|
This web page provides datasets associated
with
Chen K, Kurgan LA, 2007. PFRES: Protein fold classification
by using evolutionary information and predicted secondary structure.
Bioinformatics, 23(21):2843-2850
The supplementary files include dataset of 908 sequence that was used to test the PFRES method.
Download dataset
The dataset includes the following information:
1. PDB id
2. primary sequence using single letter encoding of amino acids
3. protein fold (SCOP fold label)
The fold names are encoded as follows (label in the dataset / name of fold in SCOP / ID in SCOP):
a Globin-like (a.1.*.*)
b Cytochrome c (a.3.*.*)
c DNA/RNA-binding 3-helical bundle (a.4.*.*)
d Four-helical up-and-down bundle (a.24.*.*)
e 4-helical cytokines (a.26.*.*)
f EF Hand-like (a.39.*.*)
g Immunoglobulin-like beta-sandwich (b.1.*.*)
h Cupredoxin-like (b.6.*.*)
i viral coat and capsid proteins (b.121.*.*)
j Concanavalin A-like lectins/glucanases (b.29.*.*)
k SH3-like barrel (b.34.*.*)
l OB-fold (b.40.*.*)
m beta-Trefoil (b.42.*.*)
n Trypsin-like serine proteases (b.47.*.*)
o Lipocalins (b.60.*.*)
p TIM beta/alpha-barrel (c.1.*.*)
q FAD/NAD (P)-binding domain (c.3.*.*)
r Flavodoxin-like (c.23.*.*)
s NAD (P)-binding Rossmann-fold domains (c.2.*.*)
t P-loop containing nucleoside triphosphate hydrolases (c.37.*.*)
u Thioredoxin fold (c.47.*.*)
v Ribonuclease H-like motif (c.55.*.*)
w alpha/beta-Hydrolases (c.69.*.*)
x Periplasmic binding protein-like I (c.93.*.*)
y beta-Grasp (d.15.*.*)
z Ferredoxin-like (d.58.*.*)
- small inhibitors toxins lectins (g.3.*.*)
The primary sequences include X, which denotes a "special" amino acid:
- a residue originally annotated as not one of the 20 amino acids
- or a residue that constitutes a connection between two distinct
segments in one sequence; for example in a.24.25.1,
{A:118-211,A:372-450} is a domain that consists of two segments and X
annotates the connection.
X should be ignored while performing predictions and preparing input data (this assumption was followed when evaluating PFRES).
|