The major challenge in the control of viral diseases in aquaculture is the long duration it takes from the first time the disease is discovered through clinical reports to identification of the etiological agent. This delay precludes our ability to develop timely disease control strategies. For example, IPNV was first reported as a clinical disease causing acute catarrhal enteritis in salmonids in 1940 (McGonigle, 1941) and yet the virus was first identified and characterized 20 years later in 1960 (Wolf et al., 1960). VHSV was first identified in 1962 (Jensen, 1965) followed by IHNV in 1969 (Wingfield et al., 1969) although reports of mortalities caused by both viruses date far back in the early 1950s. In the case of nodavirus, infections causing major problems in hatcheries as well as clinical signs linked to lethargy, anorexia, pale coloration and corkscrew swimming in barramundi (Lates calcarifer) and other fish species in Asia were first reported in the 1970s and yet the virus was first characterized in 1990 (Glazebrook et al., 1990; Munday et al., 2002). SAV was first reported as a disease causing skeletal muscle and cardiac myopathy in salmonids in 1976 and yet the virus was first identified and characterized almost two decades later in 1995 (Boucher et al., 1994; Castric et al., 1997). Outbreaks of cardiomyopathy syndrome (CMS) were first reported in 1985 (Amin and Trasti, 1988) while the virus responsible for the disease was recently characterized as piscine myocarditis syndrome virus (PMCV) in 2011 (Haugland et al., 2011). Similarly, heart and skeletal muscle inflammation (HSMI) outbreaks were first reported in 1999 (Kongtorp et al., 2004a,b) and yet the etiological agent was recently identified after a decade as piscine reovirus (PRV) now referred to as piscint orthoreovirus (Løvoll et al., 2010; Palacios et al., 2010). Table 1 shows a summary of the major fish viral diseases showing the year when clinical cases were first reported and the year when the etiological agent was identified. The general trend is that etiological agents were only identified after they caused disease outbreaks reaching epidemics proportions, which is in line with observations made by Alavandi and Poornima (2012) who pointed out that our response to viral infection has mainly been reactive in the sense that a new pathogen is usually not identified until it has reached epidemic proportions. Hence, there is need to develop proactive diagnostic tools able to identify novel viruses before they cause disease outbreaks leading to high economic losses in aquaculture.

Apart from fish farming, shrimp farming is one of the rapidly expanding aquaculture industries whose expansion has brought with it a corresponding increase in the number of novel pathogenic viruses being discovered. Similar to observations in fish, Table 2 shows that the year when the first clinical case was identified in shrimp and the year when the etiological agent was identified, depicting long intervals allowing the diseases to reach epidemic levels before the causative agents were identified. For example, mortality due to Taura syndrome (TS) reached high epidemic proportions spreading from the Taura river in Ecuador, where the disease was first reported (Jimenez, 1992), expanding to Peru, Colombia, Brazil, Honduras, and Hawaii. It spread to Asia where it infected different shrimp species before the virus was identified (Walker and Winton, 2010). The cause was initially attributed to fungicide toxicity (Lightner et al., 1994) and the subsequent identification of Taura syndrome virus (TSV) as the etiological agent of the disease led to development of molecular diagnostic tools that paved way to breeding and cultivation of specific pathogen free (SPF) stocks to ensure that all stocks used in shrimp farming were free of the disease (OIE, 2016). Similarly, white spot syndrome virus (WSSV), shrimp infectious myonecrosis virus (IMNV), Penaeus vannamei nodavirus (PvNV) and several other shrimp viral diseases reached high epidemic levels before their etiological agents were identified and characterized (Lotz, 1997; Arcier et al., 1999; Tang and Lightner, 1999; Yang et al., 2001; van Hulten et al., 2001). Identification of the causative agents for these diseases paved way to developing diagnostic tools that are currently used for the screening of breeding stocks to ensure that only SPF stocks are used for shrimp production (Lotz, 1997; Arcier et al., 1999; Tang and Lightner, 1999; van Hulten et al., 2001; Yang et al., 2001; OIE, 2016). Hence, it is important that proactive diagnostic tools able to timely identify novel pathogenic viruses are devised in order to expedite the process of developing effective disease control strategies against emerging viral disease in aquaculture.

