Skip to main content

Advertisement

Table 5 Performance metrics for the proposed approach, Blast, PSI-Blast and Delta-Blast, in discriminating the resistant proteins from non-resistant proteins, where the positive dataset consists of <40% (first) and <90% (second) pair-wise sequence identity

From: DIRProt: a computational approach for discriminating insecticide resistant proteins from non-resistant proteins

Dataset Method Sn Sp Ac Pre MCC
First Proposed 0.897 0.934 0.916 0.933 0.836
Blast 0.961 0.611 0.786 0.713 0.617
PSI-Blast 0.959 0.602 0.780 0.707 0.607
Delta-Blast 0.961 0.652 0.806 0.735 0.647
Second Proposed 0.875 0.891 0.883 0.901 0.784
Blast 0.958 0.350 0.654 0.596 0.392
PSI-Blast 0.958 0.358 0.658 0.601 0.400
Delta-Blast 0.958 0.466 0.712 0.646 0.495
  1. Here, AUC-ROC values were not computed, as in Blast algorithms accuracies are computed based on number of hits
  2. Sn Sensitivity, Sp Specificity, Ac Accuracy, Pre Precision, MCC Matthew’s correlation coefficient