Integrative analysis of single-cell genomics data by coupled nonnegative matrix factorizations.

Area of Science:

• Computational Biology
• Genomics
• Data Science

Background:

• Heterogeneous cell populations require sophisticated analysis methods.
• Integrating multiple functional genomics datasets from single cells presents analytical challenges.
• Clustering of cells across different samples should ideally be consistent.

Purpose of the Study:

• To address the challenge of coupled clustering in single-cell genomics.
• To develop a computational method for integrative analysis of diverse single-cell data.
• To improve the analysis of heterogeneous cell populations by linking clusters across samples.

Main Methods:

• Formulation of the coupled clustering problem as an optimization task.
• Development and application of coupled nonnegative matrix factorizations (coupled NMF).
• Integrative analysis of single-cell RNA-sequencing (RNA-seq) and single-cell ATAC-sequencing (ATAC-seq) data.

Main Results:

• Demonstrated the effectiveness of coupled NMF for integrative single-cell data analysis.
• Showcased the ability of coupled NMF to achieve consistent cell clustering across datasets.
• Successfully applied the method to combined RNA-seq and ATAC-seq data from heterogeneous cell populations.

Conclusions:

• Coupled NMF provides a robust framework for integrating single-cell functional genomics data.
• The proposed method enhances the analysis of complex, heterogeneous biological systems.
• This approach facilitates more accurate and biologically relevant cell population characterization.