Roman will use powerful deep learning-based algorithms to design new candidates for a promising, but difficult to identify, class of drugs.
Proteins have varied tasks to fulfil within cells. To function appropriately, proteins need to adopt the correct three-dimensional structure. Misfolding of proteins is involved in up to 50% of human diseases during disease progression.
Pharmacological chaperones are a promising class of drugs to treat the progression of these diseases, but they are difficult to identify. As a 2024 Schmidt Science Fellow, Roman will employ powerful deep learning-based protein structure prediction algorithms to design pharmacological chaperones from scratch.