Editorial board editor-in-chief

Database
In experiments described in this paper two data sets derived 
from SCOP (Structural Classiication of Proteins) database are 
used. The detailed description of these sets can be found in 
Ding and Dubchak [5]. The training set consists of 313 protein 
sequences and the testing set consists of 385 protein sequences. 
These data sets include proteins from 27 most populated different 
classes (protein folds) representing all major structural classes: 
a
,
b
, 
a
/
b
, and 
a
+ 
b
. Where 
a
are those whose structure is 
essentially formed by 
a
-helices, 
b
are those whose structure is 
essentially formed by 
b
-sheets, 
a
/
b
are those with 
a
-helices and 
b
-strands and 
a
+ b are those in which 
a
-helices and 
b
-strands 
are largely segrega-ted.
The training set was based on PDB_select sets (Hobohm 
et al. [18], Hobohm and Sander [19]) where two proteins have 
no more than 35% of the sequence identity. The testing set was 
based on PDB-40D set developed by Lo Conte et al. [8] from 
which representatives of the same 27 largest folds are selected. 
The proteins that had higher than 35% identity with the proteins 
of the training set are removed from the testing set. 
Table 1. 
The protein folds used in experiments
Fold name
Structural class
Fold index
Number of proteins in
training set
testing set
Globin-like
a
1
13
6
Cytochrome c
a
7
7
9
Dna-binding 3-helical bundle
a
4
12
20
4-Helical up-and-down bundle
a
7
7
8
4-Helical cytokines
a
9
9
9
Alpha; ef-hand
a
11
7
9
Immunoglobulin-like
b
-sandwich
b
20
30
44
Cupredoxins
b
23
9
12
Viral coat and capsid proteins
b
26
16
12
Cona-like lectins/glucanases
b
30
7
6
Sh-3 like barrel
b
31
8
8
Ob-fold
b
32
13
19
Trefoil
b
33
8
4
Trypsin-like serine proteases
b
35
9
4
Lipocalins
b
39
9
7
(Tim)-barrel
a / b
46
29
48
Fad (also nad)-binding motif
a / b
47
11
12
Flavodoxin like
a / b
48
11
13
Nad(p)-binding rossman fold
a / b
51
13
27
P-loop containing nucleotide
a / b
54
10
12
Thioredoxin-like
a / b
57
9
8
Ribonuclease h-like motif
a / b
59
10
14
Hydrolases
a / b
62
11
7
Periplasmic binding protein-like
a / b
69
11
4
฀
-Grasp
a + b
72
7
8
Ferredoxin-like
a + b
87
13
27
Small inhibitors, toxins, lectins
a + b
110
14
27
Total
313
385