Geographic and cross-species transmission of Mamastroviruses from multiple hosts

Abstract

Background

Astroviruses (AstVs) are single-stranded RNA viruses belonging to the family Astroviridae, which comprises two genera: Mamastrovirus (MaAstV) and Avastrovirus. These viruses circulate globally and infect multiple hosts, enabling cross-species transmission that poses significant risks to both humans and animals. In this study, we comprehensively analyze the evolution, global geographic transmission, and cross-species transmission of MaAstVs.

Results

Our analysis reveals that a broad distribution and diverse host range for AstVs, with a noticeable decline in detection after 2017. Phylogenetic analyses uncover complex relationships, suggesting potential host spillover events. The evolutionary rates of different MaAstVs vary. Most MaAstVs exhibited a decreasing trend in effective population size, with minor fluctuations observed in a few instances. Furthermore, our global transmission analysis demonstrates the spread of MaAstVs across Asia, Europe, the Americas, Africa, and Oceania, identifying specific countries as key nodes within the transmission network. Analyses of cross-species transmission in Groups 1, 2, and 3 reveal distinct ancestral origins and trace the cross-species transmission history of MaAstVs among domestic animals, humans, companion animals, and wildlife.

Conclusions

This study offers a detailed examination of the evolution, global transmission, and cross-species transmission of MaAstVs across multiple hosts. Our findings provide valuable insights into their circulation across species and regions, as well as their evolutionary dynamics.

Supplementary Information

The online version contains supplementary material available at 10.1186/s42523-026-00517-0.

Background

The family Astroviridae comprises single-stranded RNA (+ssRNA) viruses with genome lengths of approximately 6–8 kb [1]. Astrovirus (AstV) genomes contain three open reading frames (ORFs): ORF1a/1b encode nonstructural proteins essential for viral replication and production of infectious virions, while ORF2 encodes the capsid protein [2]. Compared to ORF1a/1b, ORF2—the most variable and immunogenic region of the genome—is the primary target for phylogenetic and evolutionary studies, as it better reflects genetic diversity and host adaptation [1, 3].

AstVs exhibit a wide geographic distribution and cause diverse clinical manifestations across a broad range of hosts. Astroviridae consists of two genera: Mamastrovirus (MaAstV), infecting mammals, and Avastrovirus, infecting birds. In geographic distribution, MaAstVs have been detected on every continent, including North America, Europe, Asia, Africa, South America, and Oceania [2]. MaAstVs infect a wide range of hosts, causing not only gastrointestinal diseases but also neurological disorders. In humans, human astroviruses (HuAstVs) primarily cause gastroenteritis in children, the elderly, and immunocompromised individuals [2], but have also been detected in patients with fatal meningoencephalitis [4], acute flaccid paralysis [5], and respiratory infections [6]. MaAstVs have been identified in domestic animals (e.g., cattle [7], pigs [8], goats [9], rabbits [10]), companion animals (cats [11], dogs [12]), and marine mammals [13], often causing diarrhea [14]. Neurological symptoms have been reported in infected cattle, sheep and minks [15–17]. In some animals (dogs [12], cats [12] and primates [18]), MaAstV infection may also be asymptomatic. In addition, MaAstVs have been found in wildlife (bats [19], foxes [20], raccoon dogs [21]), rodents (murine [22], porcupines [23], squirrels [24]) and camels [25].

MaAstVs pose significant public health and veterinary risks due to their widespread presence and diverse host range. In humans, classic HuAstVs are responsible for approximately 5% of pediatric gastroenteritis cases [6] and contribute to up to 10% of care center outbreaks in the United States [26]. The emergence of novel HuAstV genotypes has further heightened concerns about their potential health impact [6]. Similarly, in animals, MaAstVs exert considerable effects on livestock, companion animals, and wildlife. For example, outbreaks characterized by vomiting have been reported in cats in Canada [27], and in Thailand, detection rates reached 10.8% in cats and 4.2% in dogs [12]. In livestock, cattle exhibit positivity rates ranging from 0.67% to 46.34% globally [7], while pigs show infection rates as high as 95.4% in Greece [28]. Minks with shaking syndrome in China tested positive for MaAstVs at a rate of 63.3% [29], consistent with observations from Poland [30]. Additionally, MaAstVs circulate widely among wild animals, including but not limited to bats, wild boars, rodents, and foxes, although the associated clinical symptoms remain largely unclear [2].

MaAstVs are notable for their cross-species transmission potential, which occurs via direct contact, predation, environmental contamination, or shared habitats [14, 31]. This capability facilitates their emergence in new hosts and may cause severe, unpredictable effects, especially in young animals. Cross-species transmission also enables recombination between distinct viral strains within hosts, driving genetic diversity and environmental adaptation [32, 33]. When a virus enters a new host species, it may co-infect with divergent strains, creating conditions for genetic recombination as the replication machinery exchanges genetic material—as observed in coronaviruses [34]. In MaAstVs, recombination is most frequently observed at the junction between ORF1 and ORF2, whereas recombination within a single ORF is relatively rare [35]. Recombination events have been reported for HuAstV, leading to the emergence of novel strains associated with gastroenteritis in humans [36, 37]. Recombination events in MaAstVs isolated from animals occur either within the same host species, such as humans, pigs, and dogs, or between different host species, such as cat/pig, cattle/deer, pig/deer, and human/sea lion [35].

