IMA

IMA identifies protein complexes within protein-protein interaction (PPI) networks by constructing weighted PPI networks that integrate topological and biological properties to reduce noise and applying a memetic algorithm to detect complexes.


Key Features:

  • Weighted Network Construction: Combines topological and biological attributes to construct a weighted PPI network that reduces false positive and false negative interactions.
  • Initial Population Formation: Integrates various clustering results to form initial populations for the evolutionary search.
  • Fitness Function Design: Uses a fitness function based on five distinct topological properties of protein complexes.
  • Memetic Hybrid Strategy: Employs a hybrid memetic algorithm combining genetic algorithms with local optimization to perform global and local searches.
  • Selection Operator: Selects promising candidates from the population based on fitness scores.
  • Recombination Operator: Combines selected individuals to generate offspring and promote genetic diversity.
  • Local Optimization Strategy: Applies local optimization techniques to refine candidate complexes.
  • Updating Population Operator: Updates the population with newly generated and optimized candidates to maintain a dynamic search process.

Scientific Applications:

  • Protein complex detection: Detection of protein complexes from PPI networks, reported to outperform baseline and state-of-the-art methods.
  • PPI topology analysis and cellular mechanism inference: Characterization of multiple topological structures within PPI networks to uncover intricate biological interactions and inform understanding of cellular mechanisms.

Methodology:

Construction of a weighted PPI network by integrating topological and biological attributes; formation of initial populations by combining multiple clustering results; fitness evaluation based on five topological properties; and optimization via a memetic approach combining genetic algorithm operators (selection and recombination) with local optimization and population updating.

Topics

Details

License:
Not licensed
Tool Type:
workflow
Operating Systems:
Windows
Programming Languages:
Python
Added:
6/15/2022
Last Updated:
6/15/2022

Operations

Publications

Wang R, Ma H, Wang C. An Improved Memetic Algorithm for Detecting Protein Complexes in Protein Interaction Networks. Frontiers in Genetics. 2021;12. doi:10.3389/fgene.2021.794354. PMID:34970305. PMCID:PMC8712950.