Fig. 4
From: diffcyt: Differential discovery in high-dimensional cytometry via high-resolution clustering
Results for experimental dataset Anti-PD-1. Results for re-analysis of experimental dataset Anti-PD-1 using method diffcyt-DA-edgeR; testing for differential abundance (DA) of cell populations between baseline samples from responder and non-responder groups of patients. a Heatmap shows phenotype (median arcsinh-transformed marker expression profiles) of significant detected clusters at 10% false discovery rate (FDR), compared to all cells (left panel); and relative cluster abundances (proportion of cells per cluster, by sample) (right panel) for the detected clusters. Heatmap rows (clusters) are grouped by hierarchical clustering with Euclidean distance and average linkage. b Boxplot shows total abundance (combined number of cells) for the clusters matching the phenotype of interest (clusters 317, 358, and 380), by sample and group. Runtime was 32.0 s, on a 2014 MacBook Air laptop, 1.7 GHz processor, 8 GB memory, using a single processor core. NR non-responders, R responders
