Antibody Virtual Screening
Our researchers have developed a proprietary methodology and pipeline for in silico or virtual antibody screening. While working with Generative AI methods of antibody design, we recognized that we needed to filter down from a larger set of potential candidates. We developed a proprietary, robust and effective virtual screening method and pipeline that allows us to evaluate a large library of antibody candidates to arrive at a ranked order list of hits.
Virtual screening in the antibodies domain has a number of challenges. In silico protein-protein interaction analysis is difficult due to the lack of appropriate protein structure availability. Other challenges include the complex interactions between amino acid domains, the presence of large and flat interface regions, and computational complexity and cost in evaluating the context of the network in which these interactions occur. While Alphafold can predict the structure of a given protein sequence, questions on accuracy persist in addition to the time and computational cost that quickly derail it as an alternative for large scale virtual screening.
Virtual screening for antibodies requires integrating a number of complex steps. It involves identifying and focusing on an epitope that needs to be screened against a large set of paratopes i.e, CDRs. Finding the appropriate epitope on the antigen is a task that involves excluding bound regions and regions of PTM modifications. Screening for millions of CDR sequences rapidly to identify a few hits is an arduous task requiring time and computational cost. Once the hits are identified further evaluation of the entire antibody against a target antigen needs to be performed.
We developed a proprietary and robust method and pipeline that allows us to evaluate a large library of antibody candidates to arrive at a ranked order list of hits. Our technology can not only screen on binding affinity but also incorporate other developability properties such as immunogenicity, hydrophobicity, specificity, aggregation and isoelectric point. This comprehensive in silico multi-attribute screening reduces the time and cost of screening antibody candidates.
We have the team and computational capabilities (research scientists, AI engineers, AI models, software tools) that can work with you on the specific screening needs you have.