Fig. 1

Cataloging variation in the pangenome graph. A Total lengths of sequence contained in insertions (INS), deletions (DEL), and SNVs, compared between this study (“pangenome”) and read-alignment methods [38, 39]. B Distribution of lengths of insertions and deletions found in this study compared to those found by Zhang et al. [38] using long reads shows that although long-read alignment finds more short insertions (< 1 kb) than the pangenome, the larger cumulative length of insertions found by our pangenome compared to Zhang as shown in A is driven by long insertions (> 1 kb), which have a larger effect on cumulative length. C A hypothetical schematic of how nested variation can evolve: an insertion mutation is followed by a later single-nucleotide mutation, resulting in an insertion relative to the reference that contains a segregating site. A genotype against a linear reference would represent these as three different alleles, whereas a pangenome conserves the nested structure of this variation