Figure 1.
Sense codons differ in their propensity for conversion to STOP codons.
The Standard Genetic Code contains 18 fragile codons (shaded) that can be changed into a STOP codon by a single point-mutation and whose mistranscription can therefore generate nonsense errors. The remaining 43 sense codons are “robust” to such errors. Six amino acids are encoded exclusively by fragile codons (“fragile amino acids”, shaded), ten amino acids are encoded exclusively by robust codons (“robust amino acids”, unshaded) and four amino acids can be encoded either by robust or fragile codons (“facultative amino acids”, hatched shading).
Table 1.
Normalized fragile codon content (NFCU) (controlling for GC-content and amino-acid usage) of multi-exon genes and single-exon genes in the human, mouse, and fly genomes.
Table 2.
Fragile amino acid usage (FAU) of proteins encoded by multi-exon genes and single-exon genes in the human, mouse, and fly genomes.
Table 3.
Normalized fragile amino acid usage (NFAU) (controlling for GC-content) of proteins encoded by multi-exon genes and single-exon genes in the human, mouse, and fly genomes.
Figure 2.
The genome-wide correlation between transcriptional robustness strategies depends on selective constraint.
Pairwise correlation between normalized fragile codon usage (NFCU) and normalized fragile amino-acid usage (NFAU) for 17421 human genes with an ortholog in mouse. Human genes were binned by selective constraint (Ka/Ks) estimated using the pairwise alignment with their mouse ortholog and for each quartile of Ka/Ks (Q1, lowest, to Q4, highest), Spearman's correlation between normalized fragile codon usage and normalized fragile amino-acid usage (rNFCU.NFAU) was calculated. The vertical extent of the bar indicates the 95% confidence interval for each correlation.
Table 4.
Normalized fragile codon usage (NFCU) of last exons and upstream exons of multi-exon genes in the human, mouse, and fly genomes.
Figure 3.
Intragenic depletion of fragile codons commences beyond the boundary of EJC–dependent NMD activity.
Normalized fragile codon usage (NFCU) in “NMD-competent” and “NMD-compromised” regions of multi-exon genes. The schematic depicts a generic mammalian mature mRNA. The arrowhead shows the position of the last exon-exon junction. The relative efficiency of each mammalian NMD pathway predicted in three distinct regions is shown using ‘+’ symbols. Predicted inactivity of NMD is shown using a ‘-’ symbol. EJC-dependent NMD is expected to be active >50–55 nts 5′ of the last exon-exon junction and inactive 3′ of the last exon-exon junction. Its activity is uncertain in the intervening 50–55 nts region (hatched shading). PABP-dependent NMD is expected to increase in efficiency with distance from the poly-A tail. Note that the efficiency of PABP-dependent NMD is predicted to be much lower than that of EJC-dependent NMD. NFCU was determined in two windows of 50 codons positioned on either side of the last exon-exon junction. The example shows the last intron in “phase 0” (i.e. the intron is positioned between codons) and the depicted nucleotide coordinates for each window are specific to this case.
Figure 4.
Fragile codon depletion is not due to reduced efficiency of PABP–dependent NMD in shorter mRNAs.
Normalized fragile codon usage (NFCU) of human single- (S) and multi-exon (M) genes binned by transcript CDS length. For each quartile of transcript length (Q1, shortest, to Q4, longest) NFCU for single and multi-exon genes is plotted separately. The width of each bin is proportional to the square root of the number of genes it contains.