Given their broad host range and transmission potential, MaAstVs pose significant threats to animal health and the livestock industry. In this study, we compiled comprehensive MaAstV datasets from multiple species and regions, and employed various analytical approaches to investigate their evolution, global spread, and cross-species transmission. Specifically, we performed phylogenetic analyses using ORF2 sequences, estimated evolutionary rates and population dynamics with Bayesian methods, reconstructed transmission pathways via discrete phylogeography, and identified cross-species transmission patterns through ancestral host reconstruction and recombination analyses. These efforts provide critical insights into MaAstV transmission dynamics and host adaptation, informing strategies for prevention and control.

Results

Broad host, geographic and temporal distribution

We collected a total of 11,168 valid sequences (Supplemental Material 1) from NCBI, defining them as sequences that contained at least one of the following metadata: host information, sampling location, or collection date. As shown in Fig. 1, these sequences originated from 86 countries or districts, with most being from China, followed by Japan, Thailand, India, Russia, Brazil, the United States, Germany, South Korea, Gabon, Bangladesh, the United Kingdom, Italy, and Cambodia. We identified 20 host types for MaAstVs. As expected, humans represented the most frequently sampled hosts for MaAstVs, followed by pigs, bats, rodents, cattle, dogs, cats, primates, sheep, and goats. This likely reflects the higher intensity of surveillance and sequencing efforts in human populations. The highest number of all MaAstV detections was recorded in 2012 and 2016. Subsequently, after 2017, the number of detections showed a decreasing trend.

Fig. 1.

Host, geographic, and temporal distribution map of MaAstVs genome samples. A total of 11,168 valid sequences (Supplemental material 1) from NCBI. (a) Geographic; (b) countries/district; (c) temporal distribution; (d) host. MaAstVs: Mamastrovirus

Complex evolutionary history and intertwined evolutionary relationships

Phylogenetic analysis clearly revealed the complex evolution history of MaAstVs (Fig. 2). Based on the phylogenetic relationships in the ORF2 tree, we defined five groups (Groups 1–5, G1–G5) representing distinct evolutionary pathways. These groupings are not official species classifications but reflect phylogenetic clusters derived from our analysis. Except for G5, G1–G4 contained MaAstVs originating from multiple hosts. In G1, marine mammal and rodent AstVs formed the oldest evolutionary branch, which then dispersed into dogs and raccoon dogs. Interestingly, dogs re-transmitted the viruses back to marine mammals, with subsequently circulated among cats, pigs, raccoon dogs, cats, and eventually became widespread in humans. In G2, the evolutionary trajectory originated from bats and rodents, spread to domestic pigs, returned to rodents, then expanded to ruminants, and subsequently circulated among pigs and camels. G3 exhibited a similar pattern but with more intricate relationships: bat AstVs represented the oldest lineage, followed by the formation of ruminant AstVs. This was succeeded by a highly complex branch containing MaAstVs from multiple species, after which evolutionary branches of pigs and bats emerged, culminating in widespread human circulation. In G4, porcine astrovirus (PoAstVs) and BoAstVs clustered into a distinct branch, suggesting a close relationship and potential for cross-species transmission. G5 comprised independent HuAstV-MLB lineages, indicating independent evolution and an unknown origin.

Fig. 2.

Phylogenetic tree of Mamastrovirus of multispecies origin based on ORF2 (n = 1223). AstVs of different species origins are highlighted in different background colors. Phylogenetic analysis was performed based on ORF2 sequences aligned with MAFFT and trimmed with trimAl. The maximum likelihood tree was constructed using IQ-TREE with the best-fit substitution model selected by ModelFinder (GTR+F+R10) and ultrafast bootstrap in 20,000 bootstrap replicates. Turkey astrovirus was used as outgroup (not shown). The blue arrows indicate the ancestral branch for each group. ORF2: open reading frame 2

Notably, MaAstVs from different hosts often clustered within the same branch, or conversely, viruses from the same host diverged into different branches. For example, in G1, Feline (Fe) AstVs and PoAstVs formed a closely related branch; rodent (Ro) AstVs and marine mammal AstVs were clustered together; canine (Ca) AstVs and Raccoon dog (RaD) AstVs were in the same branch. CaAstVs formed common branches with ruminant AstVs and RoAstVs, respectively; camel (Cam) AstVs and PoAstVs were also closely related. Porcupine AstV and PoAstV form close evolutionary branches. Overall, our phylogenetic analyses demonstrate that even within the same host species, MaAstVs exhibit remarkable genetic diversity, which may indicate underlying host-switching processes.

Evolutionary rate and temporal changes in effective population size of MaAstvs

To evaluate the evolutionary rates and temporal changes in effective population size, we first assessed the temporal signal of the individual datasets in TempEst. Our results showed that different MaAstVs exhibited varying degrees of temporal signal (R²: 0.1480–0.9115) (Table 1). The evolutionary rates of MaAstVs varied across lineages, with most falling within the order of 10⁻³ substitutions/site/year (Table 1). PoAstV2 displayed the highest rate (1.319 × 10⁻² substitutions/site/year), whereas HuAstV-MLB had the lowest rate (1.10 × 10⁻³ substitutions/site/year). However, we excluded the evolutionary rate estimates for BoAstV2, BoAstV3, and BoAstV4 due to insufficient temporal signal and overly wide 95% HPD intervals.

