Table 2.

SD coverage of the human genome (hg19) as reported by different tools

Tool	Covers	Misses	Extra	Time (h: m)
WGAC (gold standard)	159.5	0.0	0.0	weeks
SEDEF	218.8	0.6	60.0	0:36
Minimap2	53.3	107.3	1.1	1:30
MUMmer/nucmer	142.6	30.8	13.9	≥20:00
SDDetector	30.1	130.8	1.5	≥1:00a

Note: Misses and Extra are calculated with respect to the WGAC SD calls, which are currently the gold standard of SD calls. Note that we have filtered out all calls where at least one mate is composed solely of common short repeats (Minimap2, MUMmer/nucmer and SDDetector) as we did on SEDEF. All running times were adjusted for 20 CPU cores (all tools which support parallelization were run on 20 cores).

^aAdjusted running time for 20 cores; in reality, SDDetector spends $\ge 8$ hours in the single threaded pre-processing stage. Furthermore, the reported running time only includes post-processing and does not include initial BLAST alignment calculations.