Skip to main content

Table 1 Performance of prediction of DGJ responsiveness by DecisionStump on sequence derived features.

From: Prediction of the responsiveness to pharmacological chaperones: lysosomal human alpha-galactosidase, a case of study

  accuracy precision recall
zeroR 0.240 ± 0.003 0.239 ± 0.003 1 ± 0
MUPRO 0.811 ± 0.001 0.555 ± 0.091 0.302 ± 0.017
Blosm62 0.737 ± 0.005 0.183 ± 0.057 0.176 ± 0.069
Close_homologs 0.867 ± 0.001 0.666 ± 0.125 0.489 ± 0.102
Far_homologs 0.786 ± 0.026 0.463 ± 0.033 0.364 ± 0.027
Close_orthologs 0.812 ± 0.001 0.477 ± 0.097 0.347 ± 0.062
PolyPhen 0.745 ± 0.028 0.179 ± 0.151 0.108 ± 0.058
  1. Mutations were scored by the trivial classifier ZeroR to set a baseline, by MUPRO, by Bloum62 or by PSSMs derived by sets of sequences which include distant homologs (e-value > e-3), close homologs (e-value > e-50) or close orthologs (e-value > e-13).