Table 1.

Information of sequence number, temporal signal, evolutionary rate and time to the most recent common ancestor (tMRCA) of Mamastrovirus

	Number of Sequence	Temporal Signal (R2)	Evolutionary Rate	95% HPD Interval	tMRCA
Classic Human AstV1	72	0.3174	1.15 × 10−3	[8.12 × 10−4, 1.50 × 10−3]	1943.9
Classic Human AstV3	23	0.9115	2.45 × 10−3	[1.19 × 10−3, 3.81 × 10−3]	1954.8
Classic Human AstV4	25	0.6562	3.57 × 10−3	[1.60 × 10−3, 5.89 × 10−3]	1972.4
Classic Human AstV5	27	0.6257	1.491 × 10−3	[7.02 × 10−4, 2.31 × 10−3]	1975.2
Human AstV-MLB	45	0.2585	1.10 × 10−3	[6.40 × 10−4, 1.62 × 10−3]	1183.8
Human AstV-VA1	49	0.4211	4.32 × 10−3	[2.38 × 10−3, 6.54 × 10−3]	1992.6
Human AstV-VA2	46	0.2701	2.976 × 10−3	[1.36 × 10−3, 4.89 × 10−3]	2015.2
Human AstV-VA3	35	0.2976	3.328 × 10−3	[2.25 × 10−3, 4.43 × 10−3]	2015.4
Porcine AstV2	42	0.2244	1.32 × 10−2	[5.11 × 10−3, 2.21 × 10−2]	1956.2
Porcine AstV3	33	0.1480	1.143 × 10−3	[8.25 × 10−4, 1.35 × 10−3]	1950.5
Porcine AstV4	48	0.1549	9.209 × 10−3	[8.37 × 10−3, 1.01 × 10–2]	1958.8
Porcine AstV5	35	0.2771	4.39 × 10−3	[1.03 × 10–3, 7.96 × 10–3]	1774.9
Porcine AstV-M22	33	0.4031	2.89 × 10−3	[1.30 × 10–3, 4.46 × 10–3]	1916.1
Canine AstV	60	0.1589	3.81 × 10−3	[2.50 × 10–3, 5.27 × 10–3]	1869.2
Feline AstV	52	0.1772	1.262 × 10−3	[3.51 × 10–4, 2.29 × 10–3]	1836.6
Bovine AstV2#	33	0.8537	3.59 × 10–3	[4.8619E-10, 6.2133E-3]	1299.8
Bovine AstV3#	45	0.1005	3.74 × 10–3	[1.1319E-6, 6.5602E-3]	1524.4
Bovine AstV4#	23	0.1951	4.928 × 10–4	[2.8069E-4, 7.3586E-3]	1958.4

Note: ^#The 95% HPD intervals are overly wide, indicating insufficient temporal signal and/or limited sampling power; thus, these estimates are interpreted cautiously and excluded from evolutionary-rate comparisons and discussion

To further analyze changes in effective population size, we reconstructed its temporal trends for individual MaAstVs (Fig. 3). For classical HuAstV, HuAstV-1 exhibited an upward trend in 2016 and maintained a stable level. Similarly, novel HuAstV-VA1 increased in 2017 but slightly declined thereafter. All other HuAstVs showed a decreasing trend. In PoAstVs, effective population size consistently declined across branches, particularly in PoAstV3 and PoAstV-M22. FeAstV did not exhibit significant changes in effective population size. For CaAstV, population size remained stable until 2012, after which it declined steadily. Overall, most MaAstVs showed a decreasing trend in effective population size, although some lineages displayed minor fluctuations.

Fig. 3.

Changes in effective population size of MaAstVs. The intervals represent 95% highest posterior density of the product of generation time and effective population size. The middle line is the inferred median of effective population size. We analyzed the evolutionary rate using the BEAST 2.7 package with separate GTR + Γ nucleotide substitution models, optimized relaxed model and coalescent Bayesian Skyline model were used to estimate the effective population size. AstV: Astrovirus. The classification of the same virus is based on phylogenetic relationships rather than official taxonomy

Global geographic transmission

To investigate the global geographic transmission of various MaAstVs, we reconstructed their transmission pathways using available effective sequences. We visualized the simplest global propagation path by analyzing the MCC tree (Fig. 4). For HuAstVs, transmission occurred both within and between continents, including Europe, Asia, the Americas, and Africa. We identified Italy, China, the United States (US), Russia, and Japan as key “transfer hubs” for HuAstVs. The origins of different HuAstVs varied: HuAstV1 and HuAstV2 likely originated in China, HuAstV3 in Russia, HuAstV4 in Italy, while the newly emerging HuAstV-MLB and VA1 originated in Australia and US, respectively. For animal AstVs, PoAstVs and BoAstVs exhibited global dissemination of multiple genotypes. Like HuAstVs, these viruses spread across multiple continents. In Asia, China, Japan, and South Korea acted as key nodes; in the Americas, US served as a primary hub; and in Europe, countries such as Italy, Spain, Hungary, Belgium, Switzerland, and France were positioned along the transmission routes. Regarding companion animal AstVs, CaAstV displayed a more complex transmission pathway, whereas FeAstV exhibited simpler patterns. Both viruses spread intercontinentally. Italy appeared to be the origin of CaAstV, which disseminated to the Americas, Asia, and Oceania (Australia) after circulating in Europe. China was identified as potential origin for FeAstV, with spread documented between China, the US, and Australia.

