Two numbers per dot: which cell types express a gene, and how strongly? The dot plot encodes both in a single matrix.
Once you have clusters and candidate genes, the next question is which cell types actually express each gene. A heatmap encodes one quantity per cell. In single-cell data, that average is misleading: a high mean can mean every cell expresses slightly, or a few express intensely while most stay silent.
The dot plot encodes two quantities per dot. Size is the fraction of cells expressing the gene; color is the average expression for the whole cell group, often scaled within each gene before plotting. The pair distinguishes broadly expressed housekeeping genes from sporadic but intense markers. Hovering here also reports the mean among positive cells, because that hidden number explains many ambiguous dots.
Each gene below is a known marker for a population. Look for TP53 (broadly expressed tumor suppressor) and CDKN1A (its downstream target, which tracks TP53 but varies by how vigorously each cell type activates the p53 pathway).
Figure 1. Dot plot of ten genes across eight cell types. Dot size encodes the fraction of cells expressing the gene. Color intensity encodes average expression across the group; the hover readout separates that from the mean among positive cells.
The R175H mutation in TP53 produces a misfolded protein that cannot activate its transcriptional targets. The mRNA is still transcribed and may accumulate, so single-cell RNA-seq shows TP53 mRNA at normal or elevated levels while its functional output collapses.
The dot plot compares wildtype and TP53-mutant tumor cells. CDKN1A drops in both fraction expressing and mean level across most cell types in the mutant. TP53 mRNA itself stays present, masking the downstream collapse.
Figure 2. Comparing wildtype and mutant tumors. CDKN1A shows reduced expression across most cell types in the mutant, even though TP53 levels appear similar.