Fig. 3
From: Robust principal component analysis for accurate outlier sample detection in RNA-Seq data
Comparing the performance of cPCA and rPCA on the mouse cerebellum data. a-c cPCA plot of the mouse cerebellum data set with three possible scenarios to achieve separation between groups: (a) removal WT-1, KO-1 and KO-3; (b) removal WT-1 and WT-6; (c) removal WT-1 and KO-4. Arrows point to candidate outlier samples need to be removed. d-f Outlier maps of the mouse cerebellum data set using (d) cPCA, (e) PcaHubert and (f) PcaGrid. g Relative log expression (RLE) plot of the mouse cerebellum data before removing outliers WT-1 and KO-3, (h) RLE plot after removing outliers and (i) cPCA plot after removing outliers. Black line in (a, b, c, i) indicates the line that separates WT and SnoN KO samples