Experimental Model and Subject Details

Between June 2013 and December 2013, 41 samples were collected at different depths from 20 different sites near or within the Arctic Ocean (see full list of samples in Table S3). Physicochemical measurements, sample collection, and DNA extractions were performed using the methods described in (Roux et al., 2016). Extracted DNA was prepared for sequencing using library preparation method described in (Alberti et al., 2017) for viral samples collected during the TOPC campaign (section 4.2) and sequenced using the HiSeq 2000 system (101 bp, paired end reads). Importantly, our sample collection and library preparation methods have known bias towards <0.2um dsDNA viruses (Roux et al., 2017). The TOPC samples were combined with the previously published viromes in (Brum et al., 2015; Roux et al., 2016). Of the previously published dataset, the mesopelagic samples at (Tara stations 37, 39, 56, 68, 70, 76, 78, 111, 122, 137, 138) and the Southern Ocean samples (Tara stations 82_DCM, 84, 85) were sequenced deeper. These combined samples comprise the GOV 2.0 dataset. The number of reads found in each sample can be found in Table S3.