Protein Self-Interaction in Cellular Function and Network Evolution: Molecular Mechanisms, AI-Driven Insights, and Therapeutic Potentials
Protein self-interaction to form homodimers and higher-order homo-oligomers is a ubiquitous phenomenon fundamental to living organisms. Recent structural, system-level, and computational insights reveal that self-interacting proteins dictate the specificity, rewiring, and topological complexity of cellular signaling networks and macromolecular assemblies. Beyond their physiological roles, aberrant or dysregulated homotypic interactions disrupt cellular proteostasis, driving the formation of toxic non-native oligomers, pathological amyloid fibrillization, or aberrant liquid–liquid phase separation transitions linked to neurodegenerative and systemic diseases. This review provides a comprehensive overview of the state-of-the-art experimental methodologies, including proximity labeling, as well as the advanced computational frameworks, such as deep learning architectures and protein language models, used to map the structural dynamics of SIPs. Furthermore, we dissect the evolutionary trajectories of SIPs within protein–protein interaction networks, which are underpinned by dosage-balance constraints, and highlight their diverse functional advantages, ranging from allosteric modulation to biomolecular condensation. Finally, we summarize the molecular mechanisms linking pathological self-associations to human disorders, underscoring the emerging paradigm of targeting homotypic interfaces as a promising frontier for precision therapeutics.