A Mechanism-Informed Deep Neural Network Enables Prioritization of Regulators That Drive Cell State Transitions

The study introduces regX, a deep neural network designed to prioritize regulators driving cell state transitions by incorporating gene-level regulation and gene-gene interaction mechanisms. Applied to single-cell multi-omics data on type 2 diabetes and hair follicle development, regX effectively identifies key transcription factors and candidate cis-regulatory elements. This approach not only aids in understanding cellular events but also highlights potential new therapeutic targets, drug repurposing opportunities, and causal single nucleotide polymorphisms. The research demonstrates the advantage of using interpretable neural networks to decode complex biological systems.