MOTIVATION: Unraveling the human interactome to uncover disease-specific patterns and discover drug targets hinges on accurate protein-protein interaction (PPI) predictions. However, challenges persist in machine learning (ML) models due to a scarcity of quality hard negative samples, shortcut learning, and limited generalizability to novel proteins. RESULTS: In this study, we introduce a novel approach for strategic sampling of protein-protein noninteractions (PPNIs) by leveraging higher-order network characteristics that capture the inherent complementarity-driven mechanisms of PPIs. Next, we introduce Unsupervised Pre-training of Node Attributes tuned for PPI (UPNA-PPI), a high throughput sequence-to-function ML pipeline, integrating unsupervised pre-training in protein representation learning with Topological PPNI (TPPNI) samples, capable of efficiently screening billions of interactions. By using our TPPNI in training the UPNA-PPI model, we improve PPI prediction generalizability and interpretability, particularly in identifying potential binding sites locations on amino acid sequences, strengthening the prioritization of screening assays and facilitating the transferability of ML predictions across protein families and homodimers. UPNA-PPI establishes the foundation for a fundamental negative sampling methodology in graph machine learning by integrating insights from network topology. AVAILABILITY AND IMPLEMENTATION: Code and UPNA-PPI predictions are freely available at https://github.com/alxndgb/UPNA-PPI.

Extensive experimental information supports the formation of ligand-specific conformations of G protein-coupled receptors (GPCRs) as a possible molecular basis for their functional selectivity for signaling pathways. Taking advantage of the recently published inactive and active crystal structures of GPCRs, we have implemented an all-atom computational strategy that combines different adaptive biasing techniques to identify ligand-specific conformations along pre-determined activation pathways. Using the prototypic GPCR β2-adrenergic receptor as a suitable test case for validation, we show that ligands with different efficacies (either inverse agonists, neutral antagonists, or agonists) modulate the free-energy landscape of the receptor by shifting the conformational equilibrium towards active or inactive conformations depending on their elicited physiological response. Notably, we provide for the first time a quantitative description of the thermodynamics of the receptor in an explicit atomistic environment, which accounts for the receptor basal activity and the stabilization of different active-like states by differently potent agonists. Structural inspection of these metastable states reveals unique conformations of the receptor that may have been difficult to retrieve experimentally.