Fig. 4

Sense internal TSS. A Illustration of the strategy to differentiate genuine internal TSS from background signals. iTSS represents a reproducible internal TSS (= signal = S). The relative read score is compared to the RRS of the position with the highest signal, which is not reproducibly detectable as a TSS in all replicates (= noise = N). iTSS with \(\frac{S}{N}>1.5\) in all three replicates are considered as genuine internal TSS. B Comparison of the signal strength of iTSSs and gTSSs of the corresponding annotated gene. The log₁₀ of the mean RRS of three replicates is shown as box plots within the violin plot. Outliers are indicated with dots. The violin plot visualizes the abundance of TSSs at a specific RRS values. C Schematic overview for RT-qPCR quantification of different mRNA molecules of internal and gene associated TSS of an annotated gene. The RT-primer (gray) was used for cDNA synthesis of the respective RNA transcripts. Two primer pairs (black) were designed to amplify ~ 100 bp fragments downstream of the gTSS but upstream of the iTSS (orange) and downstream of the iTSS (blue), respectively, with equal efficiencies. A lowered Cq value in RT-qPCR (earlier crossing with the threshold line) for the iTSS values is an indication for an individual transcript additionally originating from the iTSS as secondary, short mRNA transcripts (left panel). If the same mRNA is used for amplification of gTSS and iTSS fragments (i.e., iTSS is not present or weak), similar Cq values are expected (right panel). D Quantification of mRNA originating from transcripts starting at gTSS and iTSS. Two genes, EDL933_RS17005 and EDL933_RS24700, are shown here. The normalized expression (ΔΔCq) regarding the gene cysG in LB medium (exponential phase) is used as normalizing gene. Mean value and standard deviation was calculated based on three biological replicates for the two genes indicated. Statistical significance between the normalized expression was verified with a one-tailed Welch two sample t-test (* p ≤ 0.05)