Comparison of the 16 unassembled human gut viromes and the assembled dataset based on their tetranucleotide compositions. The NMDS was generated from the pairwise distances computed from the tetranucleotide frequency bias. Each virome is named according to the subject (H1, H2, L1, L2, L3) and day of sampling (day 1, 2, 7 or 8). Samples taken from the same individual are highlighted in shades of blue, yellow and red. Highlighted in green are both the control dataset (X in Figure 1) and the assembled virome described in .