Antibodies are proteins that play a critical role in our immune system for recognizing various antigens. Antibodies are increasingly used as diagnostic tools and therapeutic drugs, and an important component of the antibody engineering process is the availability of high quality three-dimensional structures. Experimental determination of antibody structures is laborious and not always possible. On the contrary, computational modeling provides a fast and inexpensive route for generating antibody structural models.
The variable domains (Fvs) of the heavy and light chain are of special interest in structural modeling, as they typically impart most or all of the specificity of an antibody for its antigen target. Modeling of the Fv region of an antibody involves prediction of both the framework (FR) region and the variable loop (CDR) region. Our protocol starts with a template search for the framework region (FR) in our curated antibody database, which is derived from the publically available crystal structures in the Protein Data Bank (PDB). After that, templates for the CDR region are chosen based on homology from canonical loop conformations that are clustered in the database. Structural models for the Fv region are then built along with refinement of the H3 loop. It is also possible to model the entire antibody including the constant region (Fc).