Fig. 3: Visualizations of stability of pseudo-design datasets.
From: A realistic FastQ-based framework FastQDesign for ScRNA-seq study design issues
a–c The UMAPs of the FastQ sample with 10% read depth of all cells, the UMI sample with 10% read depth of all cells, 10% cell and total read depth, were colored with the consistency of cell clustering match with the reference. d–f The pairwise minimal p-value's of each gene among the clusters from the samples and the reference, x is \(-\log 10(p\_val\_adj)\) in the reference, y is the same value in the sample, only the significant DE genes in either reference or samples are shown. g–i The pairwise minimal p-value's of each gene between the condition within the cluster from the samples and the reference, only the significant DE genes in either reference or samples are shown. j, k The impact of varying read depth, and cell numbers on ARI and Jaccard indexes, 10 simulations were conducted at each setting.
