SVMDO

SVMDO detects tumor/normal sample discriminating gene sets from differentially expressed genes by integrating Disease Ontology enrichment analysis with support vector machine classification.

Key Features:

• Iterative Gene Filtering: The algorithm iteratively filters genes using Disease Ontology enrichment analysis combined with Wilk's lambda criterion prior to SVM classification model construction.
• Gene Set Extraction: SVMDO extracts stable small gene sets that discriminate between tumor and normal states with high accuracy.
• Prognostic Marker Detection: The tool provides individual prognostic marker detection alongside gene set extraction.
• RNA-Seq Compatibility: The algorithm is designed for FPKM and RPKM normalized RNA-Seq transcriptome datasets.

Scientific Applications:

• Cancer Classification: SVMDO identifies discriminatory mRNA signatures for predicting cancer status in tumor versus normal samples.
• Biomarker Discovery: The method discovers novel and practical sets of discriminatory biomarkers by utilizing the association of cancer and disease.
• Survival Analysis: The generated gene sets perform well in survival analyses, indicating potential for prognosis.
• Clinical Cancer Research: By combining gene sets for both diagnosis and prognosis, the method can improve clinical applications in cancer research.

Methodology:

The approach employs SVM classification for differentially expressed genes enriched by Disease Ontology, with nondiscriminatory features filtered using Wilk's lambda criterion prior to classification. The algorithm was validated using comprehensive RNA-Seq data for adenocarcinoma of the colon, squamous cell carcinoma of the lung, and adenocarcinoma of the lung.