ConvexGating infers gating strategies from clusters in single cell cytometry data

Manual gating remains the standard for identifying specific cell populations in flow cytometry and related technologies, but it is prone to subjectivity and inconsistency, particularly in high-dimensional or multi-center studies. This project introduces and validates ConvexGating, a machine learning-based tool that learns gating strategies in a data-driven, interpretable, and reproducible manner. The approach is applied to flow cytometry, CyTOF, and CITE-seq datasets to identify both known and previously ill-defined cell types with minimal contamination. Key populations such as CD8+ TEMRAs and plasmacytoid dendritic cells (pDCs) are examined. The goal is to improve consistency in single-cell analysis and enable more effective design of marker panels for sorting.

• Type: Transcriptome Sequencing
• Archive: European Genome-phenome Archive (EGA)