CyAnno

CyAnno annotates cell types in CyTOF (Cytometry by Time-of-Flight) mass cytometry datasets by applying machine learning to integrate manually gated and 'ungated' single-cell populations and thereby improve accuracy of immune-system profiling at single-cell resolution.

Key Features:

• Semi-automated annotation: Combines manual gating information with automated classification to produce cell type annotations for mass cytometry data.
• Machine learning deconvolution: Leverages machine learning models to deconvolute unlabeled cytometry datasets.
• Use of manually gated training data: Trains models on manually gated cell populations to inform classification.
• Integration of 'gated' and 'ungated' cells: Models both previously identified ('gated') and heterogeneous undefined ('ungated') live-cell populations within the same framework.
• Improved gated-cell prediction accuracy: Enhances precision of predicting annotated ('gated') cell types compared to prior semi-automated approaches.
• Rare/single cell type detection: Demonstrates improved identification of single or rare cell types relative to state-of-the-art semi-automated methods.
• CyTOF-focused: Specifically developed and validated for CyTOF (Cytometry by Time-of-Flight) mass cytometry datasets.

Scientific Applications:

• Cell type annotation in mass cytometry: Assigns cell type labels in CyTOF datasets by integrating manual gating and automated classification.
• Immune system monitoring at single-cell resolution: Supports immune profiling in large-scale studies by improving single-cell annotation accuracy.
• Rare cell population identification: Facilitates detection and characterization of rare or single-cell phenotypes within complex cytometry samples.

Methodology:

Applies machine learning models trained on manually gated data to integratively model 'gated' and 'ungated' cells and deconvolute unlabeled CyTOF datasets, with validation reported across multiple CyTOF datasets.