Inference of differential key regulatory networks and mechanistic drug repurposing candidates from scRNA-seq data with SCANet

Abstract

Motivation: The reconstruction of small key regulatory networks that explain the differences in the development of cell (sub)types from single-cell RNA sequencing is a yet unresolved computational problem. Results: To this end, we have developed SCANet, an all-in-one package for single-cell profiling that covers the whole differential mechanotyping workflow, from inference of trait/cell-type-specific gene co-expression modules, driver gene detection, and transcriptional gene regulatory network reconstruction to mechanistic drug repurposing candidate prediction. To illustrate the power of SCANet, we examined data from two studies. First, we identify the drivers of the mechanotype of a cytokine storm associated with increased mortality in patients with acute respiratory illness. Secondly, we find 20 drugs for eight potential pharmacological targets in cellular driver mechanisms in the intestinal stem cells of obese mice. Availability and implementation: SCANet is a free, open-source, and user-friendly Python package that can be seamlessly integrated into single-cell-based systems medicine research and mechanistic drug discovery.