Decomposing multi-scale dynamic regulation from single-cell multiomics with scMagnify

Deciphering the highly coupled regulatory circuits that drive cellular dynamics remains a fundamental goal in biology. However, capturing the multi-scale time-lagged dynamics and combinatorial regulatory logic of gene regulation remains challenging. Here we present scMagnify, a deep-learning-based framework that leverages chromatin accessibility and gene expression via Granger causality to reconstruct and decompose multi-scale gene regulatory networks (GRNs). Diverse benchmarks demonstrate scMagnify’s favorable performance. scMagnify employs tensor decomposition to identify combinatorial TF modules and their activation profiles across multiple time-lags. We applied scMagnify to human hematopoiesis and mouse pancreas development, where it recovered known lineage-driving regulators and provided further insights into the combinatorial logic underlying cell fate decisions. Furthermore, in the context of kidney injury, scMagnify’s intercellular communication module mapped key signaling-to-transcription cascades linking microenvironment cues to pathological epithelial cell changes. In summary, scMagnify provides a robust computational framework for dissecting the multi-scale regulatory logic that governs complex biological processes in development and disease. Chen and colleagues present scMagnify, a framework that uses single-cell multi-omic data to reconstruct and decompose time-lagged gene regulatory networks. The study analyses regulatory programs associated with development and disease.