Fig. 2: A gene-wise test for differential isoform expression (DIE) is more sensitive than an exon-wise test at detecting splicing changes.

a Schematic of the scisorseqR approach—Barcode deconvolution, filtering, pairwise comparison, and reporting of significant results based on FDR and ΔΠ cutoffs (created with BioRender.com). b Volcano plot of bulk HIPP vs. bulk PFC differential abundance analysis, with the effect size (ΔΠ) on the X-axis. P-values derived from a χ² test and corrected for multiple testing using the Benjamini–Hochberg correction are plotted on the Y-axis. Points are colored according to the levels of significance based on FDR and ΔΠ value. Genes considered significant (pink) when FDR ≤ 0.05 and |ΔΠ| ≥ 0.1. c Scatter plot showing the ΔΨ of all exons for genes that show significant DIE between HIPP and PFC. Gray points represent non-significant exons. Points are colored according to the cell-type in which an exon is considered significant by a BY corrected p-value and a ΔΠ ≥ 0.1. Diagonal lines indicate cutoff of 0.1 ΔΨ. d Venn diagram showing the overlap of genes significant by DIE (BH correction) with genes significant by exon tests (BY correction). e Boxplot showing the maximum absolute value of ΔΨ per gene in three different categories: genes that are not significant by DIE tests (n = 1395), genes that are significant by DIE tests and also exhibit differential TSS or polyA-site usage (n = 38), and genes that are significant by DIE and do not exhibit differential TSS or polyA-site usage (n = 128). P-values were calculated using a two-sided Wilcoxon rank sum test. Data are represented as boxplots where the middle line is the median, the lower and upper hinges correspond to the first and third quartiles, the upper whisker extends from the hinge to the largest value no further than 1.5× IQR from the hinge (where IQR is the inter-quartile range) and the lower whisker extends from the hinge to the smallest value at most 1.5× IQR of the hinge. Please see function geom_boxplot in R (ggplot2). f Percentage of novel transcripts by scisorseqR that were manually validated as being novel by Gencode team, and breakdown of predicted function. g Heatmap of significant DIE genes (n = 395, BH correction) according to entire isoform between bulk HIPP and bulk PFC that also exhibit differential usage of transcription start site (TSS) and polyA-site (PolyA). Each row is a single gene. Gray represents genes that are non-significant by category (Iso/TSS/PolyA) whereas purple represents significant by category. h Isoform expression of Nsfl1c gene. Each horizontal line in the plot represents a single transcript colored according to the cell-type it is represented in. Therefore, blocks represent exons and whitespace represents intronic space (not drawn to scale). Orange exon represents alternative inclusion. i Isoform expression for the Nsmf gene. Each row represents an isoform colored by Π and each column represents a cell-type in HIPP or PFC.