PSICHIC: physicochemical graph neural network for learning protein-ligand interaction fingerprints from sequence data

Abstract

In drug discovery, determining the binding affinity and functional effects of small-molecule ligands on proteins is critical. Current computational methods can predict these protein-ligand interaction properties but often lose accuracy without high-resolution protein structures and falter in predicting functional effects. We introduce PSICHIC (PhySIcoCHemICal graph neural network), a framework uniquely incorporating physicochemical constraints to decode interaction fingerprints directly from sequence data alone. This enables PSICHIC to attain first-of-its-kind emergent capabilities in deciphering mechanisms underlying protein-ligand interactions, achieving state-of-the-art accuracy and interpretability. Trained on identical protein-ligand pairs without structural data, PSICHIC matched and even surpassed leading structure-based methods in binding affinity prediction. In a library screening for adenosine A1 receptor agonists, PSICHIC discerned functional effects effectively, ranking the sole novel agonist within the top three. PSICHIC's interpretable fingerprints identified protein residues and ligand atoms involved in interactions. We foresee PSICHIC reshaping virtual screening and deepening our understanding of protein-ligand interactions.

Multimodal protein representation learning and target-aware variational auto-encoders for protein-binding ligand generation

“…Compared with structure-based and complex-based methods, the model shows comparable performances across all metrics with lower standard deviations. Furthermore, the model performs on par with other more sophisticated state-of-the-art methods, namely PSICHIC [58] and TankBind. It is worth noting that PSICHIC also leverages the residue-level embeddings extracted from the pre-trained ESM; however, Koh et al [58] use these embeddings to construct 2D graphs of proteins.…”

“…Furthermore, the model performs on par with other more sophisticated state-of-the-art methods, namely PSICHIC [58] and TankBind. It is worth noting that PSICHIC also leverages the residue-level embeddings extracted from the pre-trained ESM; however, Koh et al [58] use these embeddings to construct 2D graphs of proteins. This does not preserve the SE(3)-symmetry (rotations and translations), which is an important property in learning three-dimensional structures.…”

Multimodal pretraining for unsupervised protein representation learning

“…The incorporation of these diverse datasets forms a comprehensive and varied tested, allowing us to thoroughly assess the predictive capabilities of our multimodal representation when estimating protein–ligand binding affinities. To ensure a fair and standardized evaluation, we meticulously followed the test/training/validation split settings as outlined in previous studies, specifically adhering to the configurations defined in the respective sources for the DAVIS, KIBA, and PDBbind version 2020 datasets [ 15 , 43 ]. By maintaining this consistency, we aimed to create a level playing field for comparisons, allowing for an equitable assessment of our multimodal representation’s performance.…”

Multimodal Pretraining for Unsupervised Protein Representation Learning

“…The incorporation of these diverse datasets forms a comprehensive and varied tested, allowing us to thoroughly assess the predictive capabilities of our multimodal representation when estimating proteinligand binding affinities. To ensure a fair and standardized evaluation, we meticulously followed the test/training/validation split settings as outlined in previous studies, specifically adhering to the configurations defined in the respective sources for the DAVIS, KIBA, and PDBbind v2020 datasets [57, 43]. By maintaining this consistency, we aimed to create a level playing field for comparisons, allowing for an equitable assessment of our multimodal representation’s performance.…”