Fig. 3

Genus-level profiles and global distribution. A, C Genus-level profiles of CB and SB/SW vOTUs, respectively. The genus-level profiles were constructed by the genome-content-based network analysis of vOTUs from this study, viral genomes from the NCBI RefSeq database, and vOTUs from 250 environmental metagenomes in the following ecosystems: global oceans (GOV2 dataset), deep ocean water, deep ocean sediment, surface layers of permafrost (IsoGenie), soil, air, glacier cryoconite, glacier ice core, and lake water (see “Methods”). SB and SW viruses were combined for the analysis (as SB/SW). B, D Environmental associations of CB and SB/SW vOTUs, respectively. For the CB (B) and SB/SW (D) vOTUs that clustered (i.e., shared VCs/genera) with viruses from environmental metagenomes, the distribution of environmental habitats was evaluated and illustrated by a heatmap. The gradient colors in the heatmap represent the percentage of vOTUs (weighted by the associated vOTUs) that were associated with the different ecosystems. Each cell in the last column (i.e., labeled as “Total”) indicates the total percentage of vOTUs detected from each of the tested environments on the right-side legend, while each cell in the other columns (i.e., the 12 columns in the left rectangle) indicates the percentage of vOTUs that were detected from both of the two environments suggested at the top- and right-side legend of the heatmap. For example, the cells associated with both Arctic and Deep ocean water showed that 10–20% (B) and 5–10% (D) of the 52 CB and 2628 SB/SW vOTUs, respectively, were detected from both of the Arctic and Deep ocean water environments. Most of the CB and SB/SW viruses were associated with seawater viruses from the Arctic