scDIG: an R Shiny application for interactive density-based gating of single-cell proteomic and transcriptomic data

Delineating biologically meaningful cell populations within single-cell embedding spaces requires methods that balance expert guidance with reproducibility. Current approaches range from subjective manual selection to fully automated algorithms that may obscure rare but biologically significant populations. We present scDIG, a Shiny-based tool that integrates bimodal index-driven feature selection, feature-weighted kernel density estimation, and interactive contour-based gating to define cell populations directly within two-dimensional projections of scRNA-seq and CITE-seq data. Applied to CITE-seq PBMC data from subjects in the Cardiovascular Assessment Virginia (CAVA) cohort, scDIG resolved transcriptionally distinct CD4⁺ T cell subpopulations within continuous embeddings not readily captured by conventional clustering and identified GZMB -enriched cytotoxic lymphocyte subsets from RNA expression data. Sensitivity analysis confirmed that defined gates correspond to coherent cell populations in high-dimensional feature space. s cDIG provides a robust, reproducible framework for density-based gating of single-cell data, bridging the interpretability of manual expert gating with the rigor of algorithmic population definition. The app is freely available for non-commercial use at https://au-cbgm-shiny.augusta.edu/gating , with source code at https://gitlab.com/pbombina/scdig .