Building dynamical models of multi-step state transitions from single cell gene expression trajectories

Area of Science:

• Computational Biology
• Systems Biology
• Genomics

Background:

• Multi-step cell state transitions are crucial in biology (e.g., differentiation, disease) but their regulatory mechanisms are poorly understood.
• Inferring gene regulatory networks and modeling dynamic processes from single-cell data presents significant challenges.

Purpose of the Study:

• To introduce NetDes, a novel computational method for inferring transcription factor regulatory networks.
• To develop ODE-based dynamical models from single-cell gene expression trajectories to understand cell state transitions.
• To provide a generalizable framework for mechanistic modeling of gene regulation in complex biological processes.

Main Methods:

• NetDes integrates top-down and bottom-up systems biology approaches.
• It infers core transcription factor regulatory networks and builds dynamical models using single-cell RNA sequencing (scRNA-seq) time-series data.
• Benchmarking involved in-silico time trajectories, gene circuit simulations, and application to human induced pluripotent stem cell (iPSC) differentiation data.

Main Results:

• NetDes successfully predicts regulatory interactions and reproduces gene expression dynamics.
• The method captures sequential state transitions within a single dynamical model, outperforming existing approaches.
• Network simulations and coarse-graining elucidated the regulatory roles of genes in driving cell state transitions.

Conclusions:

• NetDes offers a powerful, generalizable framework for mechanistic modeling of gene regulation during complex cell state transitions.
• The computational method advances our understanding of the regulatory underpinnings of biological processes like cell differentiation.
• This approach enables more accurate prediction and modeling of dynamic biological systems using scRNA-seq data.