The contigs >5.0 kb were collected, de-replicated and then piped through VirSorter2 (based on sequence similarity and other viral-like features such as GC skew) and VirFinder (based on k-mer signatures) for the identification of viral sequences. The identified viral contigs from VirSorter2 and VirFinder were merged and de-replicated with CD-HIT v4.7 at local identity of 100%. The valid 21,518 viral contigs were subjected to species-level clustering to create viral operational taxonomic units (vOTUs) using the ClusterGenomes scripts, following the MIUViG recommended criteria of 95% average nucleotide identity (ANI) and 85% alignment fraction (AF), resulting in the identification of 13,895 vOTUs.