Figure 1. NovoGraph overview.

Overview of the genome graph generation pipeline presented here. In Step 1, each genome is aligned to a reference genome (GRCh38, shown here in black). In Step 2.1, the pairwise alignments are partitioned into smaller windows for multiple sequence alignment in Step 2.2. Multiple sequence alignments are concatenated into an approximate global single alignment in Step 2.3. Finally, in Step 3, the multiple sequence alignment is converted to a single graph representation of the genome, shown in gray. Each individual genome has a single, acyclic path through the genome graph (black, green, blue, and orange paths). The magenta path represents a “mosaic” genome—that is, a path through the graph which was not observed in any genome.