logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-02132025-131258


Tipo di tesi
Tesi di laurea magistrale
Autore
RONCOLI, ANDREA
URN
etd-02132025-131258
Titolo
Impact of Pre-trained Representations in Machine Learning-Based Protein-Ligand Docking
Dipartimento
INFORMATICA
Corso di studi
INFORMATICA
Relatori
relatore Prof. Bacciu, Davide
correlatore Prof. AlQuraishi, Mohammed
Parole chiave
  • biology
  • docking
  • drug discovery
  • machine learning
  • pretraining
Data inizio appello
28/02/2025
Consultabilità
Completa
Riassunto
This work analyzes possible avenues for improving performance and generalization of machine learning-based models for protein-ligand docking. Learned representations from models pre-trained on different chemical tasks are leveraged by introducing them as features in the molecular graphs, which generative docking models are then trained on.
File