Commercial flow cytometry software, such as Cytobank, also provide automated pipelines, such as CITRUS ( 19). Furthermore, specific characteristics of spectral flow cytometry, such as removal of auto-fluorescence per cell, higher maximum fluorescence intensities and minimal requirement for spectral compensation, also hinder easy use of already published workflows for data analysis in mass cytometry or conventional flow cytometry ( 14, 15).A few groups published advanced workflows also suited for spectral flow cytometry ( 16– 18). Not knowing where to start, new analysis strategies can be intimidating to work with and mistakes are easily made.
Flowjo acquired by bd manual#
Some of these tools were also integrated in current flow cytometry software, such as FlowJo™ software and Cytobank, or new visualization software, such as Cytosplore ( 12, 13).ĭespite the need for automated analyses, laboratories specialized in flow cytometry often lack bioinformatics expertise and tend to stick to traditional manual gating. This is useful for statistical testing comparing median expression of number of cells per cluster and/or group, which is not possible in standard t-SNE or UMAP. Clustering algorithms, such as spanning-tree progression analysis of density-normalized events (SPADE) ( 10) and flow cytometry self-organizing map (FlowSOM) ( 11), can visualize the data, but also cluster data. These algorithms reduce dimensionality of the data and enable easy visualization of high-dimensional data. Examples of such analyses are t-distributed stochastic neighbor embedding (t-SNE) ( 7), hierarchical stochastic neighbor embedding (h-SNE) ( 8), and uniform manifold approximation and projection (UMAP) ( 9). In recent years, automated analysis methods have been proven useful in multicolor flow cytometry as well as mass cytometry single cell data ( 5, 6). Therefore, unbiased exploration of data using automated analyses can help find unknown cell populations and can help compare cell populations between groups in a more reproducible way. Moreover, manual gating strategies limit exploration of the data. Therewith, traditional 2D manual gating strategies fail to comprehend and depict the entirety of the data, since the number of 2D plots increases exponentially with the number of parameters measured. With the increase in number of markers, especially in spectral flow cytometry datasets, flow cytometry data have become high-dimensional. However, easy-to-use data analysis workflows for spectral flow cytometry, for the starting researcher in this field, are currently limited. Adherence to established general guidelines for key practical aspects and data analysis will help to increase reproducibility ( 4). With the current complexity of flow cytometry assays, reproducibility is a major concern. Even in conventional flow cytometry, technologies are used to increase the number of markers included in one panel ( 3). Up to date, panels of over 40 colors have been developed ( 1, 2), with larger panels expected to emerge in the near future. Not being limited to the number of channels of the instrument, spectral flow cytometry enables multicolor panels with many more parameters than ever deemed possible in conventional flow cytometry. Over the years, the number of variables measured in flow cytometry experiments has increased, especially with the recent development of spectral flow cytometry. Application of our workflow will aid spectral flow cytometry users to obtain valid and reproducible results. This methods article provides an R-based pipeline based on previously published packages, that are readily available to use. Using healthy control data as example, we will go through the concepts of quality control, data cleaning, transformation, correcting for batch effects, subsampling, clustering and data integration. Moreover, we will describe a workflow to properly prepare spectral flow cytometry data for high dimensional analysis and tools for integrating new data at later time points. In this article, we will give insight into the pitfalls of handling spectral flow cytometry datasets. However, preparing spectral flow cytometry data for high-dimensional analysis can be challenging, because of several technical aspects.
To fully explore the resulting high-dimensional single cell datasets, high-dimensional analysis is needed, as opposed to the common practice of manual gating in conventional flow cytometry. Spectral flow cytometry is an upcoming technique that allows for extensive multicolor panels, enabling simultaneous investigation of a large number of cellular parameters in a single experiment. 3Department of Clinical Immunology and Rheumatology, Maasstad Hospital, Rotterdam, Netherlands.2Department of Immunology, Erasmus University Medical Center, Rotterdam, Netherlands.1Department of Rheumatology, Erasmus University Medical Center, Rotterdam, Netherlands.Hannah den Braanker 1,2,3†, Margot Bongenaar 1,2† and Erik Lubberts 1,2*
0 kommentar(er)
