Skip to main content

Table 3 Combined runtime and accuracy results for the simulated low-complexity metagenome datasets.

From: ILP-based maximum likelihood genome scaffolding

METHOD

FRAC STAPH

RUNTIME

SCFN50

ALN50

TPN50

MCC

SILP2 0

1.00

14.3

51,775.0

20,647

34,495

67.6

SILP2 0

0.50

13.3

50,450.0

20,103

36,356

69.1

SILP2 0

0.25

12.7

47,731.0

20,761

35,323

69.3

SILP2 0

0.00

11.5

21,753.0

13,948

15,649

39.9

SILP2 1

1.00

14.3

52,557.0

20,655

33,683

67.5

SILP2 1

0.50

13.8

48,701.0

20,337

35,750

69.0

SILP2 1

0.25

13.4

49,766.0

20,752

35,146

69.0

SILP2 1

0.00

11.0

21,925.0

13,847

15,511

39.3

SILP2 2

1.00

14.3

43,144.0

20,631

31,160

66.3

SILP2 2

0.50

13.5

42,244.0

20,198

32,477

67.5

SILP2 2

0.25

13.2

45,137.0

21,161

31,562

67.6

SILP2 2

0.00

10.8

22,190.0

13,813

16,205

41.6

SILP2 3

1.00

14.1

43,646.0

19,998

28,856

65.3

SILP2 3

0.50

13.2

41,893.0

19,790

30,504

66.6

SILP2 3

0.25

13.0

42,188.0

19,945

30,449

66.4

SILP2 3

0.00

11.5

21,820.0

13,781

15,635

40.0

OPERA

1.00

2247.2

15,573.0

13,082

10,386

10.1

OPERA

0.50

1567.6

13,928.0

12,006

10,440

10.7

OPERA

0.25

884.0

14,786.0

12,617

10,507

10.5

OPERA

0.00

544.3

11,121.0

10,720

10,273

4.9

MIP

1.00

129.9

20,104.0

12,861

18,672

18.4

MIP

0.50

121.3

19,807.0

12,488

17,613

17.4

MIP

0.25

114.0

18,520.0

12,269

16,680

17.2

MIP

0.00

114.1

12,690.0

10,894

12,434

8.7

BAMBUS2

1.00

1025.89

11,251.0

11,238

-

-

BAMBUS2

0.50

1452.75

10,781.0

10,822

-

-

BAMBUS2

0.25

1676.75

10,806.0

10,834

-

-

BAMBUS2

0.00

2272

11,526.0

11,698

-

-

  1. The second column indicates the percentage of total read pairs used from the staph genome testcase, all of the rhodo pairs were used. SCFN50 is the uncorrected N50 reported by each scaffolding tool. The N50 of the contigs alone is 10,339 bp. The integer appended to SILP2 indicates the bundle weight; 0: none, 1: coverage, 2: repeat, 3:coverage * repeat. All methods were run at bundle size 1, the reported number is the average of 10 runs except for OPERA where 6 of the test cases exceeded runtime limits. TPN50 and MCC were unable to be computed for BAMBUS2 because the generated AGP had a non-standard format.