Supplementary Figure 3: Comparing trendsceek results with the identification of variable genes. | Nature Methods

Supplementary Figure 3: Comparing trendsceek results with the identification of variable genes.

From: Identification of spatial expression trends in single-cell gene expression data

Supplementary Figure 3

Trendsceek was run on the 500 most variable genes, indicated in yellow. Genes deemed significant by trendsceek (by at least one of the four statistical tests, Benjamini-Hochberg adjusted P < 0.05) are indicated in red. Blue indicates smoothed gene density histogram from all genes. Analyzed datasets were: (A) spatial transcriptomics mouse olfactory bulb (tissue section “replicate 3”, n=269 array-spots), (B) spatial transcriptomics mouse olfactory bulb (tissue section “replicate 12”, n=280 array-spots), (C) spatial transcriptomics breast cancer biopsy (layer 2, n=251 array-spots), and (D) mouse gastrulation single-cell RNA-seq data (n=481 cells). (A-C) The red line illustrates a poisson distribution fit without any overdispersion. (D) The solid red line illustrates a negative binomial fit and the dashed line marks the expected position of genes with biological coefficient of variation = 0.5.

Back to article page