Visualizing different types of matrices
The low-dimensional plots can show information stored in multiple types of matrices. To enable this, the relevant elements are associated with the plot by dragging and dropping them in the corresponding groups of the Side Panel:
- Expression Matrix (
): `Expression data' group;
- Expression Matrix with spliced and unspliced counts (
): `Expression data' group;
- Peak Count Matrix (
): `Peak count matrix' group;
- Velocity Matrix (
): `Velocity matrix' group.
Note that only one element at a time can be associated with each group, but all groups can have associations concurrently.
The Expression Matrix with spliced and unspliced counts () shares the same group as the Expression Matrix (
) because it is an extension which additionally contains separate information about the spliced and unspliced counts for each cell and gene.
The Manual Annotation section showcases how the expression of genes found in an Expression Matrix () can be used for coloring and selecting cells (see Manual annotation). The remaining matrices can display:
- Multiple types of expression values for the same feature (figure 17.21);
- Multiple types of features (figure 17.22);
- Other type of information (figure 17.23);
all of which can be chosen from a drop down menu.
Figure 17.21: The options available in the `Expression data' group for an Expression Matrix with spliced and unspliced counts.
Figure 17.22: The options available in the `Peak count matrix' group for a Peak Count Matrix.
Figure 17.23: The options available in the `Velocity matrix' group for a Velocity Matrix.
The same functionality for coloring and selecting cells available for expression data, as detailed in the Manual Annotation section, is also available for all the options in the drop down menus.
The transition probabilities present in the Velocity Matrix (figure 17.23) are the only options that are cell-based and not feature-based. To color cells after such probabilities, one cell can be selected in the plot using the lasso tool (see figure 17.5) and the option to show the probabilities can be chosen from the plot right-click menu (figure 17.24 and 17.25).
Figure 17.24: Showing transition probabilities from the plot right-click menu. Options are available when just one cell is selected.
Figure 17.25: Coloring by incoming transition probabilities from a selected cell
The `Incoming probability' colors cells that have a non-zero probability of transitioning towards the selected cell by that probability. The `Outgoing probability' colors cells that the selected cell has a non-zero probability of transitioning towards by that probability. Note that probabilities can be negative, indicating that the cell is transitioning in the opposite direction.
When a Velocity Matrix is associated with a Dimensionality Reduction Plot, the `Show arrows' option can be used to display the projected velocities at cellular level. Each arrow is obtained from the cell's `Outgoing probabilities' and summarizes them into a projected direction for the cell and its speed of movement. The arrows can reveal differences between near-terminal cells, where arrows are short, and transient cells, where arrows are longer (figure 17.26).
Figure 17.26: UMAP plot of the pancreas data set [Bastidas-Ponce et al., 2019] built-in scVelo [Bergen et al., 2020]. Arrows show the direction and speed of movement of an individual cell. The real time cells experience as they differentiate is approximated by the latent time, shown here in the 0 (black) to 1 (yellow) range.