The network characteristics of disease-associated genes across different protein-protein interaction networks

Thumbnail Image
Younis, Dawoud
Saleh, Asim
Journal Title
Journal ISSN
Volume Title
Although human interactome is incomplete but it has reached good coverage that allows us to study disease associated genes characteristics on it. Each disease consists of proteins and physical interaction between them within human interactome. Study of these characteristics is necessary due to know the main proteins of the disease which are hubs, this could lead us to analyze each disease and conclude diseases relation by knowing shared genes between them. Fourteen different PPI networks were studied and analyzed to study the robustness of these characteristics. These networks as any others of human interactome are incomplete and in order to avoid noise as well as getting more accurate results we reposed on edge removal with different percentages and compare them. For different levels of completeness of these networks we realized that the connectivity significance of disease genes slowly drops as more and more links are removed. Conversely, this trend indicates that the predictive power of the connectivity significance should continuously increase as the interactome becomes more and more complete. Case study was on gastrointestinal diseases mapped to integrated network of all networks with at least two evidence for each edge within integrated network. Analyzing and Visualizing of this case study allow us to show that most seed proteins can be removed without considerably changing the resulting outcome but there are a small number of nodes whose removal results in a drastic change of the final outcome.