Skip to main content

Table 4 Performance comparison of different methods based on the balanced subset of the independent test set in which benign variants were randomly selected from the full negative independent test set. We repeated this process 5 times with different random subsets of benign variants and averaged the results

From: Computational identification of deleterious synonymous variants in human genomes using a feature-based approach

Method Recall Precision F-measure (β = 1) AUC P-value
IDSV 0.765 ± 0.000 0.781 ± 0.022 0.772 ± 0.011 0.868 ± 0.008 *
CADD 0.320 ± 0.000 0.760 ± 0.041 0.450 ± 0.007 0.698 ± 0.018 9.452e-07
FATHMM-MKL 0.712 ± 0.000 0.660 ± 0.026 0.685 ± 0.014 0.753 ± 0.019 0.0007962
SilVA 0.490 ± 0.000 0.977 ± 0.017 0.653 ± 0.004 0.844 ± 0.017 3.211e-05
DDIG-SN 0.298 ± 0.000 0.996 ± 0.010 0.459 ± 0.001 0.853 ± 0.006 5.957e-07
TraP 0.575 ± 0.000 0.971 ± 0.012 0.723 ± 0.003 0.848 ± 0.043 0.001015
  1. The highest values are highlighted in bold. *Denotes the reference when calculating the P-value