GeneticEnsemble

Description

Genetic Ensemble is an Java implementation of a genetic algorithm-based ensemble of classifiers designed for feature subset selection https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-11-524. It is a wrapper-based approach that select a subset of features based of their overall performance across a set of classifiers.

This algorithm was originally developed for selecting a panel of interacting SNPs from genome-wide association studies (GWAS), hence the name "GEsnpxPara" where "Para" stands for the implementation of parallel computing version of the algorithm. Please refer to the original repository for more detail on how to apply this for SNP selection link. Nevertheless, the implemented software package is flexiable and can be applied for gene or protein subset selection from microarray or proteomics data.

Examples

• To obtain the general information about the program, issue following command in command line without parameters:

java -jar GEsnpxPara.jar

• Following command runs the program on the test data testset1.zip using 10 threads:

java -jar GEsnpxPara.jar -f exampleDataset.arff -t 10

• To run the above example in verbose mode:

java -jar GEsnpxPara.jar -f exampleDataset.arff -v -t 10

Note

GEsnpx.jar only accept data in ARFF format. If you have a tab-delimited data matrix with column names on the first line and the last column correspond to the instance classes (class has to be binary and coded as "0" and "1"). Please use MAT2ARFF.pl perl script to convert your data into ARFF format. The following is an example:

perl MDR2ARFF.pl [X].txt > [X].arff

where [X] is the name of the file.

Reference

1. Yang, P., Ho, J., Zomaya, A., Zhou, B. (2010). A genetic ensemble approach for gene-gene interaction identification. BMC Bioinformatics, 11(524), 1-15. [Fulltext]
2. Zhang, Z. & Yang, P. (2008). An ensemble of classifiers with genetic algorithm-based feature selection. IEEE Intelligent Informatics Bulletin, 9(1), 18-24. [Fulltext]