Skip to main content

Table 3 Comparison of running time and number of tag SNPs selected when multi-marker LD are used.

From: FastTagger: an efficient algorithm for genome-wide tag SNP selection using multi-marker linkage disequilibrium

 

max_size

min_r 2

Running time (minutes)

#tag SNPs

   

Fast-COOC

MMTagger

Fast-1vsR

MultiTag

Fast-COOC

MMTagger

Fast-1vsR

MultiTag

ENCODE CEU

2

0.95

0.038

0.041

0.048

≥10 hours

1282

1282

1291

1371

ENCODE HCB

2

0.95

0.032

0.032

0.042

≥10 hours

1305

1328

1308

1424

ENCODE JPT

2

0.95

0.029

0.028

0.038

≥10 hours

1234

1258

1240

1349

ENCODE YRI

2

0.95

0.181

0.188

0.245

≥60 hours

2575

2618

2579

2770

chr1

2

0.95

1.13

5.84

1.40

≥7 days

43202

43483

43306

43462

chr2

2

0.95

1.32

7.21

1.63

≥7 days

44135

44556

44225

49289

chr3

2

0.95

1.14

5.11

1.41

≥7 days

37881

38206

37952

39300

chr19

2

0.95

0.176

0.343

0.218

≥30 hours

11151

11192

11160

11747

chr21

2

0.95

0.287

0.473

0.359

≥60 hours

8543

8627

8564

9103

chr22

2

0.95

0.370

0.567

0.468

≥100 hours

8970

9025

8993

9533

ENCODE CEU

3

0.95

1.28

3.69

1.85

≥50 hours

972

1017

1151

1244

ENCODE HCB

3

0.95

1.26

3.40

1.93

≥80 hours

1003

1034

1170

1170

ENCODE JPT

3

0.95

1.06

2.74

1.60

≥50 hours

958

1002

1129

1244

ENCODE YRI

3

0.95

11.6

36.7

17.4

≥14 days

1848

1927

2165

2516

chr1

3

0.95

34.9

137.3

49.6

-

35556

38185

40534

-

chr2

3

0.95

42.9

166.9

60.8

-

35502

38372

41129

-

chr3

3

0.95

39.3

154.6

55.5

-

30695

33041

35305

-

chr19

3

0.95

4.34

16.6

6.25

-

9444

10032

10546

-

chr21

3

0.95

9.91

37.7

14.4

-

6929

7404

7935

-

chr22

3

0.95

16.5

65.3

24.4

-

7327

7788

8392

-

  1. Fast-COOC represents the FastTagger algorithm using the co-occurrence model, and Fast-1vsR represents the FastTagger algorithm using the one-vs-the-rest model. max_size is the maximum number of SNPs on the left hand side of a tagging rule. For the MMTagger algorithm, we divided chr1, chr2 and chr3 into 10 chunks when max_size = 3, and ran MMTagger on each chunk, and then combined the results. For the MultiTag algorithm, we divided chr1, chr2 and chr3 into 20 chunks, chr19, chr21 and chr22 into 5 chunks when max_size = 3. When max_size = 3, MultiTag took too long to finish on the 6 chromosomes, so we did not get its results on the 6 chromosomes.