Fig. 4

Widespread translation of AUG uORFs. a Box plots showing the relative footprint density on different codons (three reading frames) in 5’ UTRs for control, harringtonine, puromycin and DMDA-PatA treated samples. For the three forward reading frames footprints located on the center nucleotide of a codon were assigned to the corresponding codon for each transcript. Footprints principally accumulate on the center nucleotide of a codon with the read start offset we used (Additional file 1: Figure S1). To calculate relative footprint densities for a codon in a given 5’ UTR, the average number of footprints on a codon was divided by the average footprint density of all codons of the respective 5’ UTR. The box plots show data for transcripts with at least 50 footprints in the 5’ UTR. The right panel is a magnification of the left one. Numbers (n) indicates the number of transcripts with at least one such codon that pass the filters. A median relative density of above one indicates that a codon has an above average number of footprints in the majority of 5’ UTRs. b Cumulative frequency distribution of the fraction of drug resistant ribosome footprints for: (i) ORFs with AUG start codons in the 5’ UTRs of annotated protein coding transcripts (solid lines). Only ORFs with at least 20 ribosome footprints are shown (n = 2078, median and mean ORF sizes are 72 and 51 nt. respectively). (ii) The entire 5’ UTR of transcripts that contain above cutoff (>20 footprints) ORFs (dashed lines, n = 1256). The last 10 nucleotides of the 5’ UTRs were clipped to avoid carryover of footprints flanking the start codon of the annotated coding sequence. Inhibitors used were: puromycin (Puro), harringtonin (Harr) and DMDA-Pateamine A (PatA). c Cumulative frequency distribution of the fraction of drug resistant ribosome footprints for: annotated protein coding sequences (black); Upstream ORFs with AUG (orange) or CUG (blue) start codons; entire 5’ UTR (yellow). Only data for ORFs with at least 20 ribosome footprints (AUG, n = 2078; CUG, n = 6998) are shown. d Cumulative frequency distribution of the accumulation of ribosome footprints, on the translation start codons AUG (red) and CUG(blue) of ORFs in the 5’ UTRs of annotated protein coding sequences. The ratio of footprints per nucleotide for the translation start codon +/− 1 nucleotide and the footprints per nucleotide for the entire ORF provide a measure of footprint accumulation on the translation start codon. Cutoffs for ORFs and 5’ UTRs were as for B and C. e Boxplots showing the distribution of a Kozak consensus sequence score (methods section and [25] for all AUG uORFs of transcripts expressed in murine ES cells (>8.5 normalized transcriptome reads, n = 16246), uORFs without footprints (n = 5486), with at least 20 footprints (n = 2911), DMDA-PatA (n = 1195). f A well translated uORF in Collagen 3a1. The inset shows a zoom on the region flanking the uORF which codes for a 3 amino acid peptide. The principal peak of ribosome footprints is located on the AUG start codon of the uORF. The number of ribosome footprints on the annotated coding sequence (CDS, 823) and on the uORF (382) are shown in the histogram. g An upstream ORF that codes just for one amino acid in the small GTP-binding protein Rab10. h The uORF of TGFB-Induced Factor 1 extends into the CDS. The uORF is not in frame with the CDS. The insets show incremental zoom levels on the uORF and on the translation start codons. The main peaks of ribosome footprints are located exactly on the AUG start codons of the uORF and the CDS respectively. In f- h data are shown for control samples (no drug, blue), DMDA-PatA (red), harringtonin (green) and puromycin (yellow)