sigFeature

sigFeature identifies biologically relevant features for binary classification of gene expression data by integrating Support Vector Machine Recursive Feature Elimination (SVM-RFE) with t-statistics.

Key Features:

• Integration of SVM-RFE and t-statistic: sigFeature integrates Support Vector Machine Recursive Feature Elimination (SVM-RFE) with t-statistics to identify features differentially significant between two classes.
• Binary classification focus: Targets binary classification tasks for gene expression datasets.
• Automatic feature selection: Provides a central function in an R package that automates the feature selection process.
• Emphasis on biological significance: Prioritizes differentially expressed features to increase biological relevance of selected signatures.

Scientific Applications:

• Gene expression classification: Performs feature selection to improve classification of gene expression profiles, particularly for binary outcomes.
• Microarray data analysis: Applies to microarray datasets to identify differentially expressed genes for studying biological responses to treatments or conditions.

Methodology:

Combines SVM-RFE ranking with t-statistic assessment; performs comparative analysis on six publicly available microarray datasets from the Gene Expression Omnibus (GEO) against three other feature selection algorithms; selected features underwent gene set enrichment analysis for downstream validation.