Fig. 1

RNA G-quadruplexes within 5′-UTR alter ribosome distribution and impede translation of associated CDSs. a The transcriptome-wide distribution of translation efficiency of human mRNAs shows a skew towards inefficiently translated transcripts. b dsRNA and c rG4 associated predicted folding energies of 5′-UTRs of human transcripts binned according to their TE (first to fourth quartile of TE distribution). Folding energies are expressed as z-scores of the minimum free energies normalized by the length of the 5′-UTRs. d Hierarchical clustering analysis of human transcripts according to their TE and length-normalized dsRNA and rG4 predicted folding energies of 5′-UTRs. The heatmap reports the z-scores of the three variables. e Ribosome density within EED mRNA showing high ribosome occupancy in the 5′-UTR upstream of rG4-forming sequences (highlighted by a red box). Biophysical characterization of the EED rG4 motif is reported in Figure S2 in Additional file 1. f Ribosome distribution for transcripts of cluster 1 (blue) and cluster 2 (red), i.e., transcripts with low TE and associated with predicted stable dsRNA or rG4 structures respectively. g Ribosome distribution for transcripts of cluster 2 when binned for increasing predicted rG4 stability, − 0.5σ < ΔG⁰rG4 ≤ 0σ (blue line), − 1σ < ΔG⁰rG4 ≤ − 0.5σ (green line), and ΔG⁰rG4 ≤ −1σ (red line), where ΔG⁰rG4 is the z-score of 5′-UTR length-normalized rG4 predicted minimum free energy. Ribosome footprint coverage and transcript length are normalized; dotted lines indicates annotated translation start and stop sites and the arrows highlight the presence of RPFs in 5′-UTR. Data reported in b and c are means ± s.e.m.; P values were assessed using one-tailed Mann–Whitney nonparametric tests and represent statistical difference between the binned population and the rest of the population. *P < 0.05, ***P < 0.001