Figure 2:

Overall edit distance improvements. A: Normalized edit distance comparison of Vulcan’s (green) 90% percentile cut-off, NGMLR (red) and minimap2’s (blue) mapping result with human ONT 30X reads. We can see clear evidence that the realignment of only 10% of the reads lead to an improvement in edit distance and thus of the variant calling. B: Distribution of mappings’ normalized edit distance from Vulcan (green), NGMLR (red) and minimap2 (blue). Vulcan has a lower edit distance mapping than minimap2 with NGMLR’s refinement.