Protean 3D uses the NeoClone method, a machine learning approach developed in collaboration with
NeoClone® Biotechnology International, that improves the ability to predict linear B cell epitopes using only sequence-based information. The NeoClone method is a predictive model that considers secondary structure, flexibility, hydropathy, and antigenicity, and is based on NeoClone’s antibody development and bioinformatics approach. The NeoClone method is described further in the white paper, Precise Predictions of Linear B Cell Epitopes in Protean 3D.

A cross-validated machine learning approach was used to train two models that predict antigenic and highly antigenic regions. The antigenic models consider sequence-based helix, sheet, beta turn, and flexibility predictions as well as hydropathy, antigenicity, and stability indices when predicting antigenic regions; the highly antigenic models consider a modified combination of secondary structure predictions, hydropathy, antigenicity, and stability.

Yes. A confidence score with a value from 0 to 1 is assigned to each residue in a protein sequence. By default, residues with a confidence score greater than 0.5 are classified as putative antigenic residues. You can edit the confidence score threshold by first selecting the B-cell epitopes (DNASTAR) track from the Tracks panel, then opening the Options area of the panel. Changing this threshold value will update the Analysis View accordingly.