Skip to main content

Table 1 Khmer performance in error trimming and digital normalization using MQF and CMS

From: MQF and buffered MQF: quotient filters for efficient storage of k-mers with their counts and metadata

FPR

Memory in GB

Error trimming

Digital normalization

Error Bound in CMS

Hash func. In CMS

  

Time in min

Missed reads with errors

Time in min

Reads kept by error

  
  

MQF

CMS

MQF

CMS

MQF

CMS

MQF

CMS

  

10–1

1.8

42

39

11,011

445,817a

39

37

3253

31,143a

13,11

3

10–2

2.6

43

48

1304

404,354

41

45

416

24,987

14

5

10–3

3.4

44

61

130

311,464

42

54

58

21,000

15

7

10–4

4.5

44

75

3

292,746

42

68

4

18,449

16

10

Exact

5

45

0

43

0

  1. aPercentages of wrong decisions made by CMS at FPR = 0.1 in error trimming and digital normalization are 0.8% and 0.13% of the total number of decisions versus 0.02% and 0.01% made by MQF