A combined functional annotation score of non-synonymous coding variants. CAROL combines information from PolyPhen-2 and SIFT. We use a weighted Z method to derive the combined score. We calibrate and validate CAROL using positive (known disease-causing) and negative (postulated non-disease-causing) control variants. CAROL has high predictive power for the effect of ns variants and has the distinct advantage of high coverage, i.e. low missing data rates. The combination of annotation tools can help improve automated prediction of whole-genome/exome non-synonymous variant functional consequences.