Quick and accurate identification of microbial pathogens is essential for both diagnosis and response to emerging infectious diseases. High-throughput sequencing has recently emerged as a powerful approach to identify both known and novel viruses in clinical specimens. For example, high throughput Sanger sequencing combined with bioinformatic analysis lead to the discovery of human bocavirus [1
], KI polyomavirus [2
] and WU polyomavirus [3
]. With the advent of next-generation sequencing (NGS) technology, the increase in sequencing capability has greatly facilitated efforts to identify viruses in clinical specimens. Many novel viruses have been detected in human specimens, such as Merkel polyomavirus [4
], astrovirus VA1 [5
] and Lujo virus [6
]. Currently, the recognition that sequences are derived from a novel virus is dependent primarily upon detecting sequence similarity between a given read and known reference sequences in various databases. Because longer nucleotide sequence reads can be translated into longer amino acid sequences, they provide a higher probability of detecting divergent sequences distantly related to known viruses. In the past few years, the majority of viruses identified by Next Generation sequencing have utilized Roche/454 platform due to the longer read length compared to Illumina or SOLiD platforms. NGS instruments are now available in most research institutions and the dramatic reduction in sequencing costs has made experimental NGS broadly accessible. However the computational analysis required to identify viral sequences within the massive volumes of data generated by NGS is a barrier for many researchers. To address this limitation, we have developed a robust computational pipeline, VirusHunter, for analysis of NGS data that is suitable for detection of both novel and known virus sequences. Beta versions of VirusHunter have been used to identify both known and novel viruses from a wide array of specimen types [5
]. VirusHunter is freely available from http://www.ibridgenetwork.org/wustl/virushunter .
The World Reference Center for Emerging Viruses and Arboviruses (WRCEVA) maintains a large collection of virus isolates and provides reagents and support for investigations of virus outbreaks throughout the world [14
]. This NIAID-funded program also identifies and characterizes arboviruses (arthropod-borne viruses) and other suspected emerging viruses. The repository within the WRCEVA stores over 6000 classified arthropod-borne, rodent-borne, and other zoonotic and human viruses. In addition there are many virus samples in the collection that have not been well characterized or classified.
Salanga virus (strain AnB 904a) was originally isolated from a rodent (Aethomys medicatus
) collected in 1971 in Salanga, Central African Republic [15
] by researchers at the Institute Pasteur, Bangui, CAR. Initial analyses determined that it was chloroform-sensitive, but it did not react with any of their reagents to known African arboviruses. The virus was then sent to the Institute Pasteur, Dakar, Senegal and subsequently sent to the WRCEVA for further characterization. An early ultrastructural study by El Mekki et al. suggested that Salanga AnB 904a was a Poxvirus [16
]. Currently there are two listings of Salanga virus in the International Committee on Taxonomy of Viruses (ICTV) database. One is an unassigned species in the Bunyaviridae
(acronym SGAV) while the other is an unassigned species in the Poxviridae
(acronym SGV) [17
]. However, since no genome sequence is available; no definitive conclusion can be made about the taxonomic classification of Salanga virus and whether there may be two distinct viruses that currently share the same name.
KY-663 virus was originally isolated on July 18th
, 1965 from the blood of a bat (Myotis macrodactylus
) collected in a mine at Heramatsu, Kagoshima, Japan. The initial isolation was made by intracerebral inoculation of newborn mice. KY-663 virus was found to be partially resistant to treatment with sodium deoxycholate and did not produce a hemagglutinin. Intracranial inoculation of KY-663 into suckling mice causes lethality in 4-5 days, but it is not pathogenic to adult mice [18
]. After efforts to identify the virus in Japan were unsuccessful, it was sent to the WRCEVA for further study, but the virus remained unclassified.
In order to identify and taxonomically classify these two viruses, we used NGS technology to obtain the genomic sequences of those viruses and our customized bioinformatics pipeline, VirusHunter, to analyze the sequences. Furthermore, traditional virologic methods, including electron microscopy and serology, were also employed to provide additional modes of characterization of these viruses. Based on these analyses, Salanga virus was classified as a novel member of the Phlebovirus genus in the Bunyaviridae while KY-663 was determined to be a novel member of the Orbivirus genus in the Reoviridae. We tentatively named KY-663 Heramatsu virus (HERMV) in this manuscript.