Fig. 3

Example structural variations identified in the pangenome graph. A, B Visualisation of alternative divergent paths in the graph. For both selected examples, the left side panel shows a bandage plot indicating the reference HdrR and the alternative path. Each graph segment is color-coded according to the number of samples with at least 50% of the reference coverage for the ONT DNA-Seq dataset, from white (none) to deep red (all). Blue segments are supported by multiple samples for both Illumina RNA-Seq and ONT DNA-Seq. The right panel shows the linear structure of local assemblies for both the reference and the alternative paths. For the reference, the top blue track represents the existing Medaka HdrR annotations. The light and dark green tracks correspond to the segment layout from the graph. Finally, the heatmaps show the RNA expression intensities for all 50 medaka samples sequenced along the represented sections of the graph (grey = not found, white = less than 5 reads, dark red = more than 100 reads). C, D Visualisation of large-scale deletions in the graph. For both selected examples, the left panel shows the Medaka HdrR annotations (blue) and the graph segment layout (light and dark green), overlaid with the deletion position (grey rectangle). The bandage plots on the right are color-coded as previously described. The shaded area indicates the reference sequence deletions robustly supported by a direct connection between distant reference segments (link coverage > 50% of reference coverage for at least 9 samples)