Broader pathogenic ENG PTC variant considerations. For color codes, ENG protein domains (A) are colored as in Figure 5 and AlphaMissense74 predictions as benign (green), ambiguous (black in panels B-C; and brown in panel F), and pathogenic (red). For stop codon origins in panels E-F, these are colored purple (TAG/UAG); blue (TAA/UAA), or orange (TGA/UGA). (A) Distribution of the 2018 HHT mutation database–listed nonsense variants. Note that all were located between exons 2 and 13 of the 15 exon gene; variants were similarly spread between codons in the 4 protein domains, but the 3 most common variants (Q77X, R93X, and R171X) were all in the OR2 domain (Figure 5A; supplemental Figure 11). Amino acids flanking the domain junctions, and the PTC-distal site of intermolecular dimerization at C582 are also shown. (B) Comparison of AlphaMissense predictions for all 19 alternate amino acids at each nonsense codon site, categorized by protein domain. P value calculated by Fisher's exact test. (C) Comparison of AlphaMissense metrics for all 19 alternate amino acids at each nonsense codon site, categorized by most common (Q77X, R93X, and R171X) vs other (n = 44) HHT causal nonsense variants. P value calculated by Mann-Whitney U test. Note, paucity of pathogenic predictions for Q77X, R93X, and R171X. (D) Distribution of AlphaMissense pathogenicity scores categorized by common (Q77X, R93X, and R171X) vs other (n = 44) stop codons. ∗∗∗∗P < .0001 calculated by Dunn's test after Kruskal-Wallis test. (E) Genetic code indicating the 12 nonsense source codons in the HHT series, and the 3 resultant nonsense mutant codons. For further detail see supplemental Table 9A. (F) AlphaMissense predictions for the amino acids most likely to be substituted if translational (ribosomal) readthrough occurs, by stop codon type. For further detail see supplemental Table 8B. (G) AlphaMissense predictions for the amino acids most likely to be substituted if translational (ribosomal) readthrough occurs, by stop codon type. (i) All stop codons; and (ii) subcategorizing by more common variants (Q77X, R93X, and R171X). ∗∗∗∗P < .0001 and ∗∗∗P < .001 calculated by Dunn's test after Kruskal-Wallis test.

Broader pathogenic ENG PTC variant considerations. For color codes, ENG protein domains (A) are colored as in Figure 5 and AlphaMissense74 predictions as benign (green), ambiguous (black in panels B-C; and brown in panel F), and pathogenic (red). For stop codon origins in panels E-F, these are colored purple (TAG/UAG); blue (TAA/UAA), or orange (TGA/UGA). (A) Distribution of the 2018 HHT mutation database–listed nonsense variants. Note that all were located between exons 2 and 13 of the 15 exon gene; variants were similarly spread between codons in the 4 protein domains, but the 3 most common variants (Q77X, R93X, and R171X) were all in the OR2 domain (Figure 5A; supplemental Figure 11). Amino acids flanking the domain junctions, and the PTC-distal site of intermolecular dimerization at C582 are also shown. (B) Comparison of AlphaMissense predictions for all 19 alternate amino acids at each nonsense codon site, categorized by protein domain. P value calculated by Fisher's exact test. (C) Comparison of AlphaMissense metrics for all 19 alternate amino acids at each nonsense codon site, categorized by most common (Q77X, R93X, and R171X) vs other (n = 44) HHT causal nonsense variants. P value calculated by Mann-Whitney U test. Note, paucity of pathogenic predictions for Q77X, R93X, and R171X. (D) Distribution of AlphaMissense pathogenicity scores categorized by common (Q77X, R93X, and R171X) vs other (n = 44) stop codons. ∗∗∗∗P < .0001 calculated by Dunn's test after Kruskal-Wallis test. (E) Genetic code indicating the 12 nonsense source codons in the HHT series, and the 3 resultant nonsense mutant codons. For further detail see supplemental Table 9A. (F) AlphaMissense predictions for the amino acids most likely to be substituted if translational (ribosomal) readthrough occurs, by stop codon type. For further detail see supplemental Table 8B. (G) AlphaMissense predictions for the amino acids most likely to be substituted if translational (ribosomal) readthrough occurs, by stop codon type. (i) All stop codons; and (ii) subcategorizing by more common variants (Q77X, R93X, and R171X). ∗∗∗∗P < .0001 and ∗∗∗P < .001 calculated by Dunn's test after Kruskal-Wallis test.

Close Modal

or Create an Account

Close Modal
Close Modal