Skip to main content

Table 4 The performance on the Gram-negative datasets

From: SigUNet: signal peptide recognition based on semantic segmentation

MethodMCC (%)FPRTM (%)Precision (%)Recall (%)F1 measure (%)
The SignalP dataset
 Phobius59.922.643.994.259.9
 PrediSi30.669.019.786.532.1
 SignalP3.0-HMM47.739.231.693.347.2
 SignalP3.0-NN36.761.022.195.235.9
 PolyPhobius60.721.445.094.260.9
 Philius65.914.951.394.266.4
 SPOCTOPUS64.717.050.892.365.5
 SignalP 4.084.81.5
 TOPCONS270.813.257.295.271.5
 DeepSig81.21.788.976.982.5
 SigUNet80.61.588.876.081.9
The SPDS17 dataset
 Phobius69.518.056.495.771.0
 PrediSi35.466.325.087.038.8
 SignalP3.0-HMM65.421.351.295.766.7
 SignalP3.0-NN49.144.933.895.750.0
 PolyPhobius75.913.562.2100.076.7
 Philius88.72.284.695.789.8
 SPOCTOPUS62.520.250.091.364.6
 SignalP 4.092.50.0100.087.093.0
 TOPCONS285.95.676.7100.086.8
 DeepSig95.00.0100.091.395.5
 SigUNet97.61.195.8100.097.9
  1. The best performance is highlighted in bold