海量临床数据表明疾病之间存在着各种共病现象，即两种疾病共同存在。人们据此提出“具有共病现象的疾病间应具有共同的分子机理”假设。我们利用疾病间共享生物通路，共享基因，对应的蛋白质间的互作以及发生共病现象的疾病基因间的共表达等四个层次证实了这个假设。本项目以此假设为基础，拟对一些病因不明的疾病开展相关分子机制的研究。我们首先整合临床电子病历中的疾病共病关系，利用文本挖掘进一步扩充疾病的共病关系，建立疾病的共病网络。其次将整合的疾病相关基因的数据与疾病共病网络相结合，基于“Guilt By Association”原理，利用模拟退火算法寻找与疾病相关的新基因。随后结合蛋白质相互作用网络，生物途径，基因表达谱，疾病表型网络等多层次数据对挖掘出的与疾病相关的新基因进行更深入的筛选。最后针对每个疾病共病对，基于蛋白质相互作用网络和通路分析来探索其潜在的共病机制，挖掘导致共病现象的具多效性的基因.

Comorbidity refers to the co-occurrence of two diseases disorders or illnesses. Evidences from massive electronic medical records show that many comorbidities exist between diseases, which make people propose that the diseases with comorbidity should share the same underlying mechanism. We confirmed this hypothesis with multilevel data, including sharing biological pathways, sharing genes, direct interactions between proteins encoded by related disease genes and the coexpression between genes associated with comorbidity diseases. Based on this principle, this project plans to explore the underlying mechanisms for those undiagnosed diseases. We first generate comorbidity network by integrating comorbidity pairs extracted from electronic medical records and expanding comorbidity pairs with text mining approach from published papers; we then integrate the disease associated genes with comorbidity network, and identify new disease associated genes with simulation annealing algorithm based on “Guilt by Association ”. We further evaluate the results and eliminate those false positive associated genes with multilevel data from protein interaction network, biological pathway, gene coexpression and disease phenotype network. Last, for each comorbidity pair, we explore their underlying mechanism and identify those new pleiotropic genes associated with comorbidity pair through protein interaction network and biological pathway analysis.