Fig. 6

Machine learning and high dimensional single-cell mapping and trajectory analysis identifies the most important targets of cardiomyocyte transplantation and hints at differences in macrophage development. A The differentially expressed transcripts between cell treated mice and untreated mice were compared using the random forest (RF) algorithm to identify the features that are important for cell transplantation (‘Cells ‘) in the heart. The most effective combination of targets is marked in gold. Positive correlation is marked with red circles and negative correlation is marked with green circles (only significant correlations are shown). B Violin plots of Mfge8 and Ccl7 expression levels in the various single-cell clusters of immune cells after MI in both C57BL/6 J and Rag2^del hearts. C DE genes between the macrophage clusters from Fig. 2B were used to generate hypothetical developmental relationships using Monocle algorithm. Each line represents individual branches in Monocle with the macrophage subsets occupying distinct branches based on their development. D The varied developmental kinetics of the various macrophage subsets between C57BL/6 J and Rag2^del mice is shown