Thema der Dissertation:
Graph Representation Learning of Molecules for de novo Drug Development Thema der Disputation:
Structure-based Drug Design using 3D Equivariant Diffusion Model
Graph Representation Learning of Molecules for de novo Drug Development Thema der Disputation:
Structure-based Drug Design using 3D Equivariant Diffusion Model
Abstract: Advances in deep learning have facilitated the optimization and design of molecules to
accelerate drug discovery for tasks like property prediction, retrosynthesis planning or structure-based
ligand design.
This talk explores graph representation learning in pharmaceutical research, with a focus on de novo
small molecule design.
The first topic introduces group-equivariant architectures for processing 3D molecular data and its
employment for molecule generation using diffusion models.
The second topic discusses targeted ligand design guided by protein context and user-defined
properties addressing current limitations. Strategies to overcome data limitations and enhance model
generalization will also be discussed.
accelerate drug discovery for tasks like property prediction, retrosynthesis planning or structure-based
ligand design.
This talk explores graph representation learning in pharmaceutical research, with a focus on de novo
small molecule design.
The first topic introduces group-equivariant architectures for processing 3D molecular data and its
employment for molecule generation using diffusion models.
The second topic discusses targeted ligand design guided by protein context and user-defined
properties addressing current limitations. Strategies to overcome data limitations and enhance model
generalization will also be discussed.
Zeit & Ort
14.02.2025 | 10:00