AlphaMissense prediction correlations with cystic fibrosis patient pathogenicity metrics by diagnosis.

A. AM score plotted against sweat chloride levels in milliequivalents per liter (mEq/L) for 85 missense variants classified as CF causing. The linear correlation (Pearson Coefficient r = 0.21, Spearman Coefficient ρ = 0.33) is reduced compared to the complete data set correlation shown in Fig 2A. B. AM score plotted against sweat chloride levels for 33 missense variants classified as variants of variable clinical consequence (VVCC). The linear correlation (Pearson Coefficient r = -0.12, Spearman Coefficient ρ = -0.11) is statistically insignificant. C. AM score plotted against pancreatic insufficiency rates in percent for 83 missense variants classified as CF causing. The correlation (Pearson Coefficient r = 0.32, Spearman Coefficient ρ = 0.44) was similar to the entire dataset in Fig 2B. D. AM score plotted against pancreatic insufficiency rates for 30 missense variants classified as VVCC. The correlation for these data (Pearson Coefficient r = -0.21, Spearman Coefficient ρ = -0.22) was statistically insignificant. E. AM score plotted against pseudomonas infection rates for 82 missense variants classified as CF-causing. Linear correlation is reduced compared to the entire data set presented in Fig 2C (Pearson Coefficient r = 0.33, Spearman Coefficient ρ = 0.36). F. AM score plotted against pseudomonas infection rates in percent for 28 missense variants classified as VVCC. Linear correlation was insignificant (Pearson Coefficient r = 0.13, Spearman Coefficient ρ = 0.28).

(TIF)