The CRVC reveals a striking diversity of unreported viral genus-level clusters and connections with crop root bacterial genomes. The hierarchical clustering dendrogram was built based on the clustered viral protein families in the CRVC using the Jaccard distance with the average clustering model (STAR Methods). The first circle (I) indicates the taxonomy of host bacteria for each viral cluster at the genus level, showing that 2,199 (71.0%) of CRVC clusters exhibit connections with crop root bacteria. Connections were determined according to viral contigs and CRISPR spacers detected in bacterial genomes. The second inner circle represents the proportion of temperate phages within each viral genus (II). Black bars show the log10-transformed genome numbers in each viral genus, ranging from 1 to 144 (III). In the outermost circle (IV), the gene-sharing network analysis reveals 1,572 unreported viral genus-level clusters in the CRVC compared with publicly available databases (STAR Methods).
