Parameter inference for stochastic single-cell dynamics from lineage tree data.

Area of Science:

• Computational Biology
• Systems Biology
• Single-Cell Analysis

Background:

• Advancements in time-lapse microscopy generate extensive single-cell trajectory data.
• Existing computational methods often analyze single-cell trajectories independently, neglecting crucial lineage and population structure information.
• Lineage information is vital for understanding processes occurring at cell division or those with slow dynamics relative to the cell cycle.

Purpose of the Study:

• To develop a Bayesian framework for parameter inference using single-cell time-lapse data from lineage trees.
• To leverage mother-daughter relationships and population structure for more accurate biological process modeling.
• To provide a method that accounts for lineage information crucial for specific biological dynamics.

Main Methods:

• A Bayesian framework combining Sequential Monte Carlo (SMC) and Markov Chain Monte Carlo (MCMC).
• SMC is used for approximating the parameter likelihood function.
• MCMC is employed for efficient parameter exploration within the lineage tree data.

Main Results:

• Demonstrated the framework's efficacy on two examples where lineage information is critical.
• Successfully inferred parameters for a cell phenotype switching exclusively at division.
• Accurately modeled a system where cell state fluctuates over timescales longer than the cell cycle.

Conclusions:

• Cell ancestry data is crucial for understanding biological processes like stem cell fate and bacterial phase variation.
• Methods ignoring lineage information may yield suboptimal parameter inference for such processes.
• The proposed framework offers an efficient approach to incorporate single-cell lineage data, serving as a foundation for future advancements.