Fig. 6: Application of CAPITAL to cross-species scRNA-seq datasets of bone marrow cells.
From: Alignment of single-cell trajectory trees with CAPITAL
a A UMAP plot of clustering results along with a trajectory inferred by CAPITAL for Paul et al.'s data. An estimated trajectory is shown by a thick black line, where the solid lines indicate the linear trajectories that are compared in (c–e), whereas the dashed lines represent the others. b An alignment of the trajectory trees, where each pair of numbers in a node denotes the clusters shown in Fig. 5b and this figure (a). # denotes a space. c Heatmaps of known cell-type markers for erythrocyte, monocyte, and neutrophil with similar expression patterns in the datasets. These are ordered by pseudotime from left to right in each heatmap. Note that the markers in mouse are orthologous to the ones in human, and for clarity, these names in mouse are exactly the same as those in the human dataset. A gene name with an asterisk shows a marker that was not overlapped with the results of the computational screen (“Methods”). d Examples of different molecular patterns along pseudotime between human and mouse. e Heatmaps of enriched ontology terms across genes with different kinetics between human and mouse. They are colored by p-values computed by Metascape20, where the one-sided statistical tests based on the hypergeometric distribution were performed. The left panel indicates the terms derived from genes that showed an increasing tendency of expression in the human dataset and an decreasing tendency in the mouse dataset, and the right panel vice versa. UMAP uniform manifold approximation and projection, HSC hematopoietic stem cell, Eo/Baso eosinophil/basophil, Ery erythrocyte, Mega megakaryocyte, Mono monocyte, Neutro neutrophil.
