Departmental seminar on 26th April 2023

April 26, 2023 ( 14:00 – 15:30 )

Add to Calendar 04/26/2023 14:00 04/26/2023 15:30 Europe/Prague Departmental seminar on 26th April 2023

We have a guest, Prof. David Hoksza from Charles University, Faculty of Mathematics and Physics. The topic of his lecture is Machine learning-based detection of protein-ligand binding sites.

The presentation will cover machine-learning based approaches for protein-ligand binding sites prediction. Firstly, our state-of-the-art approach which detects binding sites based on the 3D structure of a protein will be discussed. The method utilizes a mesh-based representation of protein surface, over which a machine-learning model is trained for subsequent detection. In addition, metrics used binding site prediction methods evaluation, along with their advantages and disadvantages, will be discussed. Then, our recent work will be introduced where we will show the performance of the recently developed protein language models in predicting positions of binding sites solely from protein sequence and how it compares to existing state-of-the-art methods.

The lecture will take place in person, in the hall CH3, Faculty of Science, Hlavova 8, Praha 2.

We have a guest, Prof. David Hoksza from Charles University, Faculty of Mathematics and Physics. The topic of his lecture is Machine learning-based detection of protein-ligand binding sites.

The presentation will cover machine-learning based approaches for protein-ligand binding sites prediction. Firstly, our state-of-the-art approach which detects binding sites based on the 3D structure of a protein will be discussed. The method utilizes a mesh-based representation of protein surface, over which a machine-learning model is trained for subsequent detection. In addition, metrics used binding site prediction methods evaluation, along with their advantages and disadvantages, will be discussed. Then, our recent work will be introduced where we will show the performance of the recently developed protein language models in predicting positions of binding sites solely from protein sequence and how it compares to existing state-of-the-art methods.

The lecture will take place in person, in the hall CH3, Faculty of Science, Hlavova 8, Praha 2.