SPM

SPM implements the Subtype-specific Pathway Linear Progression Model to infer subtype-specific ordering of cancer pathway dysregulation from Cancer Cell Fraction (CCF) matrices using integer linear programming.

Key Features:

• Simultaneous identification: Uses integer linear programming to concurrently identify cancer subtypes, assign somatic variants to pathways, and infer the temporal sequence of pathway dysregulation for each subtype.
• Input data: Processes a Cancer Cell Fraction (CCF) matrix where rows are samples and columns are somatic genetic variants with measured CCF values.
• Outputs: Produces a partitioning of samples into subtypes, classification of genetic variants into pathways, and a linear progression model detailing the order of pathway dysregulation per subtype.
• Robustness and validation: Demonstrated robustness to noise on synthetic datasets and validated on glioblastoma multiforme and colorectal adenocarcinoma with results aligning to established knowledge and showing improved performance in certain scenarios.

Scientific Applications:

• Cancer research: Elucidates subtype-specific orders of pathway dysregulation to support investigations of molecular mechanisms driving cancer progression.
• Personalized medicine: Differentiates cancer subtypes based on variant-to-pathway assignments and progression order to inform subtype-targeted therapeutic hypotheses.

Methodology:

Intermediate file generation via matGenerator.cpp to prepare data structures from CCF matrices; a main SPM implementation that executes integer linear programming to simultaneously identify subtypes, classify variants into pathways, and infer linear pathway progression for each subtype.