Single-cell and spatial multiomic inference of gene regulatory networks using SCRIPro

The accurate reconstruction of gene regulation networks (GRNs) from sparse and noisy single-cell or spatial multi-omics data remains a challenge. Here, we present SCRIPro, a comprehensive computational framework that robustly infers GRNs for both single-cell and spatial multi-omics data. SCRIPro first addresses sample sparseness by a density clustering approach. SCRIPro assesses transcriptional regulator (TR) importance through chromatin reconstruction and in silico deletion, referencing 1,292 human and 994 mouse TRs. It combines TR-target importance scores with expression levels for precise GRN reconstruction. Finally, we benchmarked SCRIPro on diverse datasets, it outperforms existing motif-based methods and accurately reconstructs cell type-specific, stage-specific, and region-specific GRNs.