Fig. 4.

Geographic diffusion and transmission pathways of MaAstVs based on phylogeographic analysis. The coalescent sky grid model was used as prior of tree. The chain length of MCMC was 1 × 10⁸. The maximum Clade Credibility (MCC) tree is depicted, representing the most probable historical trajectory of phylogenetic relationships. Each node’s geographic state in the MCC tree corresponds to the most credible state among all possible histories, showcasing the definitive diffusion paths across different regions. Bayes factor rates are visualized to assess the relative importance of transition paths between geographic states. Transmission routes with Bayes factors (BF) greater than 3 are highlighted, indicating well-supported diffusion rates between regions. The support levels are categorized as follows: medium support (3 ≤ BF < 10), strong support (10 ≤ BF < 30), very strong support (30 ≤ BF < 100), and extremely strong support (BF ≥ 100). Transmission pathways with BF > 30 are defined as the main transmission pathways. The geographic spread and migration paths were visualized using Google earth, with the corresponding KML file generated by SPREAD v1.0.7

We further quantified the Bayes factor (BF) for each pathway and analyzed major transmission routes (BF > 30). In HuAstV1, HuAstV3, and HuAstV-MLB, nearly all routes received strong support (BF > 30). The primary transmission route for HuAstV4 was from the USA to Europe and Africa, with very strong support (BF > 30). For HuAstV5, the routes from Cameroon-Brazil and Cameroon-Hungary were highly supported (BF > 30). However, the transmission pathways for HuAstV-VA showed only moderate support (3 < BF < 10). The primary transmission routes for different genotypes of PoAstVs varied. For PoAstV2, the major routes (BF > 30) linked South Africa to Uganda, Spain, the USA, China, and Japan. PoAstV3 circulated primarily between Europe, America, and East Asia (BF > 30). PoAstV4 exhibited strongly supported transmission routes across four continents (BF > 30). The main routes (BF > 30) for PoAstV5 and PoAstV-M22 were simpler, primarily between the USA and Chile, and the USA and Hungary, respectively. Among BoAstVs, only the BoAstV3 lineage had a significant transmission route from Italy to Switzerland (BF > 30), while other lineages had BFs less than 30 (3 < BF < 30). The major transmission routes for CaAstV were complex, circulating mainly between Asia, Europe, Oceania, and the Americas. No supported transmission routes were identified for FeAstV (BF < 3).

Cross-species transmission analysis

To investigate the process of cross-species transmission, we analyzed three MaAstV groups (G1, G2, and G3) using the BEAST package. The MCC trees revealed that MaAstV strains originating from dolphins (G1), cattle (G2), and bats (G3) were located at the root of their respective phylogenies, suggesting these hosts may represent ancestral lineages (Fig. 5). In G1, phylogenetic clustering indicates that companion animal-associated viruses (CaAstVs and FeAstVs) may share common evolutionary origins with aquatic mammal-associated strains and may have subsequently spread from dogs to wild animals such as raccoon dogs, foxes, and tigers. This pattern reflects possible host-switching events or shared ancestral viruses. Similarly, in G2, the evolutionary relationships imply recurrent host-switching among livestock species, particularly ruminants, with occasional involvement of wild animals such as rodents and deer. G3 revealed a likely ancestral role of bat-origin viruses, which may have given rise to lineages infecting humans, livestock, and companion animals, possibly via intermediary wildlife hosts. Overall, these results support the hypothesis that MaAstVs have experienced complex host-jumping histories, as observed in the phylogenetic tree, highlighting the dynamic nature of their cross-species evolution.

Fig. 5.

Cross-species transmission patterns based on the maximum Clade Credibility (MCC) tree. The MCC tree was depicted, illustrating the phylogenetic relationships and cross-species transmission patterns across different species. Hosts were treated as discrete traits, and the analyses were conducted using BEAST 2.7 with an MCMC chain length of 1.5–2 × 10⁷ generations. The nucleotide substitution model used was GTR, with an optimized relaxed clock model and a coalescent constant prior for the tree. The posterior of each node and the host of each branch were labeled

Recombination event backtracking further reveals the process of cross-species transmission

