Skip to main content

Table 1 Prediction quality of the FAO/WHO methods.

From: Allermatch™, a webtool for the prediction of potential allergenicity according to current FAO/WHO Codex alimentarius guidelines

   

1

2

3

   

False negatives

False negatives (corrected)

False positives

Database

Method

Wordsize

Number

%

Number

%

Number

%

SwissProt

Window

n.a.

71 / 334

21.3

57 / 320

17.8

3 / 12

25.0

 

Wordmatch

6

54 / 334

16.2

n.a.

n.a.

7 / 12

58.3

  

7

69 / 334

20.7

n.a.

n.a.

6 / 12

50.0

  

8

78 / 334

23.4

n.a.

n.a.

3 / 12

25.0

WHO-IUIS

Window

n.a.

99 / 632

15.7

78 / 611

12.8

4 / 12

33.3

 

Wordmatch

6

58 / 632

9.2

n.a.

n.a.

9 / 12

75.0

  

7

98 / 632

15.5

n.a.

n.a.

8 / 12

66.7

  

8

117 / 632

18.5

n.a.

n.a.

3 / 12

25.0

SwissProt & WHO-IUIS

Window

n.a.

101 / 730

13.8

77 / 706

10.9

5 / 12

41.7

 

Wordmatch

6

55 / 730

7.5

n.a.

n.a.

9 / 12

75.0

  

7

95 / 730

13.0

n.a.

n.a.

8 / 12

66.7

  

8

115 / 730

15.8

n.a.

n.a.

3 / 12

25.0

  1. The number and percentage of false negative and false positive results are shown here for all FAO/WHO recommended method/database combinations. Result set 1 describes the number of false negatives observed in a leave-one-out approach. The next result set (2) shows the same results but corrected for those sequences that were not able to generate a hit against itself due to the short length of the sequence. The last result set (3) shows the observed number of false positives when testing 12 non-allergenic sequences with the Allermatch™ webtool. Each of the result sets consists of two columns; the first column shows the number of erroneous hits and the total number of sequences in this set. The second column shows the percentage of erroneous hits.