Supplementary figure S3.

Inter-dataset prediction using aligned reference and query matrix. A. Inter-dataset accuracy comparison, shown as heatmaps of three classification metrics (overall accuracy, ARI, and V-measure). B. Boxplots illustrating the averaged metric scores before and after alignment for each method. The x-axis is the methods, and the y-axis is the classification metrics. For each of the four pairs of experimental data used in inter-dataset evaluation, we aligned both reference and query data using CCA from the Seurat data integration function. Then we separated aligned datasets and performed the inter-dataset evaluation again.