Neutralizing antibodies are important for the body to fight any virus and might also offer a weapon against COVID. They are classically extracted from sera of recovered individuals, and ongoing trials a promising. Given the current numbers, however, it is unlikely that production can meet the needs. Another option to produce antibodies is to clone the variable regions into cell lines and let them produce the antibody. This can be scaled up easily and is already a routine for biological therapies.

The problem

Even if we are so lucky to have already an antibody (isolated from a recovered patient) that has the potential to neutralize Sars-Cov-2 viruses, this protective effect might diminish due continuously occurring mutations in the viral genome. Even if neutralized epitopes are usually less prone to mutations, since those mutations might prevent the virus entering the cell; a strategy targeting one single epitope only, favors the development of resistance.

Our proposal

In addition to exploring if critically ill patients of people at risk could benefit from a therapeutic antibodies possessing variable Immunoglobulin genes of the above cell, we suggest to evaluate additional antibody structures as well that could bind to the ACE-2 exposed surface and use a combination of therapeutic antibodies.

Progress during the weekend

We utilized pyRosetta and Snugdock to compile a pipeline that would enable screening for binding affinity of a high number of seuĆ³quence variations. We also built a Docker image (docker://fotakisg/antibody-evolver) to support parallelization of the docking task.

Challenges I ran into

In order tho screen a large number of candidate antibody sequences, we have to automatically annotated them for the variable regions. There are several nomenclatures to number the amino acid positions of these regions and one has to be aware of numbering used in the original structure.

What I learned

I learned a lot about structural modeling with Rosetta and docking simulations.

What's next for Antibody evolver

The Docker image is ready for mass testing of sequences. After integrating it with an evolutionary algorithm framework a large variety of antibody sequences could be tested for the ability of binding Sars-Cov-2 RBD.

How can it help after Crona is over

The approach can be generalized to any antigen and thus, might be able to help research regarding other virus infections as well.

Built With

Share this project: