Table 1:

Benchmarking SV recall, precision, and F1 on HG002 Human (hg19) ONT reads at varying coverages (10×, 20×, 30×)

Method	Recall, %	Precision, %	F1, %
ONT 10×
minimap2	78.31	75.59	76.93
NGMLR	77.40	76.65	77.02
Vulcan
60%	74.64	88.69	81.06
70%	76.66	87.87	81.88
80%	77.66	85.55	81.42
90%	78.29	83.31	80.72
ONT20×
minimap2	83.55	76.13	79.67
NGMLR	83.39	76.24	79.66
Vulcan
60%	83.78	77.71	80.63
70%	83.91	78.53	81.13
80%	83.50	79.57	81.49
90%	83.55	80.65	82.08
ONT 30×
minimap2	88.74	77.37	82.66
NGMLR	88.47	77.79	82.79
Vulcan
60%	89.37	79.11	83.93
70%	89.36	79.87	84.35
80%	89.24	80.71	84.76
90%	88.81	81.40	84.94

Various percentile cut-offs for Vulcan were used, including 60%, 70%, 80%, 90%. SV calls based on Vulcan mappings achieve the highest F1 score for various cut-off values.