One of the fastest growing branches in the pharmaceutical industry is drug design based on antibodies, with antibody (Ab) fragments likely to be the next important class of therapeutics. These protein pharmaceuticals promise extremely high specificity and the ability to use the body’s own immune system to kill e.g. tumors, but due to their size and complexity their computational design is challenging.
Key challenges are:
- to reliably predict 3D structures of antibodies, in particular the Complementarity-Determining Regions (CDRs)
- to model their binding mode and understand how structure and dynamics are altered in this process
- to improve their binding affinity through systematic amino acid mutations
Why is this project particularly interesting for BioExcel?
This project aims to address all these challenges through an integrative approach combining the core BioExcel applications GROMACS, HADDOCK and PMX, and with development of computational workflows to facilitate execution on Exascale high-performance computing resources.
This project is expected to be of particular interest to the pharma industry. BioExcel already has established contacts with several companies involved in antibody design who are using BioExcel core software and who stand to benefit from the project, and to achieve maximum impact we are eager to ensure that work done is informed by concrete cases of interest to industry.
What are we doing in BioExcel?
We will use GROMACS to improve antibody 3D structures and accurately sample multiple alternative conformations. The resulting conformations will be used as input in HADDOCK to model the interaction with its target antigen. The modelling of the complex using HADDOCK will be guided by a variety of experimental and predicted data. Models of the complexes will be optimized by MD using GROMACS to select a final model as input for antibody engineering using PMX to improve its binding affinity.
The designed antibodies from the HADDOCK/GROMACS pipeline will be further subjected to the single point mutation analysis by means of molecular dynamics based free energy calculations run by the PMX team. At this step a large-scale high-accuracy computational scan of amino acid mutations will be performed aimed at increasing antibody-antigen binding affinity. The procedure will be iterated if needed to identify key residues responsible for the largest changes in the binding free energy.
Utrecht University, KTH Royal Institute of Technology, Max Planck Gesellschaft, Norman Consulting