Modeling Dynamics, Cell Type Specificity, and Perturbations in Gene Regulatory Networks

TL;DR

The current state of GRN inference from single cell omic datasets is summarized, with a particular focus on dynamics and perturbations, and key open challenges that must be addressed to advance the field are outlined.

Abstract

Gene regulatory networks (GRNs) define the regulatory relationships among molecules such as transcription factors, chromatin remodelers, and target genes. GRNs play a critical role in diverse biological processes, including development, disease manifestation, and evolution. However, fully characterizing these networks across multiple cell types and states remains a significant challenge. Recent advances in single-cell omics have dramatically enhanced our ability to measure biological systems at unprecedented resolution. These technologies have opened new avenues for computational methods to infer GRNs, offering deeper insights into cell type-specific mechanisms, causality, and dynamic regulatory processes. This review summarizes the current state of GRN inference from single cell omic datasets, with a particular focus on dynamics and perturbations, and outlines key open challenges that must be addressed to advance the field.

Figures (4)

• Figure 1: A typical workflow of the main stages of inference and analysis of gene regulatory networks (GRNs) from single cell omics data: Pre-processing, Cell clustering and annotation, Cell-cell relationship inference, GRN inference and interpretation. Each stage has multiple tasks. Shown are the inputs, computational approaches (Do) and outputs produced by each task. E-P, enhancer-promoter; CRE, cis-regulatory element.
• Figure 2: Inputs and outputs for modeling dynamics in GRNs. Left box: Types of dynamics among cells inferred using single cell omics data: Pseudotime and RNA velocity, used for input into GRN inference algorithm. Either input can order cells on linear or branching trajectories. Key modeling techniques for GRN dynamics: Granger causality, transfer entropy, statistical regression, which rely on temporal ordering of cells. Right box: Different types inferred dynamics in GRNs. "Global GRN" methods infer a single network structure with a single dynamical model capturing expression of target gene as a function of expression of a regulator using expression from a previous cell. "'Cluster-specific GRN" each cell cluster can have a different network structure. Within each cluster methods may or may not additional dynamics among individual cells within a cluster. Circular nodes: transcriptional regulators; square nodes: target genes.
• Figure 3: Integration of multi-omics datasets for cis and trans-GRN inference. Various methods for integrating chromatin accessibility in gene regulatory network (GRN) inference. Chromatin Informed prior: use of peaks from chromatin accessibility assays to guide structure of trans-GRNs. Prediction with CRE: inference of CRE-gene interactions using correlation or regression models predicting target gene expression from CRE accessibility. 3D chromatin conformation: Use of long-range interactions between CREs and genes to inform GRN structure. Circular nodes: transcriptional regulators; square nodes: target genes.
• Figure 4: Inference of causal relationships in GRNs from high-throughput perturbation data, such as Perturb-seq. Approaches can be grouped into defining gene expression programs (Co-expression/modules), inferring regulators for the programs (Regulating Programs), inferring causal TF-target interactions (Causal GRNs). Circular nodes: transcriptional regulators; square nodes: target genes.