survClust

survClust classifies multi-omic samples based on their available time to event information using an outcome weighted integrative clustering algorithm that produces cross-validated, molecularly distinct and clinically meaningful clusters.

Key Features:

• Integrative clustering: Combines multiple omic data types for unified sample classification.
• Outcome weighted learning: Incorporates time to event information as a weighting factor during clustering.
• Cross-validation: Clusters are cross-validated to avoid overfitting.
• Flexible data input: Accepts binary mutation data as well as continuous data from other omic types.

Scientific Applications:

• Genomic subtyping: Identifies clinically relevant genomic subtypes from multi-omic datasets.
• Survival-based patient stratification: Classifies samples into clusters that are both molecularly distinct and clinically meaningful based on outcome information.

Methodology:

survClust uses outcome weighted integrative clustering with cross-validation to classify multi-omic samples based on time to event information.