Fig. 2

Most spatial transcriptomics datasets display overdispersion. A The proportion of genes with variance greater than mean (y-axis) across datasets (x-axis). B The median ratio of variance to mean (y-axis) across datasets (x-axis), where the horizontal line y = 1 shows where the variance equals the mean. C The log transformed variance for each gene (y-axis) is plotted against the expression mean for datasets from three techniques: seqFISH+, Slide-seqV2, and 10x_MB(C). D Mean square error (MSE) for the estimated log transformed variance (y-axis) based on either the Poisson model (blue line) or the negative binomial model (red line) across datasets (x-axis). In A, B, and D, the gray dotted line separates smFISH-based technologies from the sequencing-based technologies and solid line separates single-cell resolution technologies to spot-level resolution technologies