Table 3.

Comparison of the quality of classification of deleterious mutations in dog and mouse datasets for classifiers, trained with weights (+TL) and without weights (—) on source data

Classifier	Dog				Mouse
	Trained on HumDiv		Trained on HumVar		Trained on HumDiv		Trained on HumVar
	+TL	—	+TL	—	+TL	—	+TL	—
Random Forest	0.855	0.638	0.889	0.884	0.846	0.682	0.841	0.682
Polynomial SVM	0.874	0.657	0.715	0.454	0.764	0.764	0.812	0.528
Gaussian SVM	0.686	0.662	0.753	0.618	0.655	0.539	0.833	0.560
Logistic Regression	0.908	0.667	0.855	0.701	0.777	0.576	0.875	0.565
Linear SVM	0.672	0.672	0.715	0.715	0.597	0.597	0.852	0.568

In each cell, we present the accuracy rate reached by a classifier. The maximal accuracy values achieved are shown in bold.