Table 1.

Definitions of common performance metrics used in evaluating variant callers.

Metric	Calculation	Interpretation (Ideal)	Alternative names
Accuracy	$\frac{\text{TP+TN}}{\text{TP+FP+TN+FN}}$	Ratio of correct calls to total calls and variants (1)
Specificity	$\frac{\text{TN}}{\text{TN+FP}}$	Non-variants not called as variants relative to the total non-variants (1)
Sensitivity	$\frac{\text{TP}}{\text{TP+FN}}$	True variants called relative to all variants (1)	Recall, true positive rate (TPR), positive call rate
Precision	$\frac{\text{TP}}{\text{TP+FP}}$	True variants called relative to total calls (1)	Positive predictive value (PPV)
False positive rate	$\frac{\text{FP}}{\text{TN+FP}}$	Non-variants called relative to the total non-variants (0)