MEDICC2

MEDICC2 infers allele-specific phylogenies from somatic copy-number alteration (SCNA) data while incorporating whole-genome doubling (WGD) events to reconstruct cancer genome evolution.

Key Features:

• Allele-specific phylogeny inference: Implements a novel phylogeny inference algorithm tailored for allele-specific SCNA data.
• Whole-genome doubling (WGD) modeling: Identifies clonal and subclonal WGD events within tumor genomes.
• No infinite sites assumption: Dispenses with the infinite sites assumption to permit recurrent, overlapping, and back-mutation events.
• Parallel evolution modeling: Models parallel evolutionary events and uses multi-sample phasing to capture simultaneous changes across samples.
• SCNA timing: Times SCNAs relative to each other to order copy-number events.
• SCNA burden quantification: Quantifies SCNA burden in single-sample studies.
• Ancestral genome reconstruction: Infers ancestral genomes and reconstructs phylogenetic trees from SCNA profiles.
• Multi-sample and single-cell support: Analyzes multi-sample and single-cell sequencing data, including datasets with thousands of cells.
• Chromosomal instability and locus dependencies: Accounts for chromosomal instability (CIN) and horizontal dependencies between adjacent genomic loci.

Scientific Applications:

• Tumor phylogeny reconstruction: Reconstructs tumor evolutionary histories incorporating SCNAs and WGD events.
• Timing of WGD and SCNAs: Determines the relative timing of WGD and other SCNA events in tumor evolution.
• SCNA burden analysis: Measures SCNA burden in single-sample and single-cell studies.
• Multi-region and single-cell studies: Applied to bulk multi-region prostate cancer datasets and single-cell triple-negative breast cancer datasets.
• Large-cohort validation: Validated using 2,778 single-sample tumors from the Pan-Cancer Analysis of Whole Genomes (PCAWG).

Methodology:

Uses a phylogeny inference algorithm for allele-specific SCNA data; times SCNAs relative to each other; quantifies SCNA burden; infers phylogenetic trees and ancestral genomes from multi-sample or single-cell sequencing data and employs multi-sample phasing to detect parallel events.