JCDSA

JCDSA implements joint covariate detection and a two-step feature-selection strategy to identify miRNA and other tumor expression covariates associated with patient survival times and risk stratification in survival analysis.

Key Features:

• Two-Step Feature Selection Strategy: Employs a dual-phase approach to identify features that are consistent with survival times and that distinguish low-risk and high-risk patient groups.
• Joint Covariate Detection: Detects jointly significant covariates that may be missed by univariate methods, capturing multivariate interactions in gene expression data.
• Application to miRNA Expression Data: Applied to Level 3 miRNA expression data from 548 glioblastoma multiforme (GBM) patients to select miRNA candidates correlated with prognosis.
• Validation through Simulations: Validated via extensive simulations indicating robustness and accuracy of selected covariates in predicting survival outcomes.

Scientific Applications:

• Prognostic biomarker discovery: Identifies miRNA and other expression covariates as candidate prognostic biomarkers for cancer.
• Glioblastoma multiforme survival analysis: Facilitates selection of miRNA candidates correlated with patient survival times and risk categories in glioblastoma multiforme using Level 3 miRNA data from 548 patients.
• Risk stratification: Supports stratification of patients into low-risk and high-risk groups based on joint covariate profiles.
• Experimental prioritization: Prioritizes candidates for downstream experimental validation and investigation of molecular underpinnings of cancer progression and treatment response.

Methodology:

Two-step feature selection strategy, joint covariate detection, statistical analysis, and validation via simulations.