Fig. 2

Highly translated mRNAs are more reactive to DMS in the coding region and 3′ end of the 5′UTR. a A scatter plot of the log(e) fragments per kilobase million (FPKM) in the sub-polysomal and polysomal fractions, color coded by the top (high TE) and bottom (low TE) third of genes ranked by the translational efficiency (TE), which is calculated as a ratio of polysomal to sub-polysomal RNA. b–d Violin plots depicting the average reactivity under control conditions in the 5′UTRs, CDSs, and 3′UTRs, for the top and bottom third of mRNAs ranked by TE after filtering by coverage and 5′ end coverage and selecting the most abundant transcript per gene. Violin plots include boxplots, with the mean denoted by a dot. P values and 95% confidence intervals were calculated using an un-paired, two-sided Wilcoxon test. Each group contains 627 mRNAs. e Binned average reactivity, under control conditions, for the top and bottom third of mRNAs ranked by TE, after removing mRNAs whose 5′UTR, CDS, or 3′UTR is shorter than 100 nt, filtering by coverage and 5′ end coverage and selecting the most abundant mRNA per gene. There are 422 mRNAs in each group. The top panel plots the binned average reactivity under control conditions for all the low TE and high TE mRNAs, across the length of the UTRs (25 bins) and coding sequence (50 bins). The bottom panel plots the Δ reactivity between the low TE and the high TE group, which is calculated by subtracting the high TE from the low TE; therefore, a negative value indicates increased reactivity and therefore less structure in the high TE group, whereas a positive value indicates more structure in the high TE group. Shaded area represents 95% confidence limits for the difference in means between the two groups of mRNAs within each bin, calculated by an un-paired two-sided t test