Phylodynamics of Somatic Evolution: A Likelihood-Based Approach for Cellular Reproduction.

Area of Science:

• Evolutionary biology
• Computational biology
• Genetics

Background:

• Understanding cell population evolution requires models connecting phylogenetic patterns to cell division, death, and mutation.
• Classical phylodynamic methods often assume mutations accrue over calendar time, relying on molecular clocks, which may not suit all biological systems.

Purpose of the Study:

• To develop a novel phylodynamic framework that directly links mutations to discrete cell division events.
• To create a model that accounts for both observed and unobserved cell divisions and their impact on mutation accumulation.

Main Methods:

• Introduced a framework where mutations accumulate via a compound Poisson process tied to cell birth (division) events.
• Developed a computationally efficient dynamic programming algorithm to calculate phylogenetic likelihoods.
• Integrated latent variables like branch durations and unobserved cell divisions into the model.

Main Results:

• The new framework successfully models mutation accumulation based on cell divisions.
• The dynamic programming algorithm efficiently computes likelihoods for phylogenetic trees with mutations.
• Demonstrated the method's applicability on simulated data and real single-cell data from haematopoietic stem cells.

Conclusions:

• The proposed model provides a more accurate representation of evolutionary dynamics by linking mutations to cell divisions.
• This method is suitable for analyzing large-scale single-cell datasets, advancing phylodynamic inference.
• Offers a powerful tool for studying the evolution of cell populations, particularly in contexts like stem cell dynamics.