Flow cytometry is a laboratory technique that rapidly analyzes thousands of cells, providing detailed information about their physical characteristics and protein expression. This process generates complex, multi-parameter datasets. Translating this wealth of information into a clear scientific narrative for publication is challenging due to the multi-parameter nature of the measurements. Effective figure preparation is necessary to communicate findings accurately and ensure the study’s reproducibility. This guidance provides technical standards for creating flow cytometry figures that meet the demands of scientific journals, ensuring clarity and transparency in data presentation.
Visualizing Single and Dual Parameters
Flow cytometry analysis begins by visualizing a single measured parameter or the relationship between two parameters. One-dimensional histograms are the standard choice for displaying the expression level of a single marker or for cell cycle analysis. These plots show the cell count on the vertical axis against the fluorescent signal intensity on the horizontal axis.
When comparing two characteristics, such as cell size (Forward Scatter, FSC) against internal complexity (Side Scatter, SSC), two-dimensional scatter plots are used. Proper presentation requires that compensation, the correction for spectral overlap between fluorophores, has been applied to accurately separate signals. The choice of plot—dot, density, or contour—depends on the number and density of cells analyzed.
Simple dot plots are suitable for lower cell counts. For high-density data, density plots (using color gradients) or contour plots (using lines) are preferred to manage visual clutter and prevent saturation. A uniform approach to axis scaling is necessary to represent the wide dynamic range of fluorescence signals accurately. Intensity data, which spans several orders of magnitude, must be displayed using a logarithmic or biexponential scale. Biexponential scaling is generally preferred because it handles the log-scale presentation of bright signals and the linear representation of data near zero, aiding in separating negative populations from the axis baseline.
Communicating Gating Strategy and Subpopulation Identification
Reproducibility requires transparently documenting the sequential selection process, known as gating, used to isolate specific cell populations. A comprehensive figure must visually trace the path from the initial collected events to the final population of interest, allowing researchers to understand and replicate the analytical strategy.
The gating process typically begins with excluding cellular debris and identifying single cells (singlets) to remove doublets and aggregates. Subsequent plots isolate viable cells, often by excluding those that take up a viability dye. These initial plots are foundational and serve as the “parent” populations for all downstream analysis.
Each subsequent plot must clearly indicate the population being analyzed, referred to as the “child” population derived from its parent. For instance, if the parent population is “Viable Single Cells,” the child plot might be gated on T-cells. This visual hierarchy is typically presented as a series of small, linked scatter plots. Publishing the complete gating tree is necessary, even if some steps are moved to supplemental materials, to ensure the reader can follow the logical flow. The figure legend must explicitly define every gate and the markers used for identification. This comprehensive approach transforms a complex analysis into an easily verifiable methodological statement.
Advanced Visualization of High-Dimensional Data
When modern flow cytometry or mass cytometry experiments use eight or more markers simultaneously, traditional two-parameter plots are insufficient for capturing data complexity. High-dimensional analysis techniques are employed to reduce multiple parameters into a manageable two-dimensional space for visualization. These methods reveal underlying data structures that are not apparent from simple bivariate plots, making them necessary for complex immunophenotyping.
Two accepted dimensionality reduction techniques are t-distributed stochastic neighbor embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP). These algorithms place cells with similar marker expression profiles closer together on the resultant map, forming distinct clusters that represent unique cell phenotypes. UMAP often preserves the global data structure better than t-SNE, which focuses more on local relationships.
When presenting these maps, the plot is a single scatter plot where each point represents an individual cell. The coordinates are determined by the algorithm, not by physical or fluorescent measurements. Color-coding is applied to convey additional information, such as marker expression intensity or grouping into phenotypically distinct clusters identified by automated algorithms (e.g., FlowSOM or PhenoGraph). The method used for defining cellular clusters must be clearly stated in the figure legend.
To fully characterize identified clusters, a secondary visualization is often paired with the t-SNE or UMAP map. A heatmap displays the average expression level of every measured marker across all distinct cell clusters. This provides a quantitative summary of the immunophenotype, allowing researchers to assign biological identity to the computationally defined groups. Spider plots offer an alternative visual summary, showing the relative expression profile of multiple markers for a single cluster on a radial graph.
Reporting Statistical Metrics and Summary Data
Quantitative summary data is needed to compare results across different experimental conditions. Flow cytometry analysis generates numerical metrics, such as cell frequency (percentage of a parent population) or Mean Fluorescence Intensity (MFI), which are used for statistical comparison.
Bar graphs and box plots are the standard formats for summarizing these metrics, allowing comparison between groups (e.g., treated versus untreated cells). Box plots are particularly informative because they display the distribution of the data, showing the median, quartiles, and outliers, which is preferable to bar graphs that only show the mean and standard deviation. Individual data points from biological or technical replicates should be overlaid on the summary graph to convey underlying variability.
The figure legend must precisely define the metric on the vertical axis, specifying whether the graph represents MFI, cell percentage, or fold change. All summary graphs must include clear notation of statistical significance. The results of statistical tests, often represented by P-values or asterisks, should be directly overlaid onto the graph to indicate meaningful differences between compared groups.