Constructing and optimizing dynamical models of core transcription factor regulatory networks driving cell state transitions

A key challenge in systems biology is to uncover the gene regulatory mechanisms that govern cell state transitions during developmental processes and disease progression. Recent advancements in single-cell RNA sequencing (scRNA-seq) technology have greatly improved our understanding of genome-wide gene expression dynamics. However, deciphering the gene regulatory networks responsible for these transitions from scRNA-seq data remains challenging. Here, we present NetDes, an approach that integrates top-down bioinformatics with bottom-up systems biology, designed to computationally construct and optimize ODE-based nonlinear dynamical models of core transcription factor regulatory networks that reflect observed gene expression time trajectories. Through in-silico benchmarking and application to time series scRNA-seq data for iPSC cell differentiation, we demonstrate that NetDes has advantages in identifying true regulators and their combinations, as well as in capturing gene expression dynamics during cell state transitions from a single dynamical model. Our approach sheds light on a mechamistic understanding of gene regulation of cell state transitions.