Abstract
Nucleocytoplasmic DNA viruses (NCLDVs) are highly diverse and abundant in marine environments. However, knowledge of their hosts is limited because only a few NCLDVs have been isolated so far. Taking advantage of the recent large-scale marine metagenomics census, in silico host prediction approaches are expected to fill the gap and further expand our knowledge of virus–host relationships for unknown NCLDVs. In this study, we built co-occurrence networks of NCLDVs and eukaryotic taxa to predict virus–host interactions using Tara Oceans sequencing data. Using the positive likelihood ratio to assess the performance of host prediction for NCLDVs, we benchmarked several co-occurrence approaches and demonstrated an increase in the odds ratio of predicting true positive relationships four-fold compared with random host predictions. To further refine host predictions from high-dimensional co-occurrence networks, we developed a phylogeny-informed filtering method, Taxon Interaction Mapper, and showed it further improved the prediction performance by twelve-fold. Finally, we inferred virophage – NCLDV networks to corroborate that co-occurrence approaches are effective for predicting interacting partners of NCLDVs in marine environments.
Importance NCLDVs can infect a wide range of eukaryotes although their life cycle is less dependent on hosts compared with other viruses. However, our understanding of NCLDV– host systems is highly limited because few of these viruses have been isolated so far. Co-occurrence information has been assumed to be useful to predict virus–host interactions. In this study, we quantitatively show the effectiveness of co-occurrence inference for NCLDV host prediction. We also improve the prediction performance with a phylogeny-guided method, which leads to a concise list of candidate host lineages for three NCLDV families. Our results underpin the usage of co-occurrence approach for metagenomic exploration of the ecology of this diverse group of viruses.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
Author list updated; Supplemental files updated