Structure-based antibody paratope prediction with 3D Zernike descriptors and SVM

Background: This work presents a structure-based method for antibody interface prediction based on 3D Zernike Descriptors. Roto-translationally invariant descriptors are computed from circular patches of the antibody surface enriched with a chosen subset of physicochemical properties from the AAindex1 amino acid index set, and are used as samples for a binary classification problem. An SVM classifier is used to distinguish interface surface patches from non-interface ones. The proposed method was shown to outperform other antigen-binding interface prediction software.binaries&scripts.tar.gz - Contains the Linux binaries and Python scripts used in this study.seq_descriptors_4.5.tar.gz - Contains the sequence descriptors generated from the antibody structures in the development/validation set which were used for feature selection with the Randomized Logistic Regression algorithm.development_set_structures_descriptors.tar.gz - Contains the PDB structures of the antibody-antigen complexes in the development/validation set; the 3D Zernike Descriptors for each antibody; the predicted patch scores for each antibody; other supplementary files.training_set_structures_descriptors.tar.gz - Contains the PDB structures of the antibody-antigen complexes in the training set and the 3D Zernike Descriptors (rebalanced) for each antibody.test_set_protein_ag_structures_descriptors.tar.gz - Contains the PDB structures of the antibody-antigen complexes in the test set of antibodies complexed with protein antigens; the 3D Zernike Descriptors for each antibody; the predicted patch scores for each antibody; other supplementary files. test_set_non_protein_ag_structures_descriptors.tar.gz - Contains the PDB structures of the antibody-antigen complexes in the test set of antibodies complexed with non-protein antigens; the 3D Zernike Descriptors for each antibody; the predicted patch scores for each antibody; other supplementary files.homology_modelled_test_set_protein_ag_structures_descriptors.tar.gz - Contains the PDB structures of the homology modelled antibody structures in the test set of antibodies complexed with protein antigens; the 3D Zernike Descriptors for each antibody; the predicted patch scores for each antibody; other supplementary files. homology_modelled_test_set_non_protein_ag_structures_descriptors.tar.gz - Contains the PDB structures of the homology modelled antibody structures in the test set of antibodies complexed with non-protein antigens; the 3D Zernike Descriptors for each antibody; the predicted patch scores for each antibody; other supplementary files.