Mass cytometry is an emerging technology capable of 40 or more correlated measurements on a single cell. The complexity and volume of data generated by this platform have accelerated the creation of novel methods for high-dimensional data analysis and visualization. A key step in any high-level data analysis is the removal of unwanted events, a process often referred to as data cleanup. Data cleanup as applied to mass cytometry typically focuses on elimination of dead cells, debris, normalization beads, true aggregates, and coincident ion clouds from raw data. We describe a probability state modeling (PSM) method that automatically identifies and removes these elements, resulting in FCS files that contain mostly live and intact events. This approach not only leverages QC measurements such as DNA, live/dead, and event length but also four additional pulse-processing parameters that are available on Fluidigm Helios™ and CyTOF® (Fluidigm, Markham, Canada) 2 instruments with software versions of 6.3 or higher. These extra Gaussian-derived parameters are valuable for detecting well-formed pulses and eliminating coincident positive ion clouds. The automated nature of this new routine avoids the subjectivity of other gating methods and results in unbiased elimination of unwanted events. © 2019 International Society for Advancement of Cytometry.
Keywords: Gaussian parameters; probability state Modeling; quality control; unattended analysis.
© 2019 International Society for Advancement of Cytometry.