PPRODO: Prediction of Protein Domain Boundaries Using Neural Networks
Jaehyun Sim, Seung-Yeon Kim, and Jooyoung Lee ( Abstract PDF )
PPRODO
PPRODO predicts domain boundaries of proteins from sequence information by neural networks.
The prediction accuracy of PPRODO is about 67%.
PPRODO source code
INSTALLATION
------------
1. decompress pprodo_1.0.tar.gz
gunzip pprodo_1.0.tar.gz
tar -xvf pprodo_1.0.tar
2. cd pprodo_1.0
3. ./install
4. edit "runpprodo" script file
dbname=pfilt_nr : your database
ncbidir=/usr/local/blast : blast directory
execdir=. : PPRODO installed directory
EXAMPLE USAGE
-------------
1. first network prediction only
$ runpprodo query.fa
query.fa : fasta file format
2. first and second network predictions
$ runpprodo query.fa psipred.ss2
psipred.ss2 : PSIPRED output file
DATA SETS
NR database for PSI-BLAST used in training and evaludation: nr030621.gz (380M)
CONTACT
Send a message to jlee@kias.re.kr or jhsim2@snu.ac.kr