Applying support-vector machine learning for predicting B-cell epitopes on the spike protein of SARS-CoV-2 variants
Additional file 1: Table S1. Training dataset. " Position " indicates the positional information of residues on the peptide chain. In the "epitope" column. “1” indicates the experimentally confirmed epitope residues of WT SARS-CoV-2. “- 1” indicates the unconfirmed epitope residues of WT SARS-CoV-2. Additional file 2: Table S2. Prediction performance of the SVM model with different feature combinations in the training dataset by a 10-fold cross-validation. Additional file 3: Table S3. Twenty-five kinds of antibodies binding to the Omicron spike protein from the PDB database. Additional file 4: Table S4. Epitope information for Omicron (B.1.1.529) from the IEDB website. Additional file 5: Table S5. Omicron (B.1.1.529) RBD dataset. " Position " indicates the positional information of residues on the peptide chain. Additional file 6: Table S6. Delta spike protein dataset. Additional file 7: Table S7. The 75 real epitopes of Delta variant spike protein were obtained from the PDBePISA website. Additional file 8: Table S8. The probabilistic SVM model predicted 363 epitopes of the Delta variant spike protein. Additional file 9: Table S9. The original data of the ROC curve. Additional file 10: Table S10. Omicron subvariant prediction results using the probabilistic SVM model.