Fig. 5
From: scAlign: a tool for alignment, integration, and rare cell identification from scRNA-seq data
Alignment performance is robust to imbalance in cell type representation across conditions. a Accuracy of classifiers on the Kowalczyk et al. benchmark, when removing either LT-HSC, ST-HSC, or MPP cells from the old condition. scAlign outperforms all other methods and exhibits minimal degradation in performance as increasing numbers of cells are removed within each cell type. b Heatmap showing the pairwise similarity matrix for the young cells from Kowalczyk et al. when no cells have been removed. c Heatmap showing the pairwise similarity matrix for the young cells from Kowalczyk et al. after removing 25% of the old mouse cells from all cell types