Although calicivirus infections date as far back as 1932 when pigs that were fed raw garbage in Los Angeles in California resulted in an outbreak that caused high mortality in the infected pigs, the disease in marine mammals was first reported in 1972 from rectal swabs of Californian sea lions (Zalophus californianus) that had just aborted on San Miguel Island (Smith and Boyt, 1990). The causative agent was characterized and named San Miguel sea lion virus type I (SMSV-1). When SMSV-1 was administered in pigs, it caused classical vesicular exanthema syndrome in pigs comparable to the 1932 outbreak (Smith et al., 1973, 1998; Smith and Boyt, 1990). By 1982, calicivirus virus infections had been isolated from 11 pinnipeds and cetaceans and the host range of infected species has continued to increase (Smith et al., 1998). Marine mammal morbilliviruses were first discovered in late 1980s following a large outbreak in which approximately 18,000 harbor seals and gray seals were found dead on Northern European coasts in 1987–1988. The etiological agent was classified as phocine distemper virus (PDV). Similar high mortalities involving thousands of animals were reported in subsequent years in different species including porpoises (Phocoena phocoena) (Kennedy et al., 1988), Phoca caspisa (Kennedy et al., 2000; Ohashi et al., 2001; Kuiken et al., 2006), Stenella coeruleoalba (Domingo et al., 1990), Tursiops truncates (Lipscomb et al., 1994a,b), and Phoca vitulina (de Swart et al., 1996), harp seals (Phoca groenlandica) (Markussen and Have, 1992). The first report of influenza virus infections in marine mammals was in 1979 when more >500 harbor seals were found dead on the North Eastern Coast of the USA due to H7N7 subtype infections (Lang et al., 1981; Webster et al., 1981b). In 1982–1983, an outbreak of influenza due to subtype H4N5 caused mortality in more than 60 harbor seals in Massachusetts coastal areas in the USA (Hinshaw et al., 1984). Since then, influenza A and B viruses have been isolated from different cetaceans and pinnipeds of which postmortem handing has been linked to zoonotic infections in humans (Webster et al., 1981a,b; Hinshaw et al., 1984, 1986).

The earliest reports of adenovirus infections in marine mammals were in the late 1970s in large numbers of sea lions that had clinical signs of hepatitis and enteritis (Britt et al., 1979; Dierauf et al., 1981; Goldstein et al., 2011). Since then, adenoviruses have been isolated from California sea lions, South African fur seals (Arctocephalus pusillus), South American sea lions (Otaria flavescens) (Inoshima et al., 2013), sei whales (Balaenoptera borealis) (Inoshima et al., 2013), bowhead whales (Balaena mysticetus) (Smith et al., 1987; Inoshima et al., 2013), beluga whale (Delphinapterus leucas), and bottlenose dolphins (T. truncates) (Rubio-Guerri et al., 2015). The first report of marine mammal herpesvirus infection was from an outbreak of several harbor seals that had acute pneumonia and hepatitis in the Netherland in 1985 (Osterhaus et al., 1985). Thereafter, herpesviruses have been isolated from several species including harbor seal (P. vitulina) (Borst et al., 1986), bottlenose dolphin (Tursiops truncatus) (Manire et al., 2006; van Elk et al., 2009), orca (Orcinus orca) (Maness et al., 2011), California sea lion (King et al., 2002; Buckles et al., 2006), and gray seals (Halichoerus grypus) (Harder et al., 1996). Other viruses shown to infect marine mammals include poxviruses that have been reported in cetaceans and pinnipeds of which some are zoonotic (Van Bressem et al., 1993; Bracht et al., 2006; Waltzek et al., 2012) and astroviruses that have been associated with diarrhea in California sea lions, Steller sea lion (Eumetopias jubatus) and bottlenose dolphin (Rivera et al., 2010). Put together, these studies further consolidate the notion that our response to viral infections in aquatic organisms has mainly been reactive given that the causative agents for these diseases were only identified after they caused massive outbreaks reaching epidemic proportions. Hence, there is need for proactive diagnostic tools for the timely identification of novel pathogens having the potential to cause disease outbreaks in order to help design effective disease control strategies.

