Protein Language Models to incorporate PKPD considerations in early antibody design and generation

Starting with the disclosure of Alphafold and the ability to use protein language models for structure prediction, this deep learning approach hasbeen of great interest for design of antibody and protein therapeutics. This talk will discuss two use cases where such models are combined with PKPD considerations in early antibody design and generation. In one example, protein language models and statistical machine learning models are combined to enable risk assessment for fast clearance based on sequence data alone. This talk also discusses how traditional PKPD models can inform protein-language model based antibody design and optimization.