ANABAG: Annotated Antibody-Antigen Data Set with Unique Features for Antibody Engineering Applications.

Area of Science:

• Structural Biology
• Immunology
• Computational Chemistry

Background:

• Antibody-antigen (Ab-Ag) interactions are crucial in immunology but current analysis often neglects key structural and physicochemical factors.
• Existing datasets for Ab-Ag complexes lack standardized criteria and may exclude diverse binding modes due to sequence identity filtering.

Purpose of the Study:

• To introduce ANABAG, a novel, residue-level curated dataset for Ab-Ag complexes.
• To provide a resource incorporating detailed structural and physicochemical features for improved Ab-Ag modeling and antibody engineering.

Main Methods:

• Curated a dataset of Ab-Ag complexes with residue-level annotations.
• Enriched the dataset with UniProt sequence information, structural features, and physicochemical properties (pH, salt concentration).
• Developed flexible filtering options based on complex and residue-level descriptors.

Main Results:

• ANABAG includes comprehensive features essential for machine learning and physics-based modeling.
• The dataset facilitates the inclusion of glycosylation and varying solution conditions.
• Structural files are compatible with established antibody design and docking software.

Conclusions:

• ANABAG addresses limitations in current Ab-Ag datasets, enabling more accurate modeling.
• The dataset supports advanced antibody engineering by providing rich, actionable data.
• ANABAG is publicly available, promoting further research in Ab-Ag interactions.