Classification methods for complex diseases using molecular networks = 분자네트워크를 이용한 복합성 질환의 분류기법

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 317
  • Download : 0
DC FieldValueLanguage
dc.contributor.advisorLee, Do-Heon-
dc.contributor.advisor이도헌-
dc.contributor.authorLee, Eun-Jung-
dc.contributor.author이은정-
dc.date.accessioned2011-12-12T07:25:38Z-
dc.date.available2011-12-12T07:25:38Z-
dc.date.issued2008-
dc.identifier.urihttp://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=303564&flag=dissertation-
dc.identifier.urihttp://hdl.handle.net/10203/27064-
dc.description학위논문(박사) - 한국과학기술원 : 바이오및뇌공학과, 2008. 8., [ ix, 78 p. ]-
dc.description.abstractThe development of effective markers predicting various medical events such as disease occurrence, prognosis or treatment response is essential for precise diagnosis and delivery of tailored therapeutics to individual patients. Recently, an increasing number of disease markers have been identified through analysis of genome-wide expression profiles. Typically, marker genes are selected by measuring the power of their expression profiles to discriminate patients of different disease states. However, classification using those marker genes faces challenges to complex diseases due to cellular heterogeneity within a tissue sample and genetic heterogeneity across patients. In addition, redundant information in a set of marker genes selected independently to each other may lead to decreased classification performance because proteins in cells are known to function coordinately within protein complexes, signaling cascades, and other biological processes. In this thesis, functional modularity of genes is incorporated into disease classification procedure to address theses challenges. The proposed disease classification methods utilize human pathway databases or recently available human protein-protein interaction networks for module extraction. The activity of a module in each patient sample is calculated by summarizing expression levels of member genes in the module, and classifiers to predict the disease status of unknown samples are built based on the inferred activities of modules as a feature vector instead of the expression levels of individual markers genes. In the proposed pathway-based classification, modules are defined as a set of interacting genes in known human pathways collected from several public databases , and the markers in use are not encoded as individual genes or all member genes of pathways but as subsets of condition-responsive co-functional ``key genes``. For each pathway, an activity level is summarized from the gene expression lev...eng
dc.languageeng-
dc.publisher한국과학기술원-
dc.subjectmolecular network-
dc.subjectprotein-protein interaction-
dc.subjectpathway-
dc.subjectdisease classification-
dc.subjectPPI-
dc.subject분자 네트워크-
dc.subject단백질-단백질 상호작용-
dc.subject경로-
dc.subject질병 분류-
dc.subjectPPI-
dc.subjectmolecular network-
dc.subjectprotein-protein interaction-
dc.subjectpathway-
dc.subjectdisease classification-
dc.subjectPPI-
dc.subject분자 네트워크-
dc.subject단백질-단백질 상호작용-
dc.subject경로-
dc.subject질병 분류-
dc.subjectPPI-
dc.titleClassification methods for complex diseases using molecular networks = 분자네트워크를 이용한 복합성 질환의 분류기법-
dc.typeThesis(Ph.D)-
dc.identifier.CNRN303564/325007 -
dc.description.department한국과학기술원 : 바이오및뇌공학과, -
dc.identifier.uid020045871-
dc.contributor.localauthorLee, Do-Heon-
dc.contributor.localauthor이도헌-
Appears in Collection
BiS-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0