Skip to main content

Advertisement

Table 1 List of tools providing good SV calling results for both the simulated and NA12878 real datasets

From: Comprehensive evaluation of structural variation detection algorithms for whole genome sequencing

SV type Tools Simulated data Real data nF*1
Precision Recall Precision Recall
DEL GRIDSS 98.9 (5) 86.6 (2) 87.6 (7) 28.9 (2) 3.57 (1)
Lumpy 99.1 (4) 81.4 (6) 87.1 (8) 26.1 (4) 3.41 (2)
SVseq2 96.2 (11) 86.1 (3) 75.7 (17) 24.9 (5) 3.28 (3)
SoftSV 96.8 (10) 83.6 (4) 80.2 (13) 23.2 (8) 3.25 (7)
Manta 95.9 (12) 83.1 (5) 74.2 (20) 24.3 (6) 3.21 (5)
MATCHCLIP 99.4 (2) 71.7 (10) 91.6 (4) 20.9 (11) 3.12 (6)
inGAP-sv 91.1 (18) 78.6 (7) 78.3 (14) 22.5 (8) 3.10 (7)
DUP Wham 96.9 (4) 81.7 (4) 57.1 (4) 10.2 (5) 3.92 (1)
SoftSV 84.2 (14) 67.8 (13) 47.3 (6) 14.3 (3) 3.91 (2)
MATCHCLIP 87.6 (11) 77.5 (8) 58.0 (3) 9.9 (6) 3.79 (3)
GRIDSS 91.1 (9) 77.9 (7) 58.4 (2) 9.6 (7) 3.78 (4)
Manta 99.0 (1) 83.2 (1) 40.4 (9) 6.5 (11) 3.35 (5)
SvABA 82.6 (15) 69.6 (11) 42.7 (8) 7.2 (9) 3.02 (6)
INS [Unspecified] pbsv 89.7 (3) 38.2 (5) 72.7 (8) 27.5 (2) 6.68 (1)
inGAP-sv 99.7 (1) 58.5 (2) 85.5 (2) 11.8 (3) 6.27 (2)
Sniffles 74.8 (5) 52.5 (3) 65.9 (10) 9.0 (5) 5.08 (3)
SVseq2 70.4 (8) 64.2 (1) 38.5 (19) 7.1 (9) 4.87 (4)
INS [MEI] MELT 99.7 (3) 68.9 (3) 88.9 (1) 85.6 *2 (1) 3.21 (1)
Mobster 100 (1) 67.1 (4) 88.3 (2) 71.9 *2 (2) 3.04 (2)
INV DELLY 94.7 (8) 81.8 (4) 38.9 (4) 15.6 (2) 3.07 (1)
TIDDIT 89.2 (14) 77.9 (8) 49.1 (1) 11.7 (5) 2.89 (2)
1–2-3-SV 70.7 (19) 81.2 (5) 31.8 (9) 14.8 (3) 2.67 (3)
GRIDSS 96.6 (6) 84.7 (3) 34.2 (8) 10.4 (7) 2.67 (4)
  1. *1Sum of normalized F-measures of the simulated and the real data. Normalized F-measure = F-measure/the mean F-measure for the corresponding category
  2. *2Provisional recall value: the number of true positives was calculated by dividing by the provisional number of reference MEIs (1350), which was estimated using the data from the 1000 Genome project
  3. Ranks of tools for each result (precision, recall, or F-measure) are indicated within parentheses