Diffusion-based generation of gene regulatory network from scRNA-seq data with DigNet

Abstract

Gene regulatory network (GRN) intricately encodes the interconnectedness of identities and functionalities of genes within cells, ultimately shaping to cellular specificity. Despite decades of endeavors, reverse engineering of GRN from gene expression profiling data remains a profound challenge, particularly when it comes to reconstructing cell specific GRN that are tailored to precise cellular and genetic contexts. For alternatively approaching network reconstruction from data, we propose a discrete diffusion generation model, called DigNet, capable of generating corresponding GRN from high-throughput single-cell RNA sequencing (scRNA-seq) data. DigNet embeds the network generation process into a multi-step recovery procedure with Markov properties. Each intermediate step has a specific model to recover a portion of the gene regulatory architectures. It thus can ensure compatibility between global network structures and regulatory modules through the unique multi-step diffusion procedure. Furthermore, through the meta-cell integration and non-Euclidean discrete space modeling, DigNet can robustly be resistant to the noise of scRNA-seq data and the sparsity of GRN. Benchmark evaluation results against dozens of state-of-the-art network inference methods demonstrate that DigNet achieves superior performances across various single-cell GRN reconstruction experiments. Furthermore, DigNet provides unique insights into the immune response in breast cancer, derived from differential gene regulations identified in T cells. As an open-source software, DigNet offers a powerful and effective tool for generating cell specific GRN from scRNA-seq data.