Fig. 4

Characterization of InSETT and InSETG structures by Nanopore sequencing. a Odds ratios of enrichment of the phenotypic (depleted or enriched) and non-phenotypic insertions in the intronic and intergenic GENSCAN-specific exons. b Odds ratios of enrichment of IACFs, insertions in CACFs and non-phenotypic insertions in the intronic GENSCAN-specific exons or in intronic regions outside of the GENSCAN-specific exons. Error bars indicate SD based on the 4 systems (K562 growth, HepG2 growth, imatinib survival challenge, and etoposide survival challenge). Asterisks indicate significant differences per two-sided paired Student’s t test (p value < 0.05). c, d Overlaps of InSETes detected by the IACF and CACF analysis in the K562 drug survival assays genome-wide (c) and among the 131 exons selected for the RACE assays (d). e Average TPM of all protein-coding transcripts and those with insertions in their exons, and of all GENSCAN-specific exons and those with insertions. Error bars indicate SD based on 2 biological replicates. Asterisks indicate significant differences under two-sided paired Student’s t test (p value < 0.05). f Sensitivity of detection of novel functional transcripts with PacBio, Illumina, or with the RACE enrichment followed by Nanopore sequencing. Note: the relatively short Illumina reads overlapping InSETes could be derived from transcripts different from those detected by RACE/Nanopore. g Summary of the types of novel transcripts detected by the RACE combined with Nanopore sequencing for the 131 InSETes. Transcript classification was done based on RACE, unless only RT-PCR results were available. h CAGE signal in the real and simulated InSET TSS clusters. For each InSET TSS cluster (left) or intergenic InSET TSS cluster (right), the number of tissue or primary cell samples with overlapping CAGE tags (upper) and the normalized abundance of the CAGE signal (lower) were calculated. In the boxplots, center red lines indicate median; box limits indicate upper and lower quartiles; whiskers extend from the box limits no more than 1.5× interquartile range. In the violin plots, all violins were scaled to have the same maximum width; three horizontal white lines indicate upper, median, and lower quartiles of the density estimate respectively. All data including TSS clusters with 0 CAGE tags were plotted. To plot on the log scale, 0.1 was added to all values. The p values were calculated using Wilcoxon rank-sum test. Source data are provided in Additional file 2: Tables S8, S9, and S11-14