Scattergram

The scattergram is a diagram, created with Add Diagram, that draws interactive two-measurement plots:

1. Axis selection. Choose any two measurements as the X and Y axes: marker intensities, area or derived values.
2. Population display. Each cell is a point at its (X, Y) coordinates. Coloring by phenotype, region or another classification brings out the population structure.
3. Heat map. When points crowd together, press Heat Map on the scattergram toolbar and the dots are converted into a color map. The heat map covers only the events within the visible axis range, so press Fit first to include them all.
4. Gate drawing. Rectangular, elliptical or polygonal gates can be drawn directly on the scattergram, each selecting the events inside it, and cutoffs on the two axes divide the plot into four quadrants.
5. Linked visualization. Selecting cells on the scattergram highlights their positions in the tissue image, and selecting them in the image highlights them on the plot. That ties a statistical pattern to a place in the tissue.

Bivariate analysis (Dilbilir): A scattergram shows the joint distribution of two variables. Each variable's histogram is a marginal distribution, and a clustering in the joint distribution can be invisible in both marginals. The joint distribution also shows correlation and non-linear association.

Colocalization metrics (Pawley): In confocal microscopy, scattergrams of two fluorescence channels are used to assess colocalization. Pearson's correlation coefficient R measures intensity correlation, from -1 to +1. Manders' coefficients M1 and M2 measure the proportion of each channel that overlaps the other.

Pawley's caution is that colocalization quantifies codistribution, not molecular interaction. At the microscope's resolution of about 200 nm, two molecules in the same voxel could be hundreds of nanometers apart. The same holds for tissue cytometry scattergrams: two markers co-expressed in one cell are both present, which does not mean the proteins interact.

The compensation analogy: In flow cytometry, spectral overlap between fluorophores puts false correlations into raw scattergrams: cells positive for only one marker appear to express both. Compensation (spectral unmixing) removes the artifact. Spectral unmixing does the same job in tissue cytometry, so a scattergram of unmixed signals shows true co-expression where a raw one may show bleed-through.

Identifying immune cell subtypes in a CD3-vs-CD8 scattergram, with CD3 on the Y axis and CD8 on the X axis:

• Upper right cluster (CD3+CD8+): cytotoxic T cells, high in both markers.
• Upper left cluster (CD3+CD8−): helper T cells, high CD3 and low CD8.
• Lower left cloud (CD3−CD8−): non-T cells, negative for both.
• Lower right (CD3−CD8+): rare events or artifact, such as NK cells or spectral contamination.

Cutoffs set at the valleys on each axis divide the plot into four quadrants and classify every cell at once. The linked tissue view then shows where each population lives: helper T cells in the stroma, cytotoxic T cells at the tumor border.