IOFS-SA

IOFS-SA performs interactive feature selection and sample grouping for survival analysis to identify prognostic gene signatures and assess survival outcomes in high-dimensional, small-sample genomic datasets including TCGA cancer cohorts.

Key Features:

• Automatic and manual feature selection: Provides both automatic and manual feature selection applied to user datasets and TCGA data.
• Interactive sample grouping: Implements interactive risk score splits combined with hierarchical clustering and an automatic re-clustering strategy for sample grouping.
• Survival analysis metrics: Uses Kaplan-Meier survival curves and log-rank tests to evaluate differences in survival between groups.
• Visualizations: Produces tree views, heat maps, and scatter maps of selected genes and sample groupings.
• High-dimensional, small-sample suitability: Designed to operate on high-dimensional genomic datasets with small sample sizes.

Scientific Applications:

• Oncology survival analysis: Analysis of cancer patient survival and postoperative treatment outcomes using TCGA cohorts.
• Prognostic biomarker discovery: Identification and evaluation of prognostic gene signatures from high-dimensional genomic data.
• Risk stratification and grouping refinement: Refinement of sample grouping and risk stratification strategies via interactive splitting, hierarchical clustering, and re-clustering.

Methodology:

Feature selection (automatic and manual) on user or TCGA data; interactive risk score splitting followed by hierarchical clustering and an automatic re-clustering strategy for sample grouping; Kaplan-Meier survival curves and log-rank tests for statistical assessment.