HDG-select

HDG-select identifies and classifies genes in high-dimensional datasets to detect gene-disease associations using filter methods combined with Genetic-Based Particle Swarm Optimization (GBPSO) and Support Vector Machines (SVM).

Key Features:

• Advanced algorithm integration: Integrates filter methods with Genetic-Based Particle Swarm Optimization (GBPSO) and Support Vector Machines (SVM) for gene selection and classification.
• Dimensionality reduction: Employs filter methods to reduce dimensionality and remove irrelevant features prior to optimization and classification.
• Optimization-driven selection: Uses GBPSO to explore the search space and identify relevant gene subsets.
• Robust classification: Applies SVMs for categorization of genes based on association with specific conditions.
• Validation and data formats: Validated across eleven high-dimensional datasets, including CSV and GEO soft formats, with performance reported relative to existing methods in the literature.

Scientific Applications:

• Gene-disease association analysis: Identification and classification of genes associated with specific diseases.
• Biomarker discovery: Support for detecting potential biomarkers relevant to disease states.
• Therapeutic target identification: Prioritization of gene targets for therapeutic investigation.
• High-dimensional genomic studies: Application to genomic datasets requiring feature selection and classification in high-dimensional settings.

Methodology:

HDG-select integrates filter methods with a GBPSO-SVM framework: filter methods reduce dimensionality and remove irrelevant features; GBPSO performs optimization to identify relevant gene subsets; SVMs provide classification of selected genes.