xBind: an integrated webserver for large language model-enabled cross-molecular protein binding site prediction.

xBind predicts cross-molecular protein binding sites using large language model embeddings with high configurability and interpretability.

• Why it matters: Accurate identification of protein binding sites is essential for understanding molecular interactions, but existing methods lack integration of LLM embeddings and flexible, user-friendly interfaces.
• What they did: The webserver employs LLM embeddings from ESM-2, combined with sequence and structure features, to predict residue-level binding sites for protein-protein, protein-DNA, and protein-RNA interactions, accepting FASTA or PDB/mmCIF inputs, and includes AlphaFold-based structure prediction and threshold calibration.
• The result: xBind offers interactive, customizable visualizations and analysis tools, enabling users to interpret binding likelihoods effectively, and is freely accessible online to facilitate molecular interaction research.

AI summary of the abstract · may contain errors · not medical advice · read the paper.