A Systematic Review of Topology-based Models for Protein–Protein Interaction Networks: Methods, Architectures, and Future Research Directions

Abstract

Protein–Protein Interaction Networks (PPINs) have emerged as a fundamental framework for understanding complex biological systems, enabling the modelling of cellular processes through graph-theoretic representations. In recent years, topology-based models have played a crucial role in uncovering structural and functional properties of PPINs, facilitating applications such as disease gene prediction, drug target identification, and functional annotation. This systematic review explores the evolution of topology-based models in PPINs, focusing on methodologies, computational architectures, and emerging trends between 2018 and 2023. The study categorizes models into classical graph-theoretic approaches, probabilistic and statistical models, and advanced deep learning-based frameworks, including graph neural networks. Additionally, it analyses key topological properties such as degree distribution, centrality measures, modularity, and network motifs that underpin biological insights. Comparative analysis highlights the strengths and limitations of each model, particularly in terms of scalability, robustness, and biological interpretability. The review also identifies challenges such as data incompleteness, noise, and integration of heterogeneous biological datasets. Finally, future research directions are proposed, emphasizing hybrid models, explainable AI, and multi-modal biological network integration. This review provides a comprehensive resource for researchers aiming to advance topology-based modeling in PPINs.