Marker selection strategies for circulating tumor DNA guided by phylogenetic inference

Area of Science:

• Computational Biology
• Genomics
• Cancer Research

Background:

• Liquid biopsies using circulating tumor DNA (ctDNA) show promise for non-invasive cancer diagnosis and monitoring.
• Current computational methods need enhancement for robust quantitative analysis of tumor clonal evolution.
• Accurate characterization of clonal dynamics is crucial for early diagnosis and guiding treatment decisions.

Purpose of the Study:

• To develop novel computational methods for refining tumor phylogeny models using longitudinal ctDNA data.
• To quantify changes in clonal frequencies from ctDNA, indicative of treatment response or progression.
• To establish a probabilistic framework for optimal marker identification and characterization of clonal evolution.

Main Methods:

• Estimating density over clonal tree models using bootstrap samples from pre-treatment tissue data.
• Refining clonal tree models using successive longitudinal ctDNA samples.
• Developing optimization problems for marker selection to reduce phylogenetic uncertainty and quantify clonal frequencies.

Main Results:

• The proposed methods effectively refined tree densities and inferred clonal frequencies on synthetic data.
• Application to real tumor data demonstrated improved lineage model refinement and clonal frequency assessment.
• The computational framework enhances marker selection, lineage reconstruction, and clonal dynamics profiling.

Conclusions:

• Computational advancements are key to making liquid biopsies a robust quantitative assay for tumor clonal evolution.
• The developed methods improve the precision and quantitation of somatic evolution and tumor progression analysis.
• This work paves the way for more accurate non-invasive cancer diagnostics and personalized treatment strategies.