Analysis of ribosome profiling and RNA sequencing data

Ribosome profiling and RNA sequencing (RNA-seq) experiments were described previously (23). For ribosome profiling, adapter sequence at the 3′ end and first nucleotide at the 5′ end of each FASTQ read were trimmed. Trimmed reads were mapped to human ribosomal (rRNA) reference sequences (NR_003285.2, NR_003286.1, NR_003287.1, and NR_023363.1) with Bowtie (49). rRNA-unmapped reads were aligned to hg19 refGene and genome sequence with TopHat (50). Only uniquely mapped reads were used for further analysis. Ribosome occupancy was used to measure gene expression and was calculated as reads per kilobase of transcript per million mapped reads (RPKM) in defined regions of gene open reading frames (ORFs), which excluded 15 codons downstream of start codons and 5 codons upstream stop codons as well as the coding regions that overlapped with upstream ORFs we annotated previously (23) to improve accuracy of the measurement. For RNA-seq, sequencing reads were mapped with the same procedure as described in ribosome profiling. RNA expression levels of genes were calculated as RPKM. For reliable measurement of RNA levels and fold changes, at least five raw reads for each transcript were required.