Fig. 3

Fragment distribution analysis performed using a fragment length organisation similarity score (FLOSS). With a view to ascertaining the conformity to ribosome profiling read fragment length distributions for coding sequences (CDS), normalised histograms of fragment lengths corresponding to putative coding ORFs and non-coding genes were compared to a reference distribution represented by the mean of the normalised histograms of fragment lengths for all coding features. Curves are representative of outlier thresholds for standard (thinner curve) and extreme (thicker curve) cases using Tukey’s method. Higher FLOSS suggest less conformity. FLOSS are plotted as a function of total reads aligned to each feature. Marginal densities are also given for read counts (x-axis) and FLOSS (y-axis)