Fig. 3: Modeling using SBNA and BLOSUM62 is superior to SBNA network scores and BLOSUM62 scores alone. | npj Genomic Medicine

Fig. 3: Modeling using SBNA and BLOSUM62 is superior to SBNA network scores and BLOSUM62 scores alone.

From: Structure-based network analysis predicts pathogenic variants in human proteins associated with inherited retinal disease

Fig. 3

a ROC curves for network scores alone (red), BLOSUM62 scores alone (blue), and the difference between network scores and BLOSUM62 scores (purple). ROC curves were determined using all variants with available clinical phenotype data for all 47 inherited retinal disease proteins. AUC values are shown for each curve. b Pooled comparison between the difference between network scores and BLOSUM62 scores for variants with available clinical phenotype data for all 47 inherited retinal disease proteins. c ROC as shown in (a) with added comparison to EVE scores (orange) and the sum of EVE scores and the scaled difference between network scores and BLOSUM62 scores (light purple). ROC curves were determined using all variants with available clinical phenotype data for all 47 inherited retinal disease proteins. AUC values are shown for each curve. d Pooled comparison between the sum of EVE scores and the scaled difference between network scores and BLOSUM62 scores (“combined score”) for variants with available clinical phenotype data for all 47 inherited retinal disease proteins.

Back to article page