Fig. 3: SCALPEL predictions on iPSC and NPC data can be experimentally validated and reflect miRNA function.

a Clustering analysis identifies two main populations corresponding to iPSCs (salmon) and NPCs (cyan). b Distribution of the number of UMIs, Genes and % of mitochondrial UMIs (MT) of the two samples analyzed. iPSC violin plots show measurements for 1233 cells and the NPC violin plots show measurements for 1302 cells. Overlaid boxplots indicate the median (center line), the 25th and 75th percentiles (box limits), and the most extreme values within 1.5 times the IQR (whiskers). iPSCs: median = 7047, box = [4166, 11,495], whiskers = [597, 19,973]; NPCs: median = 2952, box = [1513, 5771], whiskers = [414, 19,975]. c SCALPEL identifies a significant change in isoform usage in EIF1 gene between iPSCs and NPCs. Coverage plots show the distribution of filtered reads along isoforms in both conditions. Relative expression of isoform abundance by SCALPEL is provided on the custom tracks. The location of the oligo dT primer (gray) and the gene-specific primer (purple) used for experimental validations are shown on top of the isoform diagrams. d Experimental validation in bulk of the isoform changes predicted by SCALPEL using 3’ RACE. EIF1-204 isoform is more highly expressed in iPSCs than in NPCs while the expression of the EIF1-201 isoform is similar in both conditions. e Quantification of JPT1 isoforms by SCALPEL. f Relative abundance of JPT1−203 in NPCs is higher than in iPSCs while JPT1-207 is more similar across conditions. g Cumulative distribution plot of log2FCs showing the difference in the expression of isoforms from DIU genes where one of them contains miR-128-5p target sites (pink) and the other does not (gray), indicating that changes in isoform usage between iPSCs and NPCs can be attributed to miRNAs known to be implicated in neurogenesis (two-tailed Kolmogorov–Smirnov test, FDR = 0.0000115). Source data are provided as a Source Data file.