Skip to main content

Table 3 Ordered list of discriminative features.

From: Word correlation matrices for protein sequence analysis and remote homology detection

#

Family 1.27.1.1

Family 1.27.1.2

Family 1.36.1.2

Family 1.36.1.5

1

Leu@5, Leu@5

Leu@6, Leu@6

Thr@1, Val@5

Ala@1, Lys@5

2

Leu@6, Leu@6

Leu@5, Leu@5

Thr@2, Val@6

Ala@2, Lys@6

3

Leu@1, Leu@1

Leu@1, Leu@1

Val@1, Ser@2

asp@2, asp@2

4

Leu@2, Leu@2

Leu@2, Leu@2

Val@2, Ser@3

asp@3, asp@3

5

Leu@4, Leu@4

Leu@4, Leu@4

Val@5, Ser@6

asp@1, asp@1

6

Leu@3, Leu@3

Leu@3, Leu@3

Val@4, Ser@5

asp@4, asp@4

7

Leu@1, Leu@5

Leu@1, Leu@5

Val@3, Ser@4

asp@6, asp@6

8

Leu@2, Leu@6

Leu@2, Leu@6

Val@2, Thr@6

asp@5, asp@5

9

Glu@6, Glu@6

Glu@1, Glu@1

Val@1, Thr@5

Ala@1, Leu@2

10

gly@1, gly@1

Glu@2, Glu@2

Ser@1, Thr@4

Ala@2, Leu@3

  1. List of 10 most discriminative features for four superfamilies associated with the SCOP class "All alpha proteins". Features are sorted in descending order according to their absolute discriminative weight (not shown). The first column corresponds to the rank of a feature and the succeeding columns contains the description of the feature in the word correlation feature space in terms of a pair of amino acids (in IUPAC three-letter code) at particular word positions. Features that are associated with negative discriminative weights are printed with lowercase first letters.