Methods for discovery and characterization of cell subsets in high dimensional mass cytometry data.

摘要: The flood of high-dimensional data resulting from mass cytometry experiments that measure more than 40 features individual cells has stimulated creation new single cell computational biology tools. These tools draw on advances in the field machine learning to capture multi-parametric relationships and reveal are easily overlooked traditional analysis. Here, we introduce a workflow for high dimensional emphasizes unsupervised approaches visualizes both population level views. This includes three central components common across analysis approaches: (1) distinguishing initial populations, (2) revealing subsets, (3) characterizing subset features. In implementation described here, viSNE, SPADE, heatmaps were used sequentially comprehensively characterize compare healthy malignant human tissue samples. use multiple methods helps provide comprehensive view results, largely facilitates automation researchers avoid missing populations with unusual or unexpected phenotypes. Together, these develop framework future identity.