VirusDetect: Efficient and Accurate Discovery of Virus By Deep Sequencing of Small RNAs

Accurate detection and identification of virus infection in plants and animals is critical for agriculture production and human health. Conventional methods such as PCR or microarrays are useful but they require the prior knowledge and sequence information of the potential pathogens, thus they are highly inefficient in detecting novel or emerging viruses. RNA silencing constitutes a fundamental antiviral defense mechanism in both plants and animals in which host enzymes cut and amplify viral RNA into pieces of 20-24 nt. Deep sequencing of these virus-derived small RNAs (sRNAs) and proper assembly or alignment of these sRNA sequences can reconstruct genomic sequence information of the viruses being targeted in the plants and animals. This approach is independent of the ability to culture or purify the virus and does not require any specific amplification or enrichment of viral nucleic acids as it automatically enriches for small RNAs of viral origin by tapping into a natural antiviral defense mechanism.

ViruDetect is a software package that can efficiently and exhaustively analyze large-scale sRNA datasets for virus identification. The program aligns the sRNA reads to the known virus reference sequence database (GenBank gbvrl) as well as assembles the sRNA sequences de novo using Velvet with automatic parameter optimization. The alignment derived contigs and de novo assembly derived contigs are concatenated to remove redundancy and the resulted unique contigs are compared to the reference virus sequences for virus identification.