Recombination history can provide indirect evidence of cross-species transmission, as recombination typically requires co-infection of distinct viral strains, often originating from different hosts [35]. To further analyze and verify the existence of cross-species transmission processes, we conducted a large-scale recombination analysis uncovered several previously undiscovered recombination relationships. In total, we identified 105 recombination events, which demonstrated substantial diversity, encompassing both intra-species and cross-species recombination (Supplemental material 2). PoAstVs were the most affected, displaying both intra- and cross-species patterns. In cross-species recombination events, we detected recombination between PoAstVs and AstVs from rodents, bovines, humans, sheep/goats, and porcupines, suggesting that MaAstVs can circulate among animals of these species. In contrast, most of HuAstVs exhibited intra-species recombination. BoAstVs were found to recombine with CaAstVs as well as with ruminant AstVs. Additionally, recombinant strains of BoAstV and PoAstV could infect species such as mink and sheep/goat. In cats, beyond the reported recombination events of FeAstV and PoAstV, we identified recombinant strains of RaccoonDog AstV and HuAstV in cats. Recombination events between SlAstV and HuAstV have been reported, and we also found that SlAstV and CamAstV recombinant strains can infect Porcupine. These cross-species transmission events amply demonstrate that MaAstVs has the ability to spread across multiple species. To more clearly indicate the potential diversity of MaAstVs cross-species transmission, we mapped the relationships of AstV cross-species transmission based on the current results (Fig. 6). Our findings expanded the spectrum of cross-species transmission of AstV.

Fig. 6.

Mapping of recombination events of MaAstVs between different hosts. Strains and recombination events from different hosts are labeled with different colors. Detailed recombination events were shown in the Supplemental material 2

Discussion

AstVs, which are recognized for their zoonotic potential, exhibit a broad host range and genetic diversity [36]. Globally, MaAstVs continue to be reported in a wide range of hosts, and frequent cross-species transmission has been elucidated directly or indirectly [33]. In addition, the high rate of evolution of RNA viruses has resulted in an extreme diversity of genotypes of AstV and has led to taxonomic difficulties [33]. In this study, we analyzed in detail the evolution, global spread, and cross-species transmission of AstVs using the available data, with the aim of fully elucidating these important features of MaAstV, that is, the evolutionary rate and dynamics, global transmission process, and cross-species transmission.

The distribution of MaAstV was uneven. Geographically, MaAstVs were omnipresent, with the highest detection rates reported in China, Japan, Thailand, India, Russia, Brazil, and the United States. On the time scale, after peaking in 2012 and 2016, the number of detections has been declining in recent years. Several factors may contribute to this decline, including improved sanitation, enhanced surveillance and prevention measures, and natural fluctuations in virus transmission dynamics. Additionally, the apparent reduction may also reflect changes in sampling strategies, reduced research focus, or underreporting in certain regions in recent years. It is important to note that these data reflect the distribution of available sequenced genomes, rather than the actual number of MaAstV incidence in each country. The number of available genomes of viruses, including MaAstVs, are influenced by a range of factors such as a country’s economic capacity, research resources, sampling intensity, geographic and climatic conditions, and level of technological development, as has been discussed in the context of other viruses such as SARS-CoV-2 [38]. A more systematic epidemiological investigation is needed to accurately assess the incidence and impact of MaAstVs within each region.

The wide range of hosts undoubtedly provided a vast evolutionary venue for MaAstVs, with hosts encompassing a variety of species, including humans, wildlife, and domestic animals. Humans are undoubtedly the most frequently monitored hosts of AstVs. However, as MaAstVs infect a wide range of species, they encounter diverse host immune systems, tRNA abundances, and cellular environments. These inter-host differences impose distinct selective pressures on the virus, influencing codon usage and driving viral evolution, potentially contributing to the genetic diversity of MaAstVs [39–41].

To further reflect the changes in the genetic diversity and historical dynamics of MaAstVs, we analyzed the changes in their effective population size over time using sequences from multiple hosts. The results were similar to those summarized using data from NCBI, with a decreasing trend in the effective population size of most MaAstVs and fluctuations in a few. A study showed that the effective population of classical HuAstV was in a state of decline, consistent with the results of the present study [42]. As novel subtypes of HuAstV, MLB and VA, MLB was similarly on a downward trend, while VA rose followed by a decline in 2017. This study revealed for the first time the trend of effective population size of animal AstVs. Similar to HuAstV, most of them were in a decreasing trend, and some effective population sizes remained stable.

In addition, different viruses exhibit distinct evolutionary rates, potentially shaped by their evolutionary origins and host environments [43]. For example, PoAstV2 displayed the fastest evolutionary rate (1.32 × 10⁻² subst/site/year), whereas HuAstV-MLB (1.10 × 10^{− 3} subst/site/year), PoAstV3 (1.143 × 10⁻³ subst/site/year) and HuAstV1 (1.15 × 10⁻³ subst/site/year) evolved more slowly (Table 1). The variability in evolutionary rates may reflect the adaptation of viruses to different host species, each providing unique selective pressures, such as immune responses, cellular environments, and codon usage biases [44]. These factors can influence viral replication fidelity, mutation rates, and selection dynamics, collectively driving the observed differences in evolutionary speed. Moreover, frequent cross-species transmissions expose MaAstVs to novel host environments, which may accelerate their evolution by imposing new constraints and opportunities for adaptation [45].

