Fig. 5: DIU analyses with simulated and real data.

A Bar plot showing the DIU calling accuracy using simulated data with different methods. B Bar plot showing alternative isoform types for up-regulated and down-regulated DIU in Primed hESC compared with Naïve hESC detected using different methods. The analysis includes long-read RNA-seq datasets from Naïve and Primed hESC generated in this study, as well as NGS RNA-seq datasets from Naïve and Primed hESC. A3: alternative 3’ splice site, A5: alternative 5’ splice site, ATSS: alternative transcript start site, ATTS: alternative transcript terminated site, ES: exon skipping, IR: intron retention, MEE: mutually exclusive exon, MES: mutually exclusive splicing. C Bar plot showing the number of up-regulated and down-regulated DIU isoforms in Primed hESC with different switching consequences. NMD: nonsense-mediated decay, ORF: open reading frame. UpSet plot showing the number of overlapped DIU genes up-regulated (D) and down-regulated (E) in Primed hESC compared with Naïve hESCs identified by different methods and from the NGS data. F IGV screenshot displaying long-read RNA-seq coverages and splicing junctions of RPL39L gene isoforms in Naïve and Primed hESC. The gene model for different transcript isoforms of RPL39L is shown under the tracks. RPL39L-L and RPL39L-S represent isoforms on the gene reference, and RPL39L-UN represents the novel isoform detected in this study. Red arrows indicate primers used in RT-qPCR validation experiments. G Bar plot representing the LRS read proportions of the three isoforms of RPL39L in Naïve and Primed hESC. H Bar plot representing the RT-qPCR results of the three different isoforms of RPL39L. RT-qPCR experiments were conducted to evaluate the three different isoform levels of RPL39L in primed and naïve hESC. Data were gathered from three independent experiments and results were presented as mean ± standard deviation of fold change, with RPL39L-S in Primed hESC serving as the control (the raw data presented in Supplementary Data 4). Each experiment’s data was represented by dots. p values were determined using a two-sided two-sample T-test (*p < 0.05). Source data underlying (A–E, H) are provided as a Source Data file.