Data-driven biological network alignment that uses topological, sequence, and functional information.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Traditional network alignment (NA) assumes topological similarity indicates functional relatedness, but this is often inaccurate.
• Recent findings show functionally unrelated proteins can be topologically similar.
• A data-driven approach, TARA, was developed to learn functional relatedness from network and protein data.

Purpose of the Study:

• To develop an improved data-driven network alignment method (TARA++) that integrates both within-network topological and across-network sequence information.
• To enhance protein functional prediction accuracy by leveraging diverse data types.

Main Methods:

• Adapted social network embedding techniques for biological network alignment.
• Integrated within-network topological information with across-network sequence information.
• Developed a data-driven approach (TARA++) that learns functional relatedness.

Main Results:

• TARA++ significantly outperforms existing network alignment methods in protein functional prediction accuracy.
• The new method demonstrates the effectiveness of combining topological and sequence information.
• The data-driven approach learns nuanced relationships between network features and protein function.

Conclusions:

• Combining knowledge from different research domains, like network topology and sequence data, is a promising strategy.
• Improvements in protein functional prediction have significant biomedical implications, aiding research in areas like cancer progression and aging.
• The TARA++ method offers a more accurate tool for understanding protein function and biological networks.