Predicting Protein-Ligand binding Residue with Token classification

Accurately predicting ligand-binding residues in proteins is crucial for understanding molecular interaction mechanisms, providing valuable insights that drive drug discovery and design. A lot of studies for the prediction of protein-ligand binding residues have been actively conducted in machine learning. The computational methods for predicting protein-ligand binding residues can be classified into sequence-based and structure-based approaches. Although structure-based methods have shown better performance than sequence-based methods for predicting protein residues where ligands bind. The structure-based methods rely on protein structure data, making their application to large-scale protein datasets both time- and resource-intensive. Due to this limitation, sequence-based approaches have recently gained attention. In this paper, we propose a new machine learning model for predicting ligand-binding residues in proteins using token classification techniques from the natural language processing field.