• Q-BioLiP represents protein structures as quaternary structures, capturing interactions across multiple chains that are missed by single-chain tertiary representations.
• It properly pairs DNA/RNA chains, addressing the issue of single-chain representation in BioLiP.
• The resource includes both experimental and predicted binding affinities, and retains both biologically relevant and irrelevant interactions to reduce misclassification.
• A new quaternary structure-based algorithm for protein–ligand complex modeling is provided, enhancing the utility for structure-based studies.