Tumor haplotype assembly algorithms for cancer genomics

Area of Science:

• Genomics
• Computational Biology
• Cancer Research

Background:

• High-throughput sequencing enables deep tumor population analysis.
• Reconstructing tumor haplotypes is crucial for understanding cancer evolution, driver mutations, and gene fusions.
• Existing haplotype phasing methods are insufficient for complex cancer genomes.

Purpose of the Study:

• To develop a computational framework for haplotype assembly in copy number variable cancer genomes.
• To address challenges like haplotype heterogeneity, contamination, polyploidy, and complex variations.
• To provide a novel algorithmic approach for accurate cancer haplotype reconstruction.

Main Methods:

• Extended a polyploid haplotype assembly model.
• Developed novel algorithms for complex variations (e.g., copy number changes) represented as disjoint paths.
• Incorporated methods to handle variable haplotype frequencies and contamination.
• Utilized simple cycles of a compass graph to constrain haplotype assembly solutions.

Main Results:

• Introduced HapCompass-Tumor, a computational modeling and algorithmic framework.
• The framework effectively handles complex variations, including copy number changes.
• Successfully addressed challenges of variable haplotype frequencies and contamination.
• Enabled computation of tumor haplotypes by constraining the solution space.

Conclusions:

• HapCompass-Tumor provides a robust solution for cancer genome haplotype assembly.
• The framework advances the ability to study cancer genome evolution and mutational phenomena.
• HapCompass-Tumor is available as a software package for researchers.