New approaches indicate constant viral diversity despite shifts in assemblage structure in an Australian hypersaline...

It is widely stated that viruses represent the most significant source of biodiversity on Earth, yet characterizing the diversity of viral assemblages in natural systems remains difficult. Viral diversity studies are challenging because viruses lack universally present, phylogenetically informative genes. Here, we developed an approach to estimate viral diversity using a series of functional and novel conserved genes. This approach provides direct estimates of viral assemblage diversity while retaining resolution at the level of individual viral populations in a natural system. We characterized viral assemblages in eight samples from hypersaline Lake Tyrrell (LT), Victoria, Australia, using 39,636 viral contigs. We defined viral operational taxonomic units (OTUs) in two ways. First, we used genes with three different functional predictions that were abundantly represented in the data set. Second, we clustered proteins of unknown function based on sequence similarity, and we chose genes represented by three clusters with numerous members to define OTUs. In combination, diversity metrics indicated between 412 and 735 sampled populations, and the number of populations remained relatively constant across samples. We determined the relative representation of each viral OTU in each sample and found that viral assemblage structures correlate with salinity and solution chemistry. LT viral assemblages were near-replicates from the same site sampled a few days apart but differed significantly on other spatial and temporal scales. The OTU definition approach proposed here paves the way for metagenomics-based analyses of viral assemblages using ecological models previously applied to bacteria and archaea.