Reference-free inference of tumor phylogenies from single-cell sequencing data

Area of Science:

• Genomics
• Computational Biology
• Cancer Research

Background:

• Cancer evolution is complex due to tumor heterogeneity.
• Single-cell sequencing offers insights into intra-tumoral heterogeneity.
• High-throughput sequencing data presents challenges for phylogenetic inference, especially with chromosomal abnormalities.

Purpose of the Study:

• To develop a reference-free computational strategy for analyzing single-cell tumor genomic data.
• To delineate intra-tumoral heterogeneity and reconstruct tumor evolutionary models.
• To track differences in tumor cell genomic content during cancer progression.

Main Methods:

• A reference-free approach using k-mer counts from single-cell genome sequences.
• Utilizing differences in normalized k-mer frequencies as a proxy for evolutionary distance.
• Bypassing challenges posed by massive genome rearrangements in tumor genomes for reference-based methods.

Main Results:

• The proposed method allows for predictive classification of tumors into heterogeneous cell types.
• K-mer features accurately classify tumor cells by their stage of progression.
• Phylogenetic trees constructed from k-mer distance matrices show statistically significant partitioning of cells at different cancer stages.

Conclusions:

• A computational approach for learning tumor progression from single-cell sequencing data using k-mer counts has been demonstrated.
• This method accurately classifies tumor cells based on their progression stage.
• The generated phylogenies provide statistically significant insights into cancer cell partitioning and evolution.