Skip to main content

Table 1 Detailed performance of various SVM based modules and EuPSI-BLAST on the RH2427 dataset

From: ESLpred2: improved method for predicting subcellular localization of eukaryotic proteins

Approaches Cytoplasm Mitochondria Nuclear Extracellular Overall Average
  ACC MCC ACC MCC ACC MCC ACC MCC ACC MCC ACC MCC
AAC-NTerm (A) 82.2 0.76 83.5 0.82 90.3 0.80 86.2 0.86 86.5 0.80 85.5 0.81
PSSM based (B) 84.1 0.79 71.3 0.74 95.2 0.85 93.2 0.95 88.6 0.83 86.0 0.83
*EuPSI-BLAST (C) 77.6 --- 54.8 --- 84.5 --- 86.7 --- --- --- --- ---
Hybrid1 (A+B) 86.1 0.84 89.4 0.89 95.2 0.87 93.9 0.94 91.7 0.88 91.1 0.89
Hybrid2 (A+B+C) 89.6 0.87 90.7 0.91 96.4 0.91 95.7 0.96 93.6 0.90 93.1 0.91
  1. ACC is accuracy; MCC is Matthew correlation coefficient
  2. ACC is calculated in percentage
  3. *The results are obtained from ESLpred method 20