Recently identified pathogenic viruses of fish include the circovirus isolated from barbell fish (Barbus barbus), which causes mortality within 4–6 days after hatching (Lőrincz et al., 2011). Phylogenetic analysis of two complete genomes classified as Barbell circovirus 1 and 2, (BaCV1 and BaCV2) showed that these viruses belong to two new genetic groups within the Circoviridae family, which are distinct from previously known circoviruses (Lőrincz et al., 2011). Apart from barbell fish, circoviruses have been detected in common bream (Abramis brama), asp (Aspius aspius), round goby (Neogobius melanostomus), monkey goby (Neogobius fluviatilis), and roach (Rutilus rutilus) (Tarján et al., 2014). Novel circoviruses have also been identified from the European catfish (Silurus glanis) (Lőrincz et al., 2012) and European eel (Anguilla anguilla) showing clinical signs of a cauliflower-like disease (Fichtner et al., 2013). Reuter et al. (2015) identified a novel posavirus designated as Fisavirus 1 (FisaV1) from freshwater carp (Cyprinus carpio) and a novel seadornavirus resembling the mammalian Banna virus from freshwater carp (Reuter et al., 2013). In another study, Boros et al. (2011) identified a novel (+)ssRNA virus with a di-cistronic genome in carp while Mor and Phelps (Mor and Phelps, 2016b) identified a novel totivirus from Gold shiner (Notemigonus crysoleucas) baitfish using NGS. The majority of viruses shown in Table 3 have only been discovered in the last 4 years unlike viruses in Table 1, which took long to discover using traditional methods. Hence, the rate at which viruses in Table 1 were discovered compared to viruses in Table 3 supports our notion that NGS and viral metagenomics are proactive diagnostic methods that have significantly accelerated our ability to identify novel viruses of fish.

Ng et al. (2013) identified two nodaviruses from the pink shrimp (Frafantepenaeus duorarum hepatopancreas), which is a commercially important farmed shrimp species. One virus clone had a 403 nt insert encoding the RNA-dependent-RNA polymerase (RdRp) partial sequence while the other contained a 236 nt insert encoding a capsid protein partial sequence. Phylogenetic analysis showed that the capsid sequence was 43–51% similar to the shrimp nodaviruses, Macrobrachium rosenbergii nodavirus (MrNV) and P. vannamei nodavirus (PvNV), which causes white tail disease in prawns (M. rosenbergii), and muscle necrosis disease in shrimp (Litopenaeus vannamei) (Table 2). Apart from FdNV, Ng et al. (2013) also identified the shrimp hepatopancreas associated circular virus (ShrimpCDV) from the digestive tract of the pink shrimp. Based on these findings, it can be concluded that viral metagenomics can be used to identify novel viral pathogens of shrimps. Once the genomic sequence of the novel virus has been determined, specific molecular biology diagnostic tools such as PCR can be designed and used for the screening of brood stock in order to ensure that only SPF stocks are used for shrimp production. In this way viral metagenomics is expected to play an important role in developing effective disease control strategies for novel viral infections of crustaceans before they cause devastating economic losses in aquaculture.

Apart from fish and crustaceans, the farming of different species of mollusks, turtle, and starfish has also increased tremendously in the last few decades. Interestingly, novel viruses for these aquatic organisms have also been identified using viral metagenomics (Table 3). Ng et al. (2009a) identified a novel sea turtle tornovirus 1 (STTV1) as the cause of fibropapilloma tumors in Florida sea green turtles. Hewson et al. (2014) identified a novel Forbes sea star (Asterias forbes) virus (AFSSV) as the causative agent of a wasting disease condition linked to high mortality in the Forbes sea stars. On the other hand, Andrade et al. (2015) showed that oysters are the hotspots for Mimivirus isolations which are the largest viruses found in the world with >1,100,000 bp length. Overall, these findings show that viral metagenomics analysis is a powerful tool able to identify novel pathogenic viruses infecting different aquatic organisms.

