Here you can find the code to reproduce the differential expression (DE) analysis to replicate the figures of the paper: JIND: Joint Integration and Discriminationfor Automated Single-Cell Annotation.
- plot_DEwithtSNE.R: Runs a DE analysis between a correct classification cell type and the misclassified selected label. This script will produce as output two heatmaps with the results of the DE analysis, one with all the genes, and a second one with a few selection of genes. It also outputs the results of the DE analysis as a xlsx file and a tSNE with depicting the probabilities of the studied cells to belong to the correct label.
- plot_DE_NC.R: Runs a negative control DE analysis between the same selected type of cells, maintaining the same proportion of cells than in the misclassification DE analysis.
- plot_cfmt_allmethods.R: Computes the confusion matrices for all the methods studied on the paper.
- plot_cfmt_ggplot.R: Computes the confusion matrices for Seurat, JIND and JIND with rejection.