Extended Data Fig. 6: Comparison between BayesPrism and the two different modes of expression inference by CIBERSORTx. | Nature Cancer

Extended Data Fig. 6: Comparison between BayesPrism and the two different modes of expression inference by CIBERSORTx.

From: Cell type and gene expression deconvolution with BayesPrism enables Bayesian integrative analysis across bulk and single-cell RNA sequencing in oncology

Extended Data Fig. 6

Scatter plot shows Spearman’s correlation (left), and Pearson’s correlation (right), between gene expression estimated by BayesPrism (red), total bulk (blue), CIBERSORTx group mode (orange) or CIBERSORTx high resolution mode (purple) and the average expression from malignant cells in scRNA-seq as a function of the fraction of malignant cells in the dataset containing the 270 simulated samples. The correlation coefficient was calculated on 53 imputable genes by the high-resolution mode out of the top 1000 most variable genes in malignant cells. Spearman’s correlation coefficients were calculated on untransformed gene expression values, while Pearson’s correlation coefficients were calculated on variance-stabilizing transformed gene expression values.

Back to article page