BRIDGE: Biological Evidence Refinement and Heterogeneous Dynamic Gating for Gene Regulatory Networks

The article introduces BRIDGE, a novel framework for gene regulatory network inference from single-cell RNA sequencing data, which addresses the challenges of sparse and noisy measurements. BRIDGE employs contrastive learning and heterogeneous gated encoding to enhance information transfer between genes and cells, achieving state-of-the-art AUROC and AUPRC metrics across multiple datasets, with a notable 5% improvement in AUPRC over the baseline GCLink. This advancement is significant for practitioners as it enhances the robustness and accuracy of transcription factor-target gene predictions, crucial for understanding cell-state-specific transcriptional programs.