Network Alignment in Biological Systems

Network alignment is the comparative mapping of nodes and edges across biological networks to identify conserved functional and structural modules. By aligning protein–protein interaction networks or multilayer representations of molecular activity, researchers reveal core pathways that persist across species, support annotation transfer and guide translational applications. Global alignment seeks a comprehensive one-to-one correspondence that maximises both sequence and topological similarity, while local alignment focuses on highly conserved subgraphs or complexes. Heterogeneous network alignment further incorporates diverse molecule types and interaction modalities, reflecting the complex interplay of genes, proteins, metabolites and regulatory elements. Recent advances leverage graph-based heuristics, ensemble strategies and machine learning to reconcile computational tractability with biological fidelity. Embedding methods translate network topology into continuous vector spaces, enabling cross-species functional inference and drug repurposing. These approaches collectively enhance our understanding of evolutionary conservation, species-specific innovation and the molecular basis of disease.

Research from Nature Portfolio

One study introduced a unified framework that integrates leading global aligners to produce complete mappings of protein–protein interaction networks. By harmonising multiple algorithms, the new tool achieves topological and biological coherence beyond what individual methods offer, and yields soft-clusterings that refine annotation transfer. Another work extended classical alignment methods to heterogeneous networks by developing coloured graphlets that account for node and edge types. This innovation improves edge conservation and robustness to noise in biological data by enabling alignments that respect molecular diversity. A third investigation focused on local alignment of heterogeneous networks, formulating the problem in terms of node-coloured graphs. The proposed algorithm uncovers conserved subnetworks across multi-modal data, delivering high-quality alignments validated on real-world examples of molecular interplay.

Network Alignment in Biological Systems publication trend

The graph below shows the total number of articles in network alignment in biological systems across all publications each year (not limited to Nature Index journals).

Technical terms

Network alignment: The process of finding a mapping between nodes and edges of two or more biological networks to reveal conserved modules.

Protein–protein interaction (PPI) network: A graph where nodes represent proteins and edges represent physical or functional interactions.

Global alignment: A network alignment approach that seeks a one-to-one mapping across entire networks.

Local alignment: A network alignment strategy that focuses on aligning highly similar subgraphs or complexes.

Heterogeneous network: A network containing multiple types of nodes or edges, representing diverse biomolecules or interaction modalities.

Embedding: A representation of network nodes in a continuous vector space that preserves topological properties for machine learning applications.

Graphlet: A small, connected subgraph used to characterise local topology in network analysis.

References

1. Unified Alignment of Protein-Protein Interaction Networks. Scientific Reports (2017).
2. From homogeneous to heterogeneous network alignment via colored graphlets. Scientific Reports (2018).
3. L-HetNetAligner: A novel algorithm for Local Alignment of Heterogeneous Biological Networks. Scientific Reports (2020).
4. Joint embedding of biological networks for cross-species functional alignment. Bioinformatics (2023).
5. Multilayer network alignment based on topological assessment via embeddings. BMC Bioinformatics (2023).
6. Boosting-based ensemble of global network aligners for PPI network alignment. Expert Systems with Applications (2023).