Fig. 2: Identification of novel IPA sites in HEK293 cells using InPACT.

a, b Two examples of InPACT-identified IPA sites in HEK293 cells. The Sashimi plots depict the RNA-seq reads aligned to the ZNF771 (a) and TERF2 loci (b), with the annotation of Refseq isoforms and InPACT-identified novel IPA isoforms. The densities of RNA-seq reads from two RNA-seq replicates of HEK293 cells are shown in purple. Splice junctions are displayed as arcs connecting exons. The number of reads observed for each junction is indicated within the arc. c, d Experimental validation of InPACT-identified IPA sites in ZNF771 (c) and TERF2 (d) genes from HEK293 cells. The gel of 3’-Rapid Amplification of cDNA Ends (3’-RACE) experiments and Sanger sequencing results of the amplified transcripts by 3’-RACE experiments were depicted. Each experiment was repeated n = 3 times. e The bar plot depicts the frequencies of 18 known polyA signals detected in the regions of 60 nt upstream of the respective sites. The IPA sites identified from two RNA-seq replicates of HEK293 are compared with randomly selected genomic sites in introns. f, g The cumulative distribution curves of terminal exon length (f) and conservation scores (PhyloP) (g). The IPA terminal exons identified from two RNA-seq replicates of HEK293 are compared with annotated terminal exons or randomly selected genomic sites in introns. h The scatter plot depicts the estimated expression levels of annotated isoforms (blue) and InPACT-identified novel IPA isoforms (red) from two RNA-seq replicates of HEK293. The Pearson correlation coefficients are indicated in the respective colors. i The plots depict the read coverage of ribo-seq reads around stop codons of annotated isoforms and IPA isoforms. The upper panels depict the average read coverage. The lower panels show the read coverage for each isoform using heatmaps. j The violin plot depicts translational efficiencies of annotated terminal exons (n = 1048 for rep1 and n = 925 for rep2), IPA terminal exons (n = 382 for rep1 and n = 312 for rep2) and introns (n = 9,268 for rep1 and n = 8284 for rep2). The center lines denote the median values with the boxes are bounded by the 25th and 75th percentiles. The whiskers extend to the maximum and minimum values within 1.5 times the interquartile range (IQR) from each end of the box. k, l Two examples of InPACT-identified IPA isoforms translated in HEK293 cells. The plots show the reads coverage of RNA-seq and ribo-seq in the locus for PIGL (k) and CHRNA5 (l), with the annotation of Refseq isoforms and InPACT-identified IPA isoforms. The coding sequences and stop codons are illustrated in the annotation tracks.