Bioinformatics research and genomic sequence alignment in particular have revolutionized the understanding of how cells work. Genomic sequence alignment identifies regions of similarity between sequences of individual genes that are a likely consequence of functional or evolutionary relationships between the sequences. However, genes and other biomolecules in the cells do not function in isolation. Instead, they interact with each other to keep one alive. And this is exactly what biological networks (BNs) model: in BNs, biomolecules are represented as nodes and physical or functional interactions between the biomolecules are represented as edges. Thus, BN alignment, which aims to identify topologically and functionally similar regions between BNs of different species, is promising to give further insights into organizational principles of life, evolution, disease, and therapeutics. For example, it could guide the transfer of biological knowledge across species between the conserved (aligned) network regions. This is important, since many nodes in BNs are currently functionally uncharacterized even for well-studied model species.
Intellectual Merit: This project aims to address several issues with the current view of the problem of network alignment. First, it will develop a novel framework for fair evaluation of existing BN alignment methods, which is currently lacking. Second, it will redefine the problem of network alignment to allow for directly optimizing the amount of conserved network topology, which current methods fail to do. Third, since different types of BNs exist that capture different functional slices of the cell, and since the existing methods can align only homogeneous networks, ignoring any node or edge types, this project will extend the proposed methods to allow for alignment of heterogeneous networks encompassing the different BN types. The proposed methods will be used in two novel interdisciplinary collaborative applications: 1) studying the role of yeast S. cerevisiae and human proteasomes responsible for protein degradation, and 2) studying pathogenicity and drug resistance of malaria parasites from the Plasmodium family.
Broader Impact: BN alignment has broad applications. For example, it can be used to transfer biological knowledge from well annotated to poorly annotated species between similar network regions or to infer species' phylogenetic and evolutionary relationships based on similarities of their BNs. Besides computational biology, this project may impact other domains as well. For example, network alignment can de-anonymize online social networks and thus impact user privacy. Since network research spans many domains, a free open-source software tool implementing the proposed methods will be offered to researchers from diverse disciplines. The software will also serve as an educational tool.