Captive marine mammals such as the Californian sea lions (Z. californianus) and bottle-nose dolphins (Tursiops tuncatus) are widely used as recreation animals. And as such, these animals are kept in marine parks where they are used in circus shows and other activities because of their intelligence and easy trainability. Given their close contact with humans, it has become expedient that their disease profile is determined both for public health reasons and for the sake of providing timely healthcare when they become infected with different diseases. Recently, Ng et al. (2009b) used viral metagenomics analysis to identify a novel California sea lion anellovirus (ZcAV) having 35% amino acid homology in the open reading frame (ORF) 1 with the feline anellovirus. In another study, Ng et al. (2011) used metagenomics analysis to identify a novel seal anellovirus (SealAV) in Pacific harbor seals (Phoca vitulina richardsii). Kluge et al. (2016) carried out a viral metagenomics survey from feces of subantarctic fur seals (Arctocephalus tropicalis) and South America fur seal (Arctocephalus australis). From the South America fur seals they identified seal anellovirus 5 (SaV5), and Troque teno salophus virus 1 (TTZV) while in the Subantarctic seal fur picornaviruses were identified namely Fur seal sakobuvirus (FSSV) having 50% amino acid identity resemblance to the Feline sakobuvirus A (FSVA). In addition, rotaviruses were identified from the subantarctic fur seals with 45–69% amino acid homology with group C rotaviruses. The identification of rotaviruses from Fur seals has important public health implications given that these animals are commonly found on coastal areas of South America where they attract a lot of tourism where they could serve as a source of zoonotic infections to tourists. These findings suggest that viral metagenomics can be used to obtain disease profiles for aquatic animals as well as profiles of zoonotic pathogens found in marine mammals. Table 4 shows some of the aquatic mammalian viruses discovered using viral metagenomics and NGS. It is interesting to note that most of the viruses shown in Table 3 have only been discovered in the last 4 years further consolidating the notion that viral metagenomics is accelerating our ability to identify novel viruses infecting aquatic organisms.

Ng et al. (2009a,b, 2013) used lung tissues to identify the ZcAV infecting Sea lions in two separate mortality events. They (Ng et al., 2009b) showed that ZcAV was mainly found in the lungs and pleural-cavity and not in blood, tonsils, lymph nodes, liver, and other organs suggesting that metagenomics can be used to determine tissue tropism of novel viruses in hosts. Follow-up studies showed that the prevalence was high during outbreaks in captive sea lions (100%) and low in reservoir wild sea lions (11%) indicating viral metagenomics can be used to monitor the prevalence of the virus in marine animals. In follow-up studies, the virus from Sea lion was used to develop an enzyme linked immunosorbent assay (ELISA) and specific PCR (Fahsbender et al., 2015b). Apart from ZcAV, Ng et al. (2011) used infected lung tissues of Pacific harbor seals (P. vitulina richardsii) to identify SealAV. Hewson et al. (2014) used viral metagenomics to identify the Sea star-associated densovirus (SSaDV) as the causative agent of a wasting disease characterized by behavioral changes, lesions on the limbs, loss of tugor and death due to rapid degradation in infected sea-star (asteroid). They observed that increase in viral titers correlated with increase in pathology in the infected sea stars. Ng et al. (2009a) used tumors to identify STTV1 as the causative agent of fibropappillomas in Florida green sea turtles. Bodewes et al. (2015) identified Phocine herpesvirus 7 (PhHV-7) as the etiological agent of ulcerative gingivitis in phocines (Bodewes et al., 2015) while Enhydralutis papilloma virus 1 (EIPV-1) was shown to be the etiological agent of oral tumors diagnosed in Southern otters (Enhydra lutris Nereis) (Ng et al., 2015).

Yang et al. (2011) have pointed out that viral metagenomics has a high chance of identifying co-infections than traditional diagnostic methods. Ng et al. (2009a) showed that STTV1 existed as a co-infection of a quasispecies of variant strains from a single fibropappilloma tumor in each infected turtle. The STTV1 variants detected by viral metagenomics were identical from the majority part of the viral genome while hypervariable regions (HVRs) were extensively divergent. Based on these findings, it can be concluded that viral metagenomics does not only serve as a diagnostic tool for identifying novel viruses, but it serves as a reliable tool for identifying co-infections of different viruses working together to cause disease as well as identification of co-infections of variant strains of the same virus existing as a quasi-species in a single infection.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.