Dafina Marin-Gerdes

Dafina is a member of the Marianski Lab, where the focus is developing computational methods to improve antibody-based diagnostics and therapeutics. Her thesis project centers on identifying and modeling surface-accessible lysine environments on monoclonal antibodies to enable site-selective labeling. By applying molecular dynamics simulations, geometric deep learning tools such as MaSIF, and structure prediction platforms, she aims to classify lysine residues by their full local context—paving the way toward more predictable and reproducible antibody conjugates.
Outside of the lab, Dafina enjoys spending time with her family: cooking, walking, reading, and working on creative projects together.