Fig. 2

Validation of SQANTI-SIM approach. a Simulated transcripts for each novel SQANTI3 structural category out of 1000 requested for each type (red line). b Comparison of the distribution of expressed isoforms per gene in real human WTC11 cell line data and simulated data. c Scatter plot comparing simulated transcript reads and mapped raw counts, using triangles for transcripts where simulated reads equal or exceed mapped counts, and circles otherwise. d Quantile-quantile (Q-Q) plot comparing the expression levels, represented as log raw counts, between known and novel transcripts. e Proportion of simulated transcripts with all splice junctions (SJ) covered by sample-specific short-read (SR) data at different simulated sequencing depths. f Bivariate distribution of CAGE peak length and distance to the closest transcript transcription start site (TSS) for CAGE peaks supporting a TSS and those not supporting any TSS in real WTC11 and simulated data