Con-AAE: contrastive cycle adversarial autoencoders for single-cell multi-omics alignment and integration.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• The multi-omics era enables comprehensive cellular analysis by integrating diverse data types.
• Single-cell multi-omics present challenges due to data sparsity, high dimensionality, and noise, particularly in simultaneous scRNA-seq and scATAC-seq measurements.

Purpose of the Study:

• To develop a novel framework for aligning and integrating single-cell RNA-seq and single-cell ATAC-seq data.
• To address the challenges of sparsity, high dimensions, and noise in single-cell multi-omics data.

Main Methods:

• Proposing a novel framework: contrastive cycle adversarial autoencoders (Con-AAE).
• Utilizing adversarial autoencoders for data integration and alignment.
• Mapping data from different spaces into a coordinated subspace for easier analysis.

Main Results:

• Con-AAE efficiently maps sparse and noisy scRNA-seq and scATAC-seq data.
• The framework facilitates alignment and integration of single-cell multi-omics data.
• Demonstrated advantages of Con-AAE on several benchmark datasets.

Conclusions:

• Con-AAE provides an effective solution for integrating single-cell multi-omics data.
• The framework enhances the ability to derive comprehensive biological insights from multi-modal single-cell data.
• The proposed method overcomes key technical hurdles in single-cell multi-omics integration.