PEPPER – Protein complex Expansion using Protein-Protein intERaction networks – identifies meaningful pathways / complexes as densely connected sub-networks from seed lists of proteins derived from pull-down assays (i.e AP-MS…).

PEPPER resolves connection sub-graph discovery problems by using multi-objective optimisation involving two objective functions: (i) the coverage, a solution must contain as many proteins from the seed as possible, (ii) the density, a solution must contain as many interactions as possible. Since these objectives conflict, no single solution can be considered as dominating the others. To summarise the information from all solutions, PEPPER merges Pareto solutions into a final predicted protein complex by maximising the modularity using a greedy search. PEPPER further refines predictions by an integrated structure- and function-based post-processing pipeline to rank proteins that were added by the algorithm. This workflow orchestrates external data integration (GO annotations, known complexes matching) besides measuring connectivity features (topological coefficients).

Associated publication(s)

  • Winterhalter, C., Nicolle, R., Louis, A., To, C., Radvanyi, F., & Elati, M. (2014). Pepper: cytoscape app for protein complex expansion using protein–protein interaction networks. Bioinformatics, btu517.
  • Elati, M., To, C., & Nicolle, R. (2013, January). Multi-objective optimization for relevant sub-graph extraction. In International Conference on Learning and Intelligent Optimization (pp. 104-109). Springer Berlin Heidelberg.