HGC: fast hierarchical clustering for large-scale single-cell data.

Area of Science:

• Computational biology
• Genomics
• Bioinformatics

Background:

• Single-cell data analysis requires robust clustering to identify cellular heterogeneity.
• Existing methods often lack hierarchical information or fail to scale to large datasets.
• Classical hierarchical clustering (HC) is computationally intensive for large-scale single-cell studies.

Purpose of the Study:

• To introduce HGC, a novel Hierarchical Graph-based Clustering tool.
• To address the limitations of existing single-cell clustering methods regarding hierarchy and scalability.
• To enable multiresolution exploration of biological hierarchies in single-cell data.

Main Methods:

• HGC integrates graph-based clustering with hierarchical clustering principles.
• It constructs a hierarchical tree on the shared nearest-neighbor graph of cells.
• The method achieves linear time complexity for hierarchical tree construction.

Main Results:

• HGC enables multiresolution exploration of biological data hierarchies.
• The tool demonstrates state-of-the-art accuracy on benchmark single-cell datasets.
• HGC exhibits excellent scalability for analyzing large single-cell datasets.

Conclusions:

• HGC offers an efficient and accurate solution for single-cell data clustering.
• The method successfully captures hierarchical structures within cellular populations.
• HGC is a valuable tool for advancing single-cell data analysis and biological discovery.