Accurate Antibody Molecular Modeling Enables Epitope Determination

Webinar description

The importance of antibody discovery and engineering for pharmaceutical development has grown significantly over the past two decades. During the past few years, modeling tools for antibody discovery and engineering have been developed, taking advantage of the structural knowledge existing in the protein database (PDB). The number of structures containing an antibody VH-VL motif is approximately 2350 today, compared to 1800 two years ago. However, the antibody repertoire diversity is around 10 billion, and specific maturation extends it to 1013.

At Roche, a modeling tool is automatically producing high quality models at a very low computational cost. The tool uses an antibody structural database dissected in heavy and light chains, frameworks, and Complement Determining Regions (CDRs) segments that enables novel features in searching templates during modeling. An accurate prediction of the relative VH-VL orientation ensures the model quality. The VH-VL orientation prediction is also used to predict the CDR grafting success during humanization. A technology based on neighborhood search is used to predict the position of side chains; this represents an add-on to the classical homology modeling (sequence-based) approach.

Attend this webcast to learn how to automate the generation of high-quality antibody structure models.

Key Learning Objectives:

  • Hear industry perspectives on the importance of molecular modeling in shortening the time to market of antibody drugs.
  • What tools are needed to engineer safe, efficacious, and stable antibody drugs?
  • Learn how to leverage these tools to automate the enumeration of high quality, accurate antibody structure models.


Guy Georges, Senior Principal Scientist/Molecular Modeling and Drug Design, Roche Innovation Center, Penzberg
Hugues-Olivier Bertrand, Sr. Director BIOVIA Pre-Sales – Senior Fellow & Head of the BIOVIA Science Council, Dassault Systèmes BIOVIA
Originally aired: