Protein-protein interaction network evaluation for identifying potential drug targets

Area of Science:

• Computational Biology
• Bioinformatics
• Drug Discovery

Background:

• The rise of antibiotic resistance renders traditional single-target drug strategies increasingly ineffective.
• Pathogenic microorganisms utilize complex protein-protein interaction (PPI) networks for survival and virulence.
• Disrupting essential pathways within these networks offers a potential avenue for novel antimicrobial therapies.

Purpose of the Study:

• To develop novel computational strategies for identifying multiple drug targets in pathogenic PPI networks.
• To find minimal sets of proteins (without human orthologs) that disrupt critical pathogenic pathways or complexes.
• To optimize the trade-off between disrupted pathways and the number of targeted proteins.

Main Methods:

• Formulation of the problem as finding minimum vertex cuts in PPI networks to disrupt pathogenic pathways.
• Development of polynomial-time algorithms for the 'sparsest cut' problem with approximation factors of |S| and O(√n).
• Application of algorithms to the Escherichia coli PPI network using KEGG signaling pathways and analyzing network partitioning.

Main Results:

• Algorithms identified strategies to disrupt three essential signaling pathways in E. coli by targeting only three proteins.
• Two of the identified essential protein targets were confirmed as essential proteins.
• An alternative approach identified 28 potential drug targets in E. coli, including four known drug targets, by partitioning the PPI network.

Conclusions:

• Novel computational approaches can effectively identify multiple drug targets in pathogen PPI networks.
• These strategies offer a more robust alternative to single-target approaches for combating antibiotic resistance.
• The identified targets and network partitioning methods provide valuable insights for future antimicrobial drug development.