Fig. 2: Diversity and major types of DGR targets.

a Prevalence and sequence characteristics of the most abundant DGR target protein clusters (PCs). The 24 PCs listed here represent >92% of all identified DGR targets. These were divided into four major types (left panel; predicted localization of target proteins highlighted with red stars). Characteristics of each PC are indicated to the right, including a number of associated DGR RT OTUs, the relative proportion of viral-encoded vs. cellular-encoded DGRs (“Genome type”), distribution across DGR clades, percentage of PC member with a detected C-Lec fold around the VR region, percentage of PC members predicted as a tail structural protein (for viral targets) or membrane protein (for cellular targets), and relative position of the VR region within the target sequence. The boxplot lower and upper hinges correspond to the first and third quartiles, respectively, and the whiskers extend no further than ±1.5 times the interquartile range. For C-Lec fold VR and localization prediction data, only high-quality targets were considered (see “Methods”). For mixed PCs (PC_00009 and PC_00021), targets with an unknown origin were excluded. b Estimated host diversity for viral genomes encoding (DGR+, n = 822) or lacking DGRs (DGR−, n = 1182) matching at least 50 CRISPR spacers. The vertical axis shows the number of connected host species per 50 protospacers. The boxplot lower and upper hinges correspond to the first and third quartiles, respectively, and the whiskers extend no further than ±1.5 times the interquartile range. c Percentage of cellular targets with a predicted transmembrane domain (top) or one or more functional domain(s) associated with cell adhesion and/or carbohydrate-binding identified outside of the VR region (bottom). Target sequences are divided based on their PC membership into “membrane-bound” PCs or “other” PCs (see panel (a)).