259. doi:10.1098/rstb.2018.0259 31056057
83 Hill Sarah C, François S, Thézé J, Smith AL, Simmonds P, Perrins CM, van der Hoek L, Pybus OG. 2023. Impact of host age on viral and bacterial communities in a waterbird population. ISME J 17 :215–226. doi:10.1038/s41396-022-01334-4 36319706
84 Ling F, Steinel N, Weber J, Ma L, Smith C, Correa D, Zhu B, Bolnick D, Wang G. 2020. The gut microbiota response to helminth infection depends on host sex and genotype. ISME J 14 :1141–1153. doi:10.1038/s41396-020-0589-3 32005978
85 Stephansson O, Sandström A, Petersson G, Wikström AK, Cnattingius S. 2016. Prolonged second stage of labour, maternal infectious disease, urinary retention and other complications in the early postpartum period. BJOG 123 :608–616. doi:10.1111/1471-0528.13287 25601143
