CAPIU

CAPIU clusters samples by integrating gene-class annotations with gene expression data to identify biologically meaningful sample groupings.

Key Features:

• Integration of Biological Information: Incorporates prior biological knowledge about gene classes during clustering to enhance biological interpretability.
• Algorithmic Approach: Searches predefined gene classes to identify those exhibiting strong sample clustering and uses them iteratively to bifurcate samples until no significant splits remain.
• Decision Tree-Like Visualization: Produces a dendrogram resembling a decision tree in which nodes represent gene classes and leaves denote grouped samples.
• Complementary Methodology: Does not rely solely on high-variance genes, enabling detection of groupings when there are few differentially expressed genes given an informative mapping of genes to classes.

Scientific Applications:

• Disease Subtype Identification: Groups patient samples by gene expression to identify disease subtypes.
• Patient Stratification: Stratifies patients into groups for targeted treatments.
• Treatment Relationship Analysis: Clusters related treatments to help elucidate treatment relationships and potential efficacy.

Methodology:

Computes decision-tree-like dendrograms by searching predefined gene classes for strong sample clustering and iteratively splitting samples by selected gene classes until no significant splits remain.