Pages 444–448
Abstract
Research summary supplied by the authors
In graph theory and data mining technique, detection of quasi-clique helps to find densely connected structures in a network. Protein-protein interactions (PPI) are at the core of the entire interaction system of any living cell. We can view a PPI network (PPIN) as a complex system of proteins which are linked by interactions among themselves. It can also be used to optimize leather supply chains by finding hidden collaboration clusters among suppliers like tanners, identifying risk communities, or mapping customer segments for targeted green initiatives. A huge complex graph is formed by raw material suppliers, tanneries, chemical suppliers (e.g., chromium sulphate), manufacturers, and retailers. These highly dependent and interconnected groups of suppliers and manufacturers can be revealed by quasi-cliques. It helps in risk management and development of alternative strategies. We can also form a graph of product preferences of customer and their interactions, purchase behaviour. Here finding quasiclique can identify hidden or overlapping communities of customers with similar, dense purchasing patterns. It helps to industry to adapt specific product lines like certain leather types, finishes and designs by specific quasi-cliques, optimizing production planning and reducing excessive inventory costs. In this paper, we present a novel heuristic approach to mine all possible maximal quasi-cliques (functionally associated proteins) from the complex and sparse human PPIN and identify significant modules of co-functional proteins. Information mining from the PPIN assists in rational drug design by targeting significant hub proteins in the network. Keywords— Protein-protein interaction, leather supply chain optimization. network analysis, quasi-clique