Table 1.

Average performance of Fido-SNP on the hd-pathogenic and dog-omia data sets. All data contains pathogenic variants and an equal number of potentially benign variants randomly selected from dbSNP. Optimized performance of Fido-SNP obtained maximizing the MCC on the hd-pathogenic set. Performance on the validation set (dog-omia) considering a classification threshold of 0.1

Data set	Threshold	Q2	TNR	NPV	TPR	PPV	MCC	AUC
hd-pathogenic	0.09±0.02	0.87±0.04	0.91±0.04	0.86±0.01	0.85±0.01	0.91±0.04	0.77±0.04	0.91±0.01
dog-omia	0.10±0.01	0.88±0.02	0.92±0.03	0.85±0.01	0.84±0.01	0.92±0.03	0.77±0.04	0.91±0.01
dog-omia*	0.11±0.03	0.87±0.04	0.92±0.05	0.84±0.05	0.82±0.07	0.92±0.05	0.75±0.08	0.91±0.04

*Performance of Fido-SNP on the dog-omia data set using a 3-fold cross-validation procedure. The performance measures are defined in Supplementary Materials. The values are computed using the canfam3 assembly.