BAMSE: Bayesian model selection for tumor phylogeny inference among multiple samples.

Area of Science:

• Oncology
• Computational Biology
• Genetics

Background:

• Intra-tumor heterogeneity significantly contributes to cancer complexity and drug resistance.
• Identifying distinct tumor subclones and their evolutionary dynamics is crucial for clinical applications.
• Understanding tumor evolution remains a significant challenge in cancer research.

Purpose of the Study:

• To present BAMSE (BAyesian Model Selection for tumor Evolution), a novel probabilistic method for inferring subclonal history and lineage tree reconstruction.
• To provide a flexible and fast software tool for analyzing heterogeneous tumor samples.
• To accurately reconstruct tumor evolutionary history using somatic mutation data.

Main Methods:

• BAMSE utilizes somatic mutation read counts from multiple tumor samples as input.
• It involves scoring mutation clusters into subclones and generating evolutionary trees.
• A Bayesian model calculates posterior probabilities, integrating prior beliefs and accounting for sequencing errors.

Main Results:

• BAMSE accurately and efficiently infers subclonal history and reconstructs lineage trees.
• The method leverages multiple tumor samples for robust analysis.
• Performance was benchmarked against state-of-the-art software using simulated datasets.

Conclusions:

• BAMSE is a flexible and fast open-source software for reconstructing tumor subclonal evolution.
• The tool aids in understanding tumor evolutionary history using somatic mutation data.
• BAMSE is implemented in Python and available under GNU GLPv3.