scMoC: single-cell multi-omics clustering.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell multi-omics assays provide complementary molecular information from individual cells.
• Integrating diverse single-cell data types is crucial for comprehensive biological insights.
• Cross-modal clustering remains a challenge due to data sparsity and heterogeneity.

Purpose of the Study:

• To develop a computational approach for cross-modal clustering of single-cell RNA sequencing (scRNA-seq) and single-cell Assay for Transposase-Accessible Chromatin sequencing (scATAC-seq) data.
• To address the sparsity of scATAC-seq data by leveraging complementary scRNA-seq information.
• To identify biologically meaningful cell clusters integrating both transcriptomic and epigenomic data.

Main Methods:

• Propose Single-Cell Multi-omics Clustering (scMoC), a novel computational framework.
• Employ an imputation strategy to enhance sparse scATAC-seq data using less-sparse scRNA-seq data from the same cell.
• Merge individual clusterings derived from scRNA-seq and scATAC-seq data to achieve integrated cell populations.

Main Results:

• scMoC effectively imputes informative scATAC-seq data guided by scRNA-seq.
• The method generates biologically meaningful integrated clusters based on both RNA and ATAC profiles.
• scMoC demonstrates robust performance across datasets with varying sparsity levels and experimental protocols.

Conclusions:

• scMoC provides a powerful strategy for integrating scRNA-seq and scATAC-seq data.
• The RNA-guided imputation significantly improves the utility of sparse scATAC-seq data.
• This approach facilitates deeper understanding of cellular heterogeneity and function through multi-modal single-cell analysis.