Numerous studies have reported that MaAstVs have already spread globally [8, 33]. Based on phylogeographic reconstruction, we found different lineages of MaAstVs from the same host have different origins, e.g., Italy, China, Australia, Russia, and the United States all serve as the origin of MaAstVs, which have been dispersed to multiple countries through a number of important node countries. It should be noted that these findings are based on the currently available sequence data, and the identification of transfer hubs may shift as more ancient or newly collected samples become available. However, the reconstructed MCC trees and dispersal histories suggested that transboundary transmission events occurred repeatedly over recent decades. The global transboundary transmission of viruses appears to be widespread; however, the driving forces facilitating the spread of human and animal viruses differ. The global dissemination of human viruses is primarily driven by international travel and human movement, while the spread of animal viruses is mainly facilitated through international trade or animal migration [46]. A relevant example is the recent global spread of SARS-CoV-2, where international travel played a key role in the repeated introduction and reconstruction of transboundary chains [47]. Similar global dissemination patterns have been observed in other animal viruses. For example, the spread of highly pathogenic avian influenza has been linked to both wild bird migration and international poultry trade [48]. Similarly, the transboundary spread of foot-and-mouth disease virus and porcine epidemic diarrhea virus have been facilitated by livestock movement and trade [49, 50].

The global spread of viruses has multiple far-reaching consequences. It not only enhances the genetic diversity of viruses through repeated transboundary transmission and adaptation to different hosts or environments [51], but also poses significant threats to industries, particularly livestock and agriculture, where outbreaks of diseases like foot-and-mouth disease have led to substantial economic losses and trade restrictions [49]. Furthermore, the global dispersal of animal viruses increases the risk of zoonotic spillover events, potentially giving rise to emerging infectious diseases in humans [52, 53]. These dynamics underscore the importance of robust biosecurity and surveillance systems for MaAstVs, especially in major trading countries, to enable early detection of emerging or novel strains and to mitigate risks associated with international trade, including trade restrictions, supply chain security, public health threats, and global economic instability driven by potential cross-species transmission [54].

Host jumping occurs frequently in MaAstVs [14], as evidenced by both recombination events and phylogenetic clustering of viruses from diverse animal hosts revealed in this study. In phylogenetic analysis, AstVs from diverse animals were clustered, especially from small carnivores (including dogs, cats, minks, foxes, raccoon dogs, and red pandas) and ruminants (cattle, deer, goats/sheep, takins, and yaks), underscoring the ease of cross-species cycling within certain animal classes. Patterns of transmission between related animals are common, and in addition to AstVs, canine distemper virus can threaten a wide range of carnivores [55], and foot-and-mouth disease is a threat to a wide range of ruminants [56]. In addition, the study revealed extensive and diverse recombination patterns in AstVs, particularly concerning the prevalence of cross-species recombination. In previous studies, recombination events between PoAstV and HuAstV [57], between PoAstV and FeAstV [11], between HuAstV and FeAstV [58], and between SLAstV and HuAstV [13] have been elucidated. In the present study, we investigated novel potential cross-species recombination patterns between HuAstV and the Raccoon dog AstV, PoAstV and Porcupine AstV, Mink AstV and RPAstV, CaAstV and CamAstV, and CaAstV and BoAstV, which further revealed the diversity of the recombination. The occurrence of inter-species recombination provides further evidence for the cross-species transmission of MaAstVs. Additionally, although MaAstVs have been most frequently detected in humans, evidence of human-animal cross-species transmission is limited. To date, only two studies have reported MaAstV strains isolated from pigs and sea lions that exhibit recombination with human-derived strains [13, 57]. However, no animal-derived genetic components have been identified in MaAstVs isolated from humans. Whether interspecies transmission between animals and humans is unidirectional or bidirectional remains to be elucidated.

The potential reason for cross-species transmission may be the diversity of receptors for AstVs that allows them to enter different host cells. Currently, there are no definitive studies revealing receptor of AstVs. However, one study reveals that the receptors for classical HuAstV and HuAstV MLB are different [59]. Using the Beast package, we reconstructed the cross-species transmission process of MaAstV. In G1, where the aquatic animal AstV becomes the ancestor of multiple MaAstVs. This was followed by gradual transmission to humans, domestic animals and wildlife using dogs and cats as “springboard”. This transmission process reveals the important role of companion animals in the spread of the virus in animals and humans. In G2, MaAstVs circulated mainly in domestic animals, especially in ruminants. BoAstV became an important source, gradually spreading to other ruminants, dogs, pigs, and some wild animals. In G3, bat became an important origin. Bats act as important “virus reservoirs” and continue to transmit novel viruses to animals and humans, causing some degree of harm [60]. These results are consistent with the results of the recombination. Combining these two results, we further elucidated the cross-species transmission process of MaAstV.

Our study has certain limitations. Recombination mainly occurs at the junction of ORF1b and ORF2 [35]. In constructing our datasets, we did not exclude potential recombinant sequences, as recombination within ORF2 is relatively limited and unlikely to substantially affect the analyses. This approach also ensured sufficient dataset size. However, we acknowledge that relying solely on ORF2 may not fully capture the evolutionary complexity of the virus, and future studies using non-recombinant regions or complete genomes could provide more comprehensive insights. In addition, for some AstVs lineages, insufficient temporal signal led to wide HPD intervals, limiting reliable evolutionary rate estimation. Despite these constraints, our approach demonstrates the utility of ORF2-based analyses for uncovering broad patterns of phylogeny, recombination, and host associations. This framework could be applied to other viruses with similar evolutionary features. Moving forward, integrating full-genome data, more comprehensive temporal sampling, and recombination-aware analytical methods will be essential to refine our understanding of astrovirus evolution and cross-species transmission dynamics.

