Canopy2: Tumor Phylogeny Inference by Bulk DNA and Single-Cell RNA Sequencing.

Area of Science:

• Oncology
• Computational Biology
• Genetics

Background:

• Tumor heterogeneity, characterized by diverse cell populations, drives drug resistance and treatment failure.
• Understanding tumor cell evolution through phylogenetic trees is crucial for deciphering cancer complexity.

Purpose of the Study:

• To introduce Canopy2, a Bayesian framework for inferring tumor phylogeny and profiling mutations in subpopulations.
• To address challenges in single-cell data analysis, including data sparsity and stochasticity.

Main Methods:

• Canopy2 integrates single nucleotide variants from bulk DNA and single-cell RNA sequencing data.
• It employs Markov chain Monte Carlo methods with binomial and beta-binomial distributions.
• The framework distinguishes between non-cancerous, stochastic, and technical sources of zero counts in single-cell data.

Main Results:

• Simulations show Canopy2 outperforms existing methods in reconstructing clonal trees with high fidelity.
• The framework demonstrates robustness even with low sequencing depth, poor single-cell yield, and complex tumor structures.
• Performance was validated on breast cancer and glioblastoma datasets.

Conclusions:

• Canopy2 provides a robust and accurate method for tumor phylogeny inference and mutational profiling.
• This framework enhances understanding of tumor evolution and heterogeneity, potentially informing targeted cancer therapies.
• Canopy2 is an open-source R package, facilitating broader research application.