Fig. 4

Analysis of translation initiation. a Meta coverage plot of RCP-seq data over 5′ UTRs, CDSs, and 3′ UTRs from all mRNA transcripts [8]. Regions are scaled to be the same size and coverage is displayed as the sum of reads for the translating (80S) and scanning ribosomes (SSU). b 5′ end coverage heatmaps of SSU reads (left column) and 80S (right column) from TCP-seq relative to transcription start site. Coverage is shown before (upper row) and after (lower row) reannotation of transcription start sites with CAGE data. c Analysis of initiation context showing the median initiation rate (IR) for translation initiation sites with a specific base (x-axis) at a specific position (y-axis). The strongest consensus is displayed with a black frame. d Analysis of ribosome recruitment showing the relationship between different TSS contexts and SSU scanning efficiency. Left: empirical cumulative distribution function (ecdf) of scanning efficiency colored by the first nucleotide in the 5′UTR. Right: ecdf of scanning efficiency for 3 different motifs; a C nucleotide, the TOP motif (C, then 4 pyrimidines), and all other TSS variants. e Initiation contexts ranked by IR. Top: mean IR for all CDSs with specified sequence (x-axis) relative to TIS (− 4 to − 1). Bottom: similarity score to human Kozak sequence as defined in [48, 49]