Background
Diarrhea in pigs, especially in piglets, is a major issue affecting global swine industry. A diverse array of viral pathogens are implicated in piglet diarrhea, including porcine epidemic diarrhea virus (PEDV), transmissible gastroenteritis virus (TGEV), porcine deltacoronavirus (PDCoV), porcine rotavirus (PRoV), porcine bocavirus, Aichivirus C (formerly named as porcine kobuvirus 1 [PKV-1]), porcine sapovirus, porcine sapelovirus (PSV), and porcine astrovirus as well as others [
1]. Among them, PEDV, TGEV, PDCoV and PRoV have been well characterized for their role in clinical diarrhea and have been considered as major swine enteric viral pathogens while the causative role of the other agents in diarrheic disease has been characterized to a significantly less extent or remains to be further studied.
PEDV is an enveloped, single-stranded, positive-sense RNA virus belonging to the Order
Nidovirales, the family
Coronaviridae. Initially identified in Europe in 1970’s, PEDV was later detected in many Asian countries where it causes severe problems in pig populations. In North America, PEDV was first detected in the United States (US) in 2013, which rapidly spread across the US in the subsequent 2 years [
2‐
4]. The diarrhea caused by PEDV can result in dehydration in suckling piglets and eventually up to 100% mortality, which caused significant economic loss to the US swine industry [
5,
6].
Using virus-specific PCRs, viruses such as PDCoV and PRoV have been detected along with PEDV in some diarrheic cases submitted to the Iowa State University Veterinary Diagnostic Laboratory (ISUVDL) and other diagnostic laboratories of different states [
7,
8]. However, there have been no reports thoroughly investigating viruses concurrent co-infections with PEDV in diarrheic pigs. The random primers-based, hypothesis-free High-throughput sequencing (HTS) enables simultaneous detection of multiple microorganisms, making it possible to identify not only known pathogens, but also novel and/or uncharacterized potential pathogens. In this study, a HTS-based metagenomics approach was used to characterize the fecal RNA virome in diarrheic pigs infected with PEDV. Some previously uncharacterized viruses were detected and genetically characterized in the present study.
Discussion
Next-generation sequencing technology, characterized by massive sequence output, low cost, and short turnaround time, have significantly changed the path of new pathogen discoveries. There are various potential applications of HTS in clinical virology, such as whole viral genome reconstruction, the discovery of uncharacterized and/or novel viral pathogens, the detection of mixed infections, the identification of viral variants and quasispecies, outbreak tracking, the disease surveillance, and the characterization of host responses. These applications have revolutionized infectious disease diagnostics and clinical microbiology. Therefore, HTS technology has obvious advantages over traditional diagnostic methods. However, most clinical samples are very “dirty” and composed of both host and microbial nucleic acids; in most cases, 99% of nucleic acid contained in samples belong to either the host or other microorganisms of no interest. As a result, searching for particular microbial nucleic acids is similar to searching for “a needle in a haystack”. Therefore, metagenomics approach for mixed infection detection and novel viral discovery relies extensively on bioinformatics to tackle the huge amounts of sequence data.
Initially identified from European countries in the 1970’s, PEDV became a major issue in many Asian pig-production countries before 2013 [
22]. Since 2013, PEDV has been newly emerged in the US and other countries of North, Central and South America, and re-emerged in several European countries, suggesting that there is a global epidemic of PED [
22]. In addition, at least two different lineages of PEDV are co-circulating in North American and Asian countries [
23,
24], compromising vaccine-based control strategies. A majority of previous studies focused on PEDV evolution and pathogenicity whereas only two studies investigated the changes detected in the fecal microbiota from pigs with PEDV infection [
25,
26], both of which only focused on the changes of bacterial profiles in pigs infected with PEDV. Utilizing HTS to characterize the virome of diarrheal pigs, four previous studies reported that significantly higher percentages of RNA viral sequences were detected than that of DNA viral sequences [
1,
27‐
29]. However, the aforementioned studies only utilized limited numbers of diarrheal pigs (≤50) and the role of PEDV in the sick pigs was not clear. Here, we characterized for the first time the RNA virome of 217 PEDV-infected pigs in the US.
Utilizing the HTS and the bioinformatics workflow described in this study, the RNA virome in clinical fecal samples of diarrheic pigs included Mamastrovirus, Enterovirus, Sapelovirus, Posavirus, Kobuvirus, Sapovirus, Teschovirus, Pasivirus, and Deltacoronavirus in decreasing order of prevalence. Our data indicate the high prevalence (73%) of mixed RNA viral infections in PEDV-infected pigs and the complexity associated with mixed viral RNA infections.
All detected viruses in this study belong to four viral families (
Picornaviridae,
Coronaviridae,
Astroviridae,
Caliciviridae), and 6 out of the 10 virus genera belong to
Picornaviridae family (
Enterovirus,
Posavirus,
Sapelovirus,
Kobuvirus,
Teschovirus, and
Pasivirus). Such findings are similar to a previous report [
27]. This may have been attributed to the well-understood fact that picornaviruses are highly resistant in the environment and as such, usually survive much longer than other viruses [
30]. The environmentally resistant picornaviruses can have a longer exposure time in pigs, increasing the odds of infection and their prevalence when compared to other viruses. Furthermore, the viruses also have better odds of detection in the environment after shedding. Both of these factors can increase the detection rate of picornaviruses from clinical samples, as compared to other viruses. Although the longer surviving feature of picornaviruses makes them easier to detect, such nature may bias their proportion in mixed infections.
In addition to obtaining mixed viral infection profiles of each sample, complete genome sequences of Aichivirus C and PSV were obtained and further characterized in this study. The clinical significance of PSV has not been confirmed to date, but this complete US PSV genome sequence can facilitate PSV diagnostics and future molecular epidemiological studies of PSV in US swine. It was also demonstrated in this study that there are at least two genogroups of Aichivirus C (PKV-1) circulating in US swine. Currently, Aichivirus C-specific virological and serological diagnostic assays have not been available in most veterinary diagnostic laboratories. Availability of more Aichivirus C sequences will help understand genetic diversity of this virus and facilitate development of appropriate diagnostic assays. The clinical significance of Aichivirus C also remains to be elucidated.
Because all samples selected in this study were from PEDV-positive pigs which had PEDV-associated clinical diarrhea disease, the clinical importance of other detected viruses could not be determined in the current study. Using a similar workflow, the virome diversity between pigs with different health status using a large sample size could be investigated in the future.
The comparable sensitivity and specificity between Kraken and Kaiju has also been studied previously [
13]. Amino acid sequences are more conserved and are more tolerant to sequencing errors than underlying DNA, which may provide higher specificity (accuracy) to Kaiju program [
13]. However, the sensitivity of Kaiju is concerned to be compromised by its inability to classify nucleic acid reads in non-coding region [
13]. Therefore, we chose to use Kraken as our primary analysis program, and Kaiju program as our secondary confirmation method in the current study. It should be noted that a fecal sample negative for PEDV was not used as a negative control, which is one of our study limitations.