Tabhu

Tabhu facilitates antibody humanisation by predicting antibody-antigen contacts, selecting human templates, identifying grafting regions, optimizing back-mutations, and constructing three-dimensional models to preserve binding affinity and reduce immunogenicity.

Key Features:

• Template Selection: Selects appropriate human templates for antibody humanisation to ensure compatibility with the target antibody framework.
• Grafting Region Identification: Identifies optimal grafting regions within the antibody structure to maintain binding affinity and specificity.
• Paratope Prediction: Predicts paratopes for non-human and humanised antibodies to map antibody-antigen interaction regions.
• Back-Mutation Optimization: Identifies a minimal set of back-mutations required to retain parental binding affinity while minimizing changes from the human template.
• Three-Dimensional Model Construction: Builds and evaluates three-dimensional models of humanised antibodies for structural validation and refinement.

Scientific Applications:

• Diagnostics: Supports engineering antibodies for diagnostic assays by preserving antigen recognition while humanising antibody sequences.
• Biotechnology: Enables modification of antibodies for biotechnological applications that require human-compatible sequences.
• Therapeutics: Assists therapeutic antibody development by reducing immunogenicity while maintaining binding properties.
• Molecular docking and re-design experiments: Provides predicted contact residues and models to guide molecular docking and antibody re-design experiments.

Methodology:

Tabhu applies random forest automatic learning methods to predict antibody-antigen contact residues from sequence data alone and analyses sequence features that contribute to antigen binding, reporting recall and specificity rates up to 80%.