PRED-CLASS: Classification of proteins into one of four possible classes

Version 1.0

Classification of proteins into one of four possible classes

Pasquier, C., Promponas, V. and Hamodrakas, S.J.
Department of Cell Biology and Biophysics,
Faculty of Biology, University of Athens
Panepistimiopolis, Kouponia, Athens 15701, Greece

PRED-CLASS is a system of cascading neural networks that classifies any protein, given its aminoacid sequence alone, into one of four possible classes:
the membrane protein class,

the globular protein class,

the fibrous protein class,

the mixed (fibrous and globular) protein class and

The results obtained on several test sets are summarized in the table below :

type of proteins number of sequences percentage classified in...

the membrane protein class the fibrous protein class the globular protein class the mixed protein class

membrane
147
139 (94.5%)
0 (0.0%)
8 (5.4%)
0 (0.0%)

fibrous
73
1 (1.4%)
72 (98.6%)
0 (0.0%)
0 (0.0%)

globular
55
0 (0.0%)
1 (1.8%)
54 (98.2%)
0 (0.0%)

mixed
112
3 (2.7%)
3 (2.7%)
0 (0.0%)
106 (94.6%)

A paper describing the neural network system has been submitted to Proteins

Now, you can :

Run PRED-CLASS on a sequence

Browse detailed results obtained with the algorithm on several test sets

View lists of data used to train and evaluate PRED-CLASS

Go to the Biophysics Lab Homepage

Claude Pasquier

type of proteins	number of sequences	percentage classified in...
type of proteins	number of sequences	the membrane protein class	the fibrous protein class	the globular protein class	the mixed protein class
membrane	147	139 (94.5%)	0 (0.0%)	8 (5.4%)	0 (0.0%)
fibrous	73	1 (1.4%)	72 (98.6%)	0 (0.0%)	0 (0.0%)
globular	55	0 (0.0%)	1 (1.8%)	54 (98.2%)	0 (0.0%)
mixed	112	3 (2.7%)	3 (2.7%)	0 (0.0%)	106 (94.6%)

Classification of proteins into one of four possible classes

PRED-CLASS is a system of cascading neural networks that classifies any protein, given its aminoacid sequence alone, into one of four possible classes: the membrane protein class, the globular protein class, the fibrous protein class, the mixed (fibrous and globular) protein class and

The results obtained on several test sets are summarized in the table below :

A paper describing the neural network system has been submitted to Proteins

Now, you can :

Run PRED-CLASS on a sequence

Browse detailed results obtained with the algorithm on several test sets

View lists of data used to train and evaluate PRED-CLASS

Go to the Biophysics Lab Homepage

PRED-CLASS is a system of cascading neural networks that classifies any protein, given its aminoacid sequence alone, into one of four possible classes:
the membrane protein class,

the globular protein class,

the fibrous protein class,

the mixed (fibrous and globular) protein class and