Fig. 1

treekoR helps to extract insight from cytometry data through deriving a hierarchy of cell clusters and measuring proportions to parent. a An example t-SNE plot showing clustering of single-cell data. b Hierarchical tree constructed using HOPACH algorithm on the cluster median marker expressions. c Definition of proportions to parent and proportions to all defined according to the organization of the hierarchical tree. d Significance testing is performed using both types of proportions calculated, testing for difference between the patient clinical endpoint of interest. e Visualization of the significance testing results. On the left, a scatterplot of each node in the hierarchical tree with the test statistic calculated using the %total (x-axis) vs. the test statistic calculated using the %parent (y-axis). On the right of the scatterplot, the hierarchical tree is colored with the test statistics: the nodes colored by the test statistic using %total and the branches of the nodes colored by the test statistic using %parent. An example of a corresponding node between the two graphs is highlighted in blue. The heatmap plots the median marker expression of the leaf nodes to assist in identification of the corresponding cell clusters