Skip to main content

Table 2 Prediction of isoelectric point on the 75 % training datasets

From: IPC – Isoelectric Point Calculator

Method Protein dataset Method Peptide dataset
RMSD % Outliers RMSD % Outliers
IPC_protein 0.838 0 114 IPC_peptide 0.247 0 635
Toseland 0.898 15.0 131 Solomons 0.251 0.8 638
Bjellqvist 0.922 21.5 149 Lehninger 0.256 2.4 643
Dawson 0.920 20.9 156 EMBOSS 0.322 18.8 1088
Wikipedia 0.930 23.8 157 Wikipedia 0.413 46.3 4280
Rodwell 0.938 26.1 159 Sillero 0.426 50.9 3025
ProMoST 0.938 26.1 140 Toseland 0.427 51.2 3618
Grimsley 0.939 26.2 147 Dawson 0.432 52.9 4192
Solomons 0.947 28.5 159 Thurlkill 0.480 70.8 4017
Lehninger 0.947 28.7 160 Rodwell 0.506 81.2 4061
pIR 1.026 54.2 180 DTASelect 0.541 96.8 4902
Nozaki 1.005 47.1 169 Nozaki 0.599 124.8 4013
Thurlkill 1.018 51.5 173 Grimsley 0.611 130.9 4609
DTASelect 1.017 51.1 167 Bjellqvist 0.661 159.2 4672
pIPredict 1.057 65.9 173 pIPredict 1.024 497.8 8051
EMBOSS 1.040 59.4 189 ProMOST 1.233 867.5 7999
Sillero 1.042 60.1 188 pIR 1.862 4020.9 9921
Patrickios 2.237 2405.1 645 Patrickios 1.977 5266.8 8131
Avg_pIa 0.940 26.6 151 Avg_pI 0.451 59.7 4600
  1. aAverage from all pKa sets without the Patrickios (highly simplified pKa set) and IPC sets. Note, that the average pI is calculated on the level of individual protein or peptide
  2. Protein dataset (IPC_protein trained on 1,743 proteins with 10-fold cross-validation – data in the table above, tested on 581 proteins not used for training – data in Table 1), peptide dataset (IPC trained on 12,662 peptides with 10-fold cross-validation – data in above table, tested on 4,220 peptides not used for training – data in Table 1). Changes in method order in comparison to Table 1 are in bold
  3. Outliers correspond to the number of predictions for which the difference between the experimental pI and the predicted pI exceeded the threshold of an MSE of 3 for the protein dataset and an MSE of 0.25 for the peptide dataset