Fig. 2

Distances of perfect TATAAT -10 elements to TSS, and the relation between cluster-size and level of gene expression. a Cartoon of the transcription initiation complex bound to a promoter, with three potential TSSs. The rTSS (with the lowest p-value) is indicated in blue. b Distances from the σ^A -10 elements to TSS clusters with a single detected TSS. In order to avoid ambiguity, only TSS clusters that have the perfect σ^A TATAAT -10 element consensus sequence are included. c Distances from the σ^A -10 elements to the rTSS in TSS clusters with more than one detected TSSs. As in panel B, only perfect TATAAT elements are included. d Higher expression levels increase the chances of detecting alternative TSSs for a promoter. Expression level from each TSS cluster (a single TSS in the cluster corresponds to an isolated TSS) was estimated by taking the average number of reads in the RNAseq data that map inside a 300 nt window immediately downstream of each rTSS. RPM indicates Reads Per Million. Shapiro-Wilk test was carried out to confirm normal distribution (p < 0.05), and Students T-test was used to determine that expression was higher from TSS clusters with wobble (*** and * indicate p < 0.001 and p < 0.05, respectively). The number of clusters with 5 TSSs was too low to perform a meaningful statistical analysis