Skip to main content

Table 3 Performance of detection methods on datasets HG00514, HG00733, and NA19240

From: cnnLSV: detecting structural variants by encoding long-read alignment information and convolutional neural network

Type

Method

HG00514

HG00733

NA19240

Pre (%)

Rec (%)

F1 (%)

Pre (%)

Rec (%)

F1 (%)

Pre (%)

Rec (%)

F1 (%)

INS/DUP

Sniffles

43.278

60.934

50.610

34.907

61.731

44.596

42.219

60.949

49.883

PBSV

50.588

53.652

52.075

50.254

53.960

52.041

51.152

52.469

51.802

SVIM

44.559

61.158

51.555

42.574

62.097

50.515

45.762

61.180

52.359

cuteSV

50.980

66.202

57.602

51.610

68.674

58.932

51.963

63.710

57.240

cnnLSV

57.539

59.758

58.628

59.114

59.100

59.107

60.201

60.569

60.384

DEL

Sniffles

55.628

64.308

59.654

54.320

64.977

59.173

57.637

65.192

61.182

PBSV

63.873

62.507

63.182

63.161

63.064

63.112

65.006

61.983

63.459

SVIM

53.081

65.714

58.726

51.097

66.439

57.766

55.699

66.379

60.572

cuteSV

54.122

67.872

60.222

51.353

69.149

58.937

55.280

68.011

60.988

cnnLSV

63.460

64.843

64.144

62.813

65.416

64.088

65.299

64.797

65.047

INV

Sniffles

9.091

7.009

7.916

6.278

7.273

6.739

9.420

6.667

7.808

PBSV

32.692

9.346

14.536

26.923

7.273

11.452

32.653

9.333

14.517

SVIM

10.000

10.280

10.138

6.020

9.545

7.384

9.548

10.667

10.076

cuteSV

27.907

6.542

10.599

25.000

6.364

10.145

30.952

6.667

10.970

cnnLSV

36.170

9.813

15.438

27.451

7.273

11.499

31.481

9.778

14.921

All

Sniffles

47.653

62.141

53.941

40.856

62.839

49.517

47.844

62.647

54.254

PBSV

55.814

57.454

56.623

55.239

57.768

56.475

56.924

56.746

56.835

SVIM

52.143

66.584

58.486

45.421

63.713

53.034

49.536

63.356

55.600

cuteSV

47.653

62.931

54.236

51.476

68.477

58.772

53.306

65.419

58.745

cnnLSV

59.889

61.781

60.821

60.562

61.613

61.083

62.371

62.329

62.350

  1. The values in bold represent the best results