Skip to main content

Table 2 Comparison of HMM-GRASPx and MetaCLADE against a simulated 100-bp marine data set with uneven coverage

From: A multi-source domain annotation pipeline for quantitative metagenomic and metatranscriptomic functional profiling

Pathway

HMM-GRASPx

MetaCLADE/HMM-GRASPx-Assemblyc

MetaCLADE/GC-Assemblyd

 

TP

FP

FN

Recall

Prec

F-score

TP

FP

FN

Recall

Prec

F-score

TP

FP

FN

Recall

Prec

F-score

“Strict” domain annotationa

 KO00010

159 243

2 216

39 938

79. 9

98.6

88.3

171 168

5 384

24 845

87.3

97.0

91.9

179 972

7 397

14 028

92.8

96.1

94.4

 KO00020

182 706

11 691

74 704

71.0

94.0

80.9

201 808

24 786

42 507

82.6

89.1

85.7

223 671

20 452

24 978

90.0

91.6

90.8

 KO00030

101 449

589

17 761

85.1

99.4

91.7

107 065

833

11 901

90.0

99.2

94.4

113 446

859

5 494

95.4

99.2

97.3

 KO00051

436 432

17 970

192 918

69.3

96.0

80.5

482 994

50 714

113 612

81.0

90.5

85.5

553 728

41 566

52 026

91.4

93.0

92.2

 KO00620

237 743

11 756

100 903

70.2

95.3

80.8

263 011

27 306

60 085

81.4

90.6

85.8

297 162

21 038

32 202

90.2

93.4

91.8

 KO00680

595 304

24 101

229 802

72.1

96.1

82.4

633 158

74 709

141 340

81.8

89.4

85.4

695 625

70 569

83 013

89.3

90.8

90.1

 KO00910

219 501

16 171

86 939

71.6

93.1

81.0

238 683

30 435

53 493

81.7

88.7

85.0

264 438

25 270

32 903

88.9

91.3

90.1

 KO00920

74 851

138

37 662

66.5

99.8

79.8

82 977

1 098

28 576

74.4

98.7

84.8

105 794

603

6 254

94.4

99.4

96.9

“Clan-based” domain annotationb

 KO00010

161 459

0

39 938

80.2

100.0

89.0

176 369

183

24 845

87.7

99.9

93.4

187 308

61

14 028

93.0

100.0

96.4

 KO00020

194 397

0

74 704

72.2

100.0

83.9

226 585

9

42 507

84.2

100.0

91.4

244 114

9

24 978

90.7

100.0

95.1

 KO00030

102 038

0

17 761

85.2

100.0

92.0

107 867

31

11 901

90.1

100.0

94.8

114 275

30

5 494

95.4

100.0

97.6

 KO00051

454 402

0

192 918

70.2

100.0

82.5

533 583

125

113 612

82.4

100.0

90.4

594 999

295

52 026

92.0

100.0

95.8

 KO00620

249 492

7

100 903

71.2

100.0

83.2

288 167

2 150

60 085

82.7

99.3

90.3

317 646

554

32 202

90.8

99.8

95.1

 KO00680

618 689

716

229 802

72.9

99.9

84.3

701 356

6 511

141 340

83.2

99.1

90.5

756 885

9 309

83 013

90.1

98.8

94.3

 KO00910

235 260

412

86 939

73.0

99.8

84.3

267 624

1 494

53 493

83.3

99.4

90.7

288 068

1 640

32 903

89.7

99.4

94.3

 KO00920

74 989

0

37 662

66.6

100.0

79.9

84 053

22

28 576

74.6

100.0

85.5

106 392

5

6 254

94.4

100.0

97.1

  1. aOnly hits having same Pfam domain with respect to the ground-truth are counted as true positives.
  2. bDomain hits that belong to the same clan with respect to the ground-truth are counted as true positives.
  3. cAnnotation obtained by applying MetaCLADE on the assembled contigs of HMM-GRASPx.
  4. dAnnotation obtained by applying MetaCLADE on the gene catalog
  5. Largest values are reported in italics