geNetClassifier

geNetClassifier implements multi-class Support Vector Machine classifiers in R to classify disease subtypes and identify minimal discriminant gene markers from genome-wide microarray and RNA-Seq expression data.

Key Features:

• Multi-Class SVM training and validation: Implements training and validation of multi-class Support Vector Machine classifiers using gene expression data from microarrays and RNA-Seq.
• Gene marker selection: Selects minimal subsets of genes with transparent selection mechanisms to differentiate and classify disease subtypes.
• Coexpression network analysis: Constructs coexpression networks to explore gene-to-gene associations within specific disease contexts.
• Discriminant power measurement: Quantifies the discriminant power of selected gene markers to assess their robustness in distinguishing subtypes.
• Gene-disease mapping: Integrates gene-to-gene association, gene disease specificity, and gene discriminant power to construct gene networks and gene-disease maps.
• Implementation: Implemented in R.

Scientific Applications:

• Transcriptomic signature discovery: Identification of gene signatures associated with disease subtypes from genome-wide expression data.
• Leukemia subtype classification: Classification and molecular characterization of leukemia subtypes to distinguish pathological states relevant to diagnosis and treatment planning.

Methodology:

Implemented in R; uses multi-class SVM training and validation on microarray and RNA-Seq expression data, gene selection for minimal discriminant subsets, coexpression network construction, discriminant power quantification, and integration of gene-to-gene association and gene disease specificity to build gene-disease maps.