The human disease network in terms of dysfunctional regulatory mechanismsReport as inadecuate

The human disease network in terms of dysfunctional regulatory mechanisms - Download this document for free, or read online. Document in PDF available to download.

Biology Direct

, 10:60



BackgroundElucidation of human disease similarities has emerged as an active research area, which is highly relevant to etiology, disease classification, and drug repositioning. In pioneer studies, disease similarity was commonly estimated according to clinical manifestation. Subsequently, scientists started to investigate disease similarity based on gene-phenotype knowledge, which were inevitably biased to well-studied diseases. In recent years, estimating disease similarity according to transcriptomic behavior significantly enhances the probability of finding novel disease relationships, while the currently available studies usually mine expression data through differential expression analysis that has been considered to have little chance of unraveling dysfunctional regulatory relationships, the causal pathogenesis of diseases.

MethodsWe developed a computational approach to measure human disease similarity based on expression data. Differential coexpression analysis, instead of differential expression analysis, was employed to calculate differential coexpression level of every gene for each disease, which was then summarized to the pathway level. Disease similarity was eventually calculated as the partial correlation coefficients of pathways’ differential coexpression values between any two diseases. The significance of disease relationships were evaluated by permutation test.

ResultsBased on mRNA expression data and a differential coexpression analysis based method, we built a human disease network involving 1326 significant Disease-Disease links among 108 diseases. Compared with disease relationships captured by differential expression analysis based method, our disease links shared known disease genes and drugs more significantly. Some novel disease relationships were discovered, for example, Obesity and cancer, Obesity and Psoriasis, lung adenocarcinoma and S. pneumonia, which had been commonly regarded as unrelated to each other, but recently found to share similar molecular mechanisms. Additionally, it was found that both the type of disease and the type of affected tissue influenced the degree of disease similarity. A sub-network including Allergic asthma, Type 2 diabetes and Chronic kidney disease was extracted to demonstrate the exploration of their common pathogenesis.

ConclusionThe present study produces a global view of human diseasome for the first time from the viewpoint of regulation mechanisms, which therefore could provide insightful clues to etiology and pathogenesis, and help to perform drug repositioning and design novel therapeutic interventions.

ReviewersThis article was reviewed by Limsoon Wong, Rui Wang-Sattler, and Andrey Rzhetsky.

KeywordsHuman disease network Disease similarity Dysfunctional regulation mechanism Differential coexpression analysis Differential regulation analysis AbbreviationsDCEADifferential coexpression analysis

DEADifferential expression analysis

DCE-basedDifferential coexpression based

DE-basedDifferential expression based

dCDifferential coexpression value

DCGDifferentially coexpressed gene

DCLDifferentially coexpressed link

DEGDifferentially cexpressed gene

DDLDisease-Disease link

MeSHMedical Subject Headings

ICD-10International Classification of Diseases 10th revision

DODisease Ontology

WDWithin-network distance

PBMCPeripheral blood mononuclear cell

Electronic supplementary materialThe online version of this article doi:10.1186-s13062-015-0088-z contains supplementary material, which is available to authorized users.

Download fulltext PDF

Author: Jing Yang - Su-Juan Wu - Wen-Tao Dai - Yi-Xue Li - Yuan-Yuan Li


Related documents