About triplet-SVM classifier

This program is developed for predicting a query sequence with hairpin structure as a real miRNA precursor or not. The triplet-SVM classifier analyzes the triplet elements of the query and predicts it using a SVM classifier. The SVM classifier is previously trained based on the triplet element features of a set of real miRNA precursors and a set of pseudo-miRNA hairpins.

triplet-SVM classifier was written by Chenghai Xue and is free to all.

 

Install and Use

1. Download the package to anywhere in your computer
2. Uncompress the file: tar -zxvf triplet-svm-classifier.tar.gz
3. Read Readme.txt before running the programs

triplet-SVM classifier runs directly on Linux with Perl compiler.

Note: this package need the third-party softwares, namely RNAfold and Libsvm packages.
RNAfold can be downloaded from http://www.tbi.univie.ac.at/~ivo/RNA/
Libsvm can be downloaded from http://www.csie.ntu.edu.tw/~cjlin/libsvm/oldfiles/, the 2.36 version is used in our work and we recommend this version. (The homepage of Libsvm is http://www.csie.ntu.edu.tw/~cjlin/libsvm/).
And both softwares should be compiled in local PC.

 

Download

The package is free for all users but without any warranty.

triplet-SVM classifier contains all source codes for predicting candidate hairpins.

Material contains all results in our submitted paper.

 

Citation

If you use triplet-SVM classifier, please cite us in the following way:

Chenghai Xue, Fei Li, Tao He, Guoping Liu, Yanda Li, Xuegong Zhang, Classification of real and pseudo microRNA precursors using local structure-sequence features and support vector machine, BMC Bioinformatics, 6: 310, 2005

Please contact mailto:chenghai.xue@mail.ia.ac.cn for any problem.