Fig. 3

Identification of upstream open reading frames (uORFs) in S. cerevisiae. a Multiple datasets described previously in [19] were used to identify potential uORFs using a program described previously [35]. The potential uORFs ≥ 3 codons (N = 6061) were run through a program previously described [36] to find 1367 uORFs showing evidence of translation in the combined dataset generated in this study (each replicate, all temperatures). The differential expression at 20 °C or 37 °C with respect to 30 °C was analyzed by DESeq2. b Distribution of start sites in translated uORFs. The AUG uORFs were sorted into good context (A/G) or poor (U/C) based on the nucleotide at the − 3 position with respect to AUG. Percentages were calculated with respect to the total number of translated uORFs. c Transcriptome-wide distribution of potential upstream start sites (USSs). The USSs that can lead to formation of uORFs < 3 codons were excluded so that an appropriate comparison with the translated set of uORFs could be done. Percentages were calculated with respect to the total number of USSs (N = 58,353). d Analysis of TE_uORF at 30 °C for both NCC and AUG uORFs. All indicates all the translated uORFs identified through our pipeline (N ~ 1350); first base change: uORFs with UUG, CUG, GUG (N = 631); second base change: uORFs with AAG, ACG, AGG (N = 109); and third base change: uORFs with AUC, AUA, AUU (N = 483). AUG uORFs were sorted into good context (N = 36) or poor (N = 93). The dotted horizontal line indicates the median TE value of All. e Analysis of TE of mORFs located downstream of the various uORFs at 30 °C. All indicates the set of mORFs located downstream of all translated uORFs analyzed in this study (N = 748)