CAbiNet: joint clustering and visualization of cells and genes for single-cell transcriptomics

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• Single-cell transcriptomics analysis requires clustering cells and identifying marker genes.
• Current methods often separate these steps, inferring genes after cell clustering.
• Biclustering algorithms can jointly cluster cells and genes but struggle with large single-cell datasets.

Purpose of the Study:

• To introduce an efficient biclustering method for joint cell and gene clustering in single-cell RNA sequencing data.
• To enable integrated visualization of cell clusters and their associated marker genes.
• To overcome scalability limitations of existing biclustering approaches for large datasets.

Main Methods:

• Developed 'Correspondence Analysis based Biclustering on Networks' (CAbiNet).
• Employed correspondence analysis for efficient bicluster identification.
• Integrated network analysis to refine bicluster structure.
• Implemented non-linear embedding for joint visualization of biclusters.

Main Results:

• CAbiNet achieves efficient co-clustering of cells and marker genes.
• The method provides joint visualization of biclusters within a non-linear embedding space.
• Demonstrated effectiveness in handling the scale of single-cell RNA sequencing data.

Conclusions:

• CAbiNet offers a powerful tool for integrated analysis and visualization of single-cell RNA sequencing data.
• Facilitates interactive exploration and understanding of cell populations and their defining genes.
• Represents a significant advancement in scalable biclustering for transcriptomics.