Fig. 1: The landscape of HBV insertions in 296 HCCs from the PCAWG dataset.

a Canonical (reciprocal) insertions are represented as red diamonds, and non-canonical insertions (single-clusters) as purple and green triangles for positive and negative clusters, respectively. In total, 148 integration events are shown of which 72% represent non-canonical events. b Classical pattern of canonical HBV insertions identified with Illumina paired-end mapping data is characterised by two reciprocal clusters of discordant reads, and clipped reads, in face-to-face orientation, demarcating the boundaries of the genomic integration. The mates of these reads map onto HBV consensus sequences. Clipped reads span the insertion site allowing base-pair resolution of the insertion breakpoints. c Most HBV insertion events in HCC tumours show a non-canonical pattern in which a single cluster of paired-end reads (short-reads in red) demarcates one of the two boundaries of the insertion only, while the second cluster is missing.