scDIG: An R Shiny Application for Interactive Density-Based Gating of Single-Cell Proteomic and Transcriptomic Data.

Area of Science:

• Single-cell genomics
• Immunology
• Bioinformatics

Background:

• Delineating cell populations in single-cell data is challenging.
• Current methods may not balance expert input with reproducibility.
• Identifying biologically meaningful cell types requires advanced analytical tools.

Purpose of the Study:

• To present scDIG, a Shiny-based tool for cell population delineation.
• To enable reproducible classification of cell types in single-cell data.
• To identify immunologically relevant cell states.

Main Methods:

• scDIG integrates bimodal index-driven feature selection.
• It uses feature-weighted kernel density estimation.
• Interactive contour-based gating is applied to 2D projections of scRNA-seq and CITE-seq data.

Main Results:

• scDIG was applied to CITE-seq PBMC data from the CAVA cohort.
• It resolved transcriptionally distinct CD4+ T cell subpopulations.
• These subpopulations were within continuous embeddings not captured by conventional clustering.

Conclusions:

• scDIG facilitates robust and reproducible classification of single-cell populations.
• The tool aids in identifying immunologically relevant effector states.
• scDIG enhances the analysis of complex single-cell datasets.