Nonparametric single-cell multiomic characterization of trio relationships between transcription factors, target genes, and cis-regulatory regions

Mapping disease regulatory circuits at cell-type resolution from single-cell multiomics data

“…Existing integrative methods 9,10,14 use single-condition data only and correlate chromatin site accessibility and gene expression primarily for peak-gene looping inference. To provide a comparative assessment of the performance of MAGICAL, we restricted MAGICAL to the single-condition data analysis possible with existing methods.…”

“…For benchmarking, MAGICAL was applied to a 10X PBMC single cell multiome dataset (https://support.10xgenomics.com/single-cell-multiome-atacgex/datasets/1.0.0/pbmc_granulocyte_sorted_10k), including 108377 ATAC peaks, 36601 genes, and 11909 cells from 14 cell types. MAGICAL used the same candidate peaks and genes as selected by TRIPOD 9 for fair performance comparison. Two different ways were used to pair candidate peaks and genes: (1) the peaks and genes were within the same prior TAD from the GM12878 cell line; (2) the centers of peaks and the TSS of genes were within 500k bps.…”

“…We addressed the challenging problem of distinguishing host regulatory circuit responses to methicillin-resistant-(MRSA) and methicillin-susceptible Staphylococcus aureus (MSSA) infections. While differential expression analysis failed to show predictive value, MAGICAL identified epigenetic circuit biomarkers that distinguished MRSA from MSSA.Recent approaches [7][8][9][10] to support integrated analysis of multiomics data demonstrate the promise of this area but lack the capacity to resolve high-resolution regulation changes within individual cell types, which precludes elucidating regulatory circuits affected by the disease or different disease states.To address these, we developed MAGICAL (Multiome Accessibility Gene Integration Calling And Looping), a method that models coordinated chromatin accessibility and gene expression variation to identify circuits that differ between conditions. MAGICAL analyzes scRNA-seq and scATAC-seq data using a hierarchical Bayesian framework that improves model robustness by leveraging information on transcription factor motifs and 3D topologically associated domains.…”

“…Recent approaches [7][8][9][10] to support integrated analysis of multiomics data demonstrate the promise of this area but lack the capacity to resolve high-resolution regulation changes within individual cell types, which precludes elucidating regulatory circuits affected by the disease or different disease states.…”

Destin2: integrative and cross-modality analysis of single-cell chromatin accessibility data

“…SnapATAC [11] computes a Jaccard similarity matrix, while cisTopic [12] performs topic modeling. Moving beyond the peak matrix, Cicero [13] and MAESTRO [14] make gene expression predictions from unweighted and weighted sum of the ATAC reads in gene bodies and promoter regions, respectively; the predicted gene activities have been shown to in the ballpark recapitulate the transcriptomic profiles and discern cell populations [15]. For TF-binding motifs, chromVAR [16] computes a motif deviation score by estimating the gain or loss of accessibility within peaks sharing the same motif relative to the average cell profile; these deviation scores have also been shown to enable accurate clustering of scATAC-seq data.…”