Conclusion

In conclusion, this study systematically analyzes the characteristics of MaAstV in terms of evolution, global spread and cross-species transmission, providing important insights for further understanding of AstVs. The limitation of this study is that only the existing available sequence information was used. Further investigations are needed to explore MaAstV global spread, cross-species transmission, and evolutionary dynamics.

Methods

Sequence sample collection and analysis

Sequence information was obtained from the National Center for Biotechnology Information (NCBI). We retrieved MaAstVs sequence data using the keyword “(Astroviridae[Organism]) NOT Avastrovirus[Organism]”, which retrieved all MaAstV sequences while excluding avian AstVs. All the strains were statistically analyzed for their hosts, collection date, and countries. The retrieved sequences were grouped into two main datasets: complete genomes and ORF2 sequences, to enable genome-wide recombination analyses and ORF2-based phylogenetic and evolutionary studies, respectively.

Phylogenetic analysis

We conducted phylogenetic analysis to resolve the evolutionary relationships of MaAstVs across hosts and regions, enabling the identification of lineage diversification, host associations, and potential cross-species transmission events. We built a dataset containing ORF2 sequences (n = 1223) and performed phylogenetic analyses based on this dataset. The ORF2 sequences were aligned using multiple sequence alignment with the fast Fourier transform (MAFFT) [61]. We performed automated alignment trimming using trimAl v1.2 [62] and the sequences were concatenated using the concatenation module in PhyloSuite v1.2.1 [63]. We selected the best substitution model using ModelFinder [64] and then inferred the maximum likelihood (ML) phylogenies with IQ-TREE, employing ultrafast bootstrap and 20,000 replicates to assess branch support [65]. We visualized and modified the phylogenetic trees using Interactive Tree of Life (iTOL) [66].

Effective population size of MaAstvs

Effective population size represents the genetic diversity of the viral population and reflects its demographic history over time [43, 67]. To investigate the effective population size, we first constructed ORF2 datasets for different host species. Based on the ML phylogenetic tree, MaAstVs were categorized into subgroups for subsequent effective population size analysis. Specifically, human AstVs were divided into eight subgroups: HuAstV1, HuAstV3, HuAstV4, HuAstV5, HuAstV-MLB, HuAstV-VA1, HuAstV-VA2, and HuAstV-VA3. Porcine AstVs were classified into PoAstV2, PoAstV3, PoAstV4, and PoAstV5. Bovine AstVs (BoAstVs) were divided into BoAstV2, BoAstV3, and BoAstV4. These designations were used solely for analytical distinction between phylogenetic branches and do not represent formal taxonomic classifications.

Firstly, We used the ORF2 gene to construct the ML phylogenetic tree according to the method described above and assessed temporal signal of datasets using TempEst v1.5.3 [68]. Abnormal sequences were manually deleted. The best model through marginal likelihood estimated using path-sampling and stepping-stone estimation approaches. We analyzed the evolutionary rate using the BEAST 2.7 package with separate GTR + Γ nucleotide substitution models on the three partitions based on the codon positions and optimized relaxed model [69]. We evaluated the effective population size over time using the coalescent Bayesian Skyline model. The MCMC chain length was vary from 5 × 10⁷ to 2 × 10⁸ in different MaAstVs datasets. Then, we used Tracer software v1.7 with a burn-in period of 10% of the total chain length for Convergence examination [70]. The effective sampling size for all parameter estimates was greater than 200 [71].

Geographic diffusion and transmission pathways of MaAstVs based on phylogeographic analysis

Phylogeographic analysis aimed to infer the global transmission routes of MaAstVs by integrating phylogenetic relationships with geographic data. Using a discrete trait model and Bayesian inference, this method estimates the most probable migration pathways, with support quantified via Bayes factors. For global discrete phylogeographic analyses, we used the ORF2 dataset as the effective population size. The nucleotide substitution model was GTR + Γ，and the clock model was optimized relaxed model. The coalescent Sky Grid model was used as prior of tree [71]. The chain length of MCMC was 1 × 10⁸. Convergence was inspected using Tracer 1.7 [70] to ensure that effective sampling size values associated with relevant parameters were all >200. To comprehensively visualize the results of our discrete phylogeographic analysis, we employed two distinct methods: the Maximum Clade Credibility (MCC) tree and Bayes factor rates.

Firstly, the MCC tree, selected from Bayesian tree samples, represents the tree with the highest clade credibility. This method allows us to illustrate the most probable single historical trajectory of phylogenetic relationships. By using the MCC tree, we can clearly observe the transitions between different geographic states. This approach simplifies the complex phylogenetic history, providing an intuitive depiction of the most likely geographic diffusion process. Specifically, each node’s geographic state in the MCC tree represents the most credible state among all possible histories, thus displaying a definitive diffusion path on the map [72].

