Pyrea

Pyrea performs multi-view hierarchical ensemble clustering to identify cancer subtypes from diverse molecular patient data using the Parea method.

Key Features:

• Multi-View Hierarchical Ensemble Clustering: Employs a hierarchical ensemble approach that leverages multiple data views to capture heterogeneity in cancer datasets.
• Integration of the Parea method: Incorporates the Parea method as part of its multi-view ensemble framework.
• Integration of Diverse Fusion and Clustering Algorithms: Supports combining various fusion and clustering algorithms to create flexible ensemble workflows.
• Performance Validation: Evaluated on machine learning benchmark datasets and validated on real-world multi-view patient data from seven cancer types, outperforming state-of-the-art methods in six of the seven analyzed cancer types.

Scientific Applications:

• Cancer subtype discovery: Clusters patients by molecular profiles to identify disease subtypes.
• Patient stratification for personalized medicine: Enables molecular stratification to inform targeted therapeutic strategies.
• Cross-cancer comparative analysis: Applied across multiple cancer types to assess subtype reproducibility and method performance.

Methodology:

Implements the Parea method via a multi-view hierarchical ensemble clustering approach that integrates multiple data views and combines various fusion and clustering algorithms.