Skip to main content
. Author manuscript; available in PMC: 2018 Aug 2.
Published in final edited form as: Genet Med. 2018 Feb 1;20(9):936–941. doi: 10.1038/gim.2017.230

Figure 3.

Figure 3

Receiver Operating Characteristic (ROC) curves from Naïve Bayes classifiers trained on the same gold-standard dataset from UniProtKB (HumVar). The number of cited variants (NCV) score is generated using a fully disparate set of citation data from HGMD variants that are not recorded in UniProtKB. A classifier trained using PolyPhen-2 features (blue) is compared with one that uses only the NCV score (green). Combining the NCV score with PolyPhen-2 features improves prediction accuracy substantially (red).