Ribosome profiling

Ribosome profiling was conducted as described previously ⁷ using the Illumina TruSeq Ribo Profile (Mammalian) Kit according to manufacturer’s instructions with modifications in harvest, RNA isolation/purification (isopropanol isolation used to improve the yield) and ribosome protected fragments size selection (~20–32 nt). During harvest, media was carefully removed, and cells were immediately flash-frozen. All equipment used from hence forth was pre-chilled. Cells were quickly scraped into 1 ml of ice-cold lysis buffer (5X Mammalian Polysome Buffer, 10% Triton-X100, 100 mM DTT, DNase I, Nuclease-free water) and homogenized on ice by passing through a 26G needle 10 times. Lysate was then spun at 4°C for 10 minutes at 20,000 × g. Supernatant was aliquoted into cryovials and immediately frozen in liquid nitrogen for future use. Samples were sequenced using Illumina HiSeq 2500.

The complete ribosome profiling pipeline analysis is described in Figure 1: Sequencing data were pre-processed and aligned as described by Alexaki et al. ⁵ as well as the step by step guide found in the README.txt accessible on GitHub.

Colored arrows indicate steps that first require execution of utility script (blue and yellow) or require manual input by the user (red). Pipeline steps are represented as ovals (main step) or pentagons (validation / analysis step). Rectangles represent input / output data. UTR: untranslated region, CDS: coding sequence, RPF: ribosome protected fragments, RPKM: reads per kilobase of transcript per million mapped reads.

RPF sequences were analyzed based on fragment length ( Figure 2a), alignment distribution between coding sequences (CDSs) and 5’- and 3’-UTRs ( Figure 2b), triplet periodicity ( Figure 3a) and reading frame ( Figure 3b). RPF fragments 20–22 nt and 27–29 nt in length were used for further analysis with a P-site offset of 12 nucleotides from the 5’ end of the fragment. Pearson and Spearman correlations were used to evaluate the reproducibility between replicates using a common subset of moderately to highly expressed genes (reads per kilobase of transcript per million mapped reads, RPKM _CDS ≥10) and considering reads with the ribosome A site annotated at least 20 nt downstream of the coding sequence start codon ( Table 1). Both Pearson and Spearman coefficients show strong correlation between experimental replicates.

( a) Fragment size distribution of Ribo-seq and RNA-seq reads. The average of 6 experiments (3 WT and 3 CO F9) was plotted, s.e.m. are shown. ( b) Distribution of Ribo-seq (left) and RNA-seq (right) reads in mRNA coding regions (CDSs) and untranslated (5’UTR and 3’UTR) regions. The average of 6 experiments (3 WT and 3 CO F9) was plotted, s.e.m. are shown.

( a) Profiles of the 5′ end positions of all 20–22 nt (top) and 27–29 nt (bottom) fragments relative to the start codon of their genes. The average of 6 experiments (3 WT and 3 CO F9) was plotted. ( b) Positions of 20–22 nt and 27–29 nt fragments relative to the reading frame of the Ribo-seq (left) and RNA-seq (right) reads. The average of 6 experiments (3 WT and 3 CO F9) was plotted, s.e.m. are shown.

RPKM of each gene in the Ribo-seq and RNA-seq datasets were calculated, considering reads with the ribosome A site annotated at least 20 nt downstream of the start codon. A comparison between each pair of experiments within the 3 replicates was performed