Quantifying pluripotency landscape of cell differentiation from scRNA-seq data by continuous birth-death process

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Cell differentiation modeling from omics data is crucial in systems biology.
• Existing single-cell RNA sequencing (scRNA-seq) methods struggle to accurately infer cell pseudo-time and quantify cell potency.
• Understanding the dynamic nature of cell differentiation remains a challenge.

Purpose of the Study:

• To introduce a novel computational method, Landscape of Differentiation Dynamics (LDD).
• To accurately quantify cell potency and construct differentiation landscapes from scRNA-seq data.
• To provide a dynamic perspective on cell differentiation processes.

Main Methods:

• Developed the Landscape of Differentiation Dynamics (LDD) method.
• Utilized a continuous birth-death process to model cell differentiation.
• Employed a source-sink diffusion process to quantify the differentiation landscape based on stochastic dynamics.
• Applied LDD to analyze seven benchmark scRNA-seq datasets.

Main Results:

• LDD accurately and efficiently constructs cell evolution trees with pseudo-time.
• LDD effectively quantifies cell potency and the differentiation landscape.
• The method demonstrates superior performance compared to existing scRNA-seq analysis techniques.

Conclusions:

• LDD provides a robust computational tool for quantifying cell potency and Waddington potential landscapes.
• The method offers novel insights into the dynamic mechanisms of cell differentiation.
• This work advances the understanding of cell development through a dynamic modeling approach.