bindSC

bindSC performs bi-order integration of single-cell multi-omic datasets to align rows and columns of data matrices and generate multimodal co-embeddings for unpaired cells measured with unmatched features across modalities.

Key Features:

• Implementation: Provided as an R package for single-cell multi-omic integration.
• Bi-order alignment: Aligns both rows and columns between data matrices without approximations to preserve relationships among features and cells.
• Multimodal co-embeddings: Simultaneously generates co-embeddings by aligning entities represented by rows and columns across datasets.
• Unpaired/unmatched integration: Integrates unpaired cells measured with unmatched features across different modalities.
• scRNA-seq and scATAC-seq integration: Supports integration of single-cell RNA sequencing (scRNA-seq) and single-cell chromatin accessibility sequencing (scATAC-seq).
• 10x Multiome support: Applied to data generated with the 10x Genomics Multiome ATAC+RNA kit.
• Spatial transcriptomics integration: Integrates scRNA-seq with 10x Visium spatial transcriptomics data for spatially resolved analyses.
• RNA–protein integration: Integrates single-cell RNA and protein data to delineate immune cell types despite discordance between gene expression and protein abundance.
• Rare cell recovery: Capable of integrating and recovering both common and rare cell types, including populations with abundance <0.25%.
• Comparative accuracy: Demonstrated improvements in integration accuracy relative to Seurat, Liger, and Harmony for scRNA-seq and scATAC-seq integration.
• Biological insight: Enables discovery of regulatory elements and cellular identities in cancer cell lines and mouse cells.

Scientific Applications:

• Regulatory element discovery: Integrates scRNA-seq and scATAC-seq to uncover key regulatory elements in cancer cell lines and mouse cells.
• Retina cell atlas construction: Integrates 10x Multiome data to resolve common and rare cell types in a mouse retina cell atlas.
• Spatial gene expression mapping: Integrates scRNA-seq with 10x Visium to spatially resolve gene expression patterns in mouse brain cortex.
• Immune cell delineation: Integrates single-cell RNA and protein data to delineate immune cell types despite sparsity in scRNA-seq data and effects of post-translational modifications.

Methodology:

Performs simultaneous alignment of rows and columns between data matrices to generate multimodal co-embeddings for integration of unpaired cells with unmatched features.