Skip to main content

Table 4 Preference of structural words in W set≥30 according to the loop types as assessed by the KLD criterion and the associated loop coverage rate (on a per structural letter basis).

From: Mining protein loops using a structural alphabet and statistical exceptionality

Loop words specificity

W set≥307

UR w 7

NS w 7

OR w 7

all loops4

short loops5

long loops6

Long-loop-specific words

758 (22.9%)

23

475

260

23.2%

12.2%

33.9%

Short-loop-specific words

476 (14.4%)

23

220

233

25.7%

33.0 %

18.6%

Shared words 1

2076 (63.7%)

120

1519

437

45.9%

39.4%

56.3%

Flanking-region-specific words2

1879 (57.1%)

102

1131

646

58.6%

58.9%

58.4%

Flanking-region-unspecific words

1431 (43.2%)

64

1083

284

31.4%

21.7%

40.8%

Loop-type-specific words3

2543 (78.8%)

124

1605

814

66.3%

64.3%

68.2%

Loop-type-unspecific words

767 (23.2%)

42

609

116

16.6%

12.7%

20.5%

  1. 1: words shared by long and short loops 2: description by the four possible flanking-types (αα, αβ, βα, ββ loops), 3: description by the four flanking-types and the two length ranges (αα s , αβ s , αβ s , ββ s for short loops and αα l , αβ l , αβ l , ββ l for long loops). 4: all-loop coverage rate (on a per structural letter basis) 5: short-loop coverage rate (on a per structural letter basis) 6: long-loop coverage rate (on a per structural letter basis) 7: Number of words