Deconvolution and phylogeny inference of structural variations in tumor genomic samples

Area of Science:

• Genomics
• Computational Biology
• Evolutionary Biology

Background:

• Phylogenetic reconstruction is vital for understanding complex cancer genomic data.
• Tumor evolution presents unique challenges compared to species evolution, particularly regarding structural variations (SVs).
• Inference of SVs has been limited in tumor phylogenetics due to detection and interpretation complexities.

Purpose of the Study:

• To develop a novel method for reconstructing evolutionary trajectories of structural variations (SVs) in tumors.
• To infer clonal sub-populations and their ancestry from bulk whole-genome sequence data.
• To address the underrepresentation of SVs in tumor phylogenetics.

Main Methods:

• A novel likelihood model for joint deconvolution and phylogenetic inference on bulk SV data.
• An associated optimization algorithm to reconstruct evolutionary trajectories.
• Application to simulated data and The Cancer Genome Atlas breast cancer genomic data.

Main Results:

• The method efficiently and accurately reconstructs SV evolutionary trajectories in realistic scenarios.
• It successfully infers clonal sub-populations and their ancestry.
• Demonstrated practical effectiveness in reconstructing SV-driven evolution in single tumors using real breast cancer data.

Conclusions:

• The developed method provides a robust approach for analyzing SVs in cancer evolution.
• It enables deeper insights into tumor heterogeneity and evolutionary dynamics.
• The tool is available for broader application in cancer genomics research.