CEPICS

CEPICS evaluates and compares computational cancer subtyping methods by integrating multi-omics molecular data to assess clustering performance and clinical relevance.

Key Features:

• Integration Method Evaluation: Evaluates up to five built-in multi-omics integration methods to generate cancer subtype predictions.
• Customizable Subtype Range (k-max): Allows specification of a maximum number of subtypes (k-max) and analysis across subtype counts from two up to k-max.
• Performance Metrics: Computes clustering metrics including Normalized Mutual Information (NMI) and Adjusted Rand Index (ARI) against a user-provided gold standard or via inter-method comparisons.
• Clinical Survival Analysis: Performs survival analysis to assess the clinical relevance of identified subtypes.
• Robust Subtype Aggregation: Analyses and aggregates subtyping results across methods to identify consistent and robust subtype predictions.
• User-Defined Result Comparison: Accepts external subtyping results for direct comparison with platform-generated results.

Scientific Applications:

• Method Benchmarking: Systematically compares and ranks multi-omics integration methods using NMI and ARI across specified subtype counts.
• Subtype Discovery and Validation: Identifies and validates cancer subtypes by aggregating results from multiple integration methods.
• Clinical Relevance Assessment: Links identified subtypes to clinical outcomes through survival analysis to evaluate potential therapeutic implications.
• Multi-omics Integration Evaluation: Assesses the effect of integrating diverse molecular data types on subtype reproducibility and concordance.

Methodology:

Apply up to five integration methods to multi-omics data across subtype counts from 2 to k-max; compute NMI and ARI against a provided gold standard or via inter-method comparisons; perform clinical survival analysis; aggregate method-specific subtyping results and compare to user-uploaded results.