Fig. 1: Validation of CS-CORE using permuted snRNA-seq data from ref. 17.
From: Cell-type-specific co-expression inference from single cell RNA-sequencing data
Results from permuted data with varying and constant sequencing depths are colored with light red and blue, respectively. a Scatter plots with fitted curves showing mean expression levels (x-axis) and average co-expression ( y-axis) of each gene with co-expression estimated using baredSC, locCSN, Noise Regularization, Normalisr, Pearson correlation of log normalized data (Pearson), propr, Spearman correlation of log normalized data (Spearman), SpQN, ρ-analytic PR, ρ-sctransform and CS-CORE. Average co-expressions are re-scaled by the maximum value to aid comparison. The mean expression levels are plotted at the scale of \({\log }_{10}{\mu }_{j}+3\) for μj defined in Eq. (1). b Q-Q plots comparing p values for testing co-expressions of gene pairs against Uniform(0,1) using seven methods with statistical tests, including Noise Regularization, Normalisr, Pearson correlation of log normalized data (Pearson), Spearman correlation of log normalized data (Spearman), ρ-analytic PR, ρ-sctransform and CS-CORE. Source data are provided as a Source Data file.
