Table 1.

SIFT Prediction on Human Variant Databases

Datasets Sequence coverage Amino acid substitution coverage % Predicted as damaging Weighted false positive error
Amino acid substitutions annotated to be involved in disease76% (462/606)71% (5218/7397)69% (3626/5218)Hypothesis not applicable
Polymorphisms detected in normal individuals (WI-nsSNPs)77% (53/69)62% (115/185)19% (22/115)20%
Nonsynonymous changes from dbSNP60% (1789/3005)53% (3084/5780)25% (757/3084)19%

[i] Sequence coverage is the percentage of proteins whose alignments contained enough diverse homologues so that predictions could be made. Amino acid substitution coverage is the percentage of substitutions predicted on. The weighted false positive error is the percentage of substitutions shown experimentally to be neutral that were incorrectly predicted as damaging.