In silico discovery of signaling pathway

• Main Authors: Kuang-Yu, Sher, 佘光堉
• Other Authors: Ueng-Cheng, Yang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/04165731397489101790

碩士 === 國立陽明大學 === 生物醫學資訊研究所 === 96 === Pathway regulates cellular functions and mechanism. The researches about biomedical are actually related to pathway. Traditionally, to discover a pathway, researchers first identify the components that are involved in a pathway. Then, the interactions of these components are uncovered. Only the order of the interactions is specified, a complete pathway is defined. To save time and effort, some bioinformatics approaches are proposed. Most of them integrated microarray expression data and protein-protein interaction (PPI) data set. However, these methods could only detect components of inducible pathway, which express differentially in different condition. There is no general way for discovering constitutive expressed pathway components. Here, we proposed a method for adjust these problems. We used protein set involved in a pathway as input because this was the first-step to discover pathway. And we evaluated relationship by interactions between pair of proteins. To decide the order, we defined start nodes or end nodes in the interaction network composed of our input proteins. According to the shortest path length to start nodes and end nodes, the order of sequential interactions were defined. If we missed some pathway proteins in the input, we included nearest neighbors’ PPI for collecting other pathway components and noise reduction methods were applied. In brief, we have developed a method which took gene set as input. In six human pathways partial gene set (80% participant included), our method had F-measure: 0.36, Precision: 0.31, Recall: 0.48. Comparing with previous two methods in same data set, we showed our method was good at reducing noise.