Finding and identifying the viral needle in the metagenomic haystack: trends and challenges

Collectively, viruses have the greatest genetic diversity on Earth, occupy extremely varied niches and are likely able to infect all living organisms. Viral infections are an important issue for human health and cause considerable economic losses when agriculturally important crops or husbandry animals are infected. The advent of metagenomics has provided a precious tool to study viruses by sampling them in natural environments and identifying the genomic composition of a sample. However, reaching a clear recognition and taxonomic assignment of the identified viruses has been hampered by the computational difficulty of these problems. In this perspective paper we examine the trends in current research for the identification of viral sequences in a metagenomic sample, pinpoint the intrinsic computational difficulties for the identification of novel viral sequences within metagenomic samples, and suggest possible avenues to overcome them.