DendroSplit

DendroSplit provides interpretable clustering of single-cell RNA-Seq (scRNA-Seq) datasets to identify biologically meaningful cell populations.

Key Features:

• Interpretability: Emphasizes interpretability, providing a framework to understand and justify clustering results from scRNA-Seq data.
• Addressing Subjectivity: Implements a structured approach intended to reduce subjectivity and enhance reproducibility of clustering outcomes.
• Feature selection and multi-level populations: Integrates feature selection to identify multiple levels of biologically meaningful populations aligned with definitions of "cell type".
• Computational Efficiency: Demonstrates computational efficiency suitable for large-scale single-cell datasets.
• Comparative Performance: Achieves clustering accuracy and speed comparable to existing methods while prioritizing interpretability.

Scientific Applications:

• Single-cell genomics: Applied to analyze cellular heterogeneity in scRNA-Seq datasets.
• Cell type and state identification: Used to identify distinct cell types and cell states from scRNA-Seq data.
• Developmental biology: Facilitates resolution of cell-type heterogeneity in developmental studies.
• Cancer research: Supports investigation of cellular heterogeneity relevant to cancer biology and disease mechanisms.
• Immunology: Enables resolution of immune cell types and states in immunological studies.

Methodology:

Integrates feature selection into the clustering process to uncover biologically meaningful clusters at multiple levels.