Secondly, to comprehensively assess the transition paths and their relative importance between geographic states, we employed Bayes factor rates. Bayes factor rates measure the relative support for transition rates between different geographic states by comparing them. High Bayes factors indicate strong evidence supporting a particular transition path, while low Bayes factors indicate weaker support. Visualizing Bayes factor rates allows us to display multiple possible migration paths and their relative importance on the map, revealing the complexity and dynamic nature of the geographic diffusion process [73].

The combination of these two methods enables us to comprehensively present and understand the geographic diffusion process from different perspectives: the MCC tree provides a simplified and most credible diffusion history, while Bayes factor rates reveal potential multiple migration paths and their support. This dual visualization approach ensures that we can accurately and thoroughly explain the results of our discrete phylogeographic analysis.

The plausible trees were summarized by TreeAnnotator 1.10.5 [69]. SPREAD V1.0.7 [74] was used to summarize the geography spread in a KML file. The KML file was visualized by Google earth. The transmission routes for which Bayes factor (BF) is greater than 3 were considered to have well-supported diffusion rates, indicating well-supported migration routes between different geographical areas. Further, 3 ≤ BF < 10: medium support; 10 ≤ BF < 30: strong support; 30 ≤ BF < 100: very strong support; BF ≥ 100: extremely strong support [75]. We defined the transmission pathway at BF > 30 as the main transmission pathway.

Cross-species transmission analysis

Cross-species transmission analysis aimed to trace host-switching events by reconstructing ancestral host states using a Bayesian discrete trait model. Recombination analysis complemented this by identifying genetic exchanges indicative of interspecies transmission. To investigate the process of cross-species transmission, we established three datasets, G1 (N = 78), G2 (N = 114) and G3 (N = 78), containing multiple species based on the phylogenetic tree. We treated the host as a new discrete trait and conducted ancestral state reconstruction using BEAST 2.7 with an MCMC chain length of 1.5–2 ×10⁷. The nucleotide substitution model employed was GTR + Γ, and the clock model was an optimized relaxed model. The coalescent constant model was used as the prior for the tree. By running the MCMC chain and analyzing the output with Tracer software to assess parameter estimation effectiveness, we ultimately extracted the MCC tree from the MCMC samples using TreeAnnotator to interpret cross-species transmission patterns and visualized the MCC tree by Figtree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/).

Recombination reflects cross-species transmission history as it requires co-infection of divergent viruses within the same host, enabling genetic exchange [76]. Thus, interspecies recombination provides molecular evidence of historical host switching events. We therefore constructed a complete sequence dataset (n = 528) used for recombination analysis. Multiple alignment was performed using MAFFT. The pairwise homoplasy index (PHI) test was performed to predict the occurrence of potential recombination events [77]. Thereafter, recombination events were evaluated using nine methods (RDP, GENECONV, Bootscan, Maxchi, Chimaera, SiSscan, PhylPro, LARD, and 3Seq) in the Recombination Detection Program (RDP) v5.34 [78]. Recombination events detected by more than four methods were considered credible. Strains that may have been caused by evolution were excluded. Simplot v3.5.1 was used to further determine reorganization events [79].

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

MaAstV

Mamastroviruses

ORFs

Open reading frames

AstV

Astrovirus

HuAstV

Human astrovirus

PoAstV

Porcine astrovirus

BoAstV

Bovine astrovirus

RoAstV

Rodent astrovirus

CaAstV

Canine astrovirus

RaDAstV

Raccoon dog astrovirus

CamAstV

Camel astrovirus

Tmrca

The most recent common ancestor

Subst

Substitutions

MCC

The Maximum Clade Credibility

BF

Bayes factor

Author contributions

Y.W. conceptualized the study, initially analyzed the data and validated as well as drafted the first manuscript. H.W. collected data and performed host and geographic area distribution analyses, phylogenetic analyses, evolutionary rates, and global transboundary geographic analyses. X.L. and H.M. performed recombination analysis and cross-species transmission analysis. A.L. and Y.L. validated the data, with Y.L. receiving funding support. X.G. conceptualized and methodologized this study and conducted the phylogenetic analyses, evolutionary rates, and global geographic transmission analyses. Y.W., Y.L. and X.G. drafted and revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Dabie Mountain Comprehensive Experiment Station Industry Union Project of Anhui Agricultural University (CYLM2024-10/11), Ningbo Top Medical and Health Research Program (No. 2023020713), the Ningbo Key Laboratory of Virus Research (20221CXJD030031), and Science and Technology Project of Zhejiang Provincial Center for Disease Control and Prevention (Grant No. 2025JK071). The authors declare that there are no conflicts of interest associated with these funding, and funding bodies have no influence on the independence of the research.

Data availability

All data generated or analysed during this study are included in this published article and its supplementary files. The information on the strains used in the study is available in the NCBI database (https://www.ncbi.nlm.nih.gov/nucleotide/).

Declarations

Ethics approval and consent to participate

Not applicable. This is a secondary analysis of publicly available NCBI viral genome data and metadata; no human/animal research, clinical specimens, or live virus experiments were conducted. Ethics approval and informed consent were therefore not required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.