疾病间相关关系的研究及其研究方法的开发

发布时间：2018-04-03 22:17

本文选题：疾病相似性　切入点：疾病网络　出处：《华东理工大学》2015年博士论文

【摘要】：如何从疾病组织和正常组织的表达谱数据中挖掘转录调控机制并比较其间的差异一即差异调控-成为人们迫切想揭示的问题之一。本文基于差异共表达分析结果,对差异共表达分析结果进行调控关系的鉴定,并将差异调控信息进行全方位的图形化展示,最后,对差异调控基因进行重要性排序,将更有能力捕获差异共表达基因或基因对的差异调控基因给予更高的重要性打分,以上方发被成功实现为R工具包-DCGLv2(本工作对应本文第2章内容)。如今,人们开始关注疾病之间错综复杂的关系,因为这不仅有助于了解疾病谱的全貌,还提供了全新的视角进行疾病病因、发病机制的研究,以及药物的开发和治疗策略的探索。利用上述开发的用于疾病组织和正常组织间的差异调控分析工具-DCGLv2,我们在108种疾病中鉴定出1,326对显著的疾病相关关系,并发现由差异共表达属性得到的疾病间相关关系比由差异表达属性得到的疾病间相关关系更符合已知的分子生物学发现,同时我们还对来自同一组织的多种疾病和来自不同组织的同一疾病进行了分析,发现疾病间相似性同时受疾病的种类和发病组织两方面影响。另外,我们通过一个子疾病网络详细示例了如何挖掘疾病关系对中共有的失调机制,且试图证明共有的失调机制是引起疾病的共同原因(本工作对应本文第3章内容)。为了方便研究者使用自产数据挖掘疾病间相关关系,我们将上述鉴定疾病间相似性(即疾病相关关系)的方法实现为DSviaDRM工具包。该工具包主要包含了五个鉴定疾病相关关系的函数(DCEA、DCpathway、DS、comDCGL和comDCGLplot),为首个基于基因转录组数据间功能失调机制的相似性来鉴定疾病间相关关系的工具,为临床医生和生物医学研究者提供了研究疾病的工具(本工作对应本文第4章内容)。另外,在基因表达研究领域,位点特异性表达参与了许多重要的生物学机制,也越来越受到人们重视。本文介绍了两种基于二代测序数据鉴定位点特异性表达的方法(pDNAar和mREF),并比较了两种方法的优劣,发现在理想条件下mREF更高效,这项工作为今后位点特异性表达的鉴定提供了方法学支持(本工作对应本文第5章内容)。总体而言,我们开发了一种能精确挖掘差异调控信息的R工具包-DCGLv2,并利用DCGL v2计算得到大量疾病的差异调控信息,然后比较疾病间差异调控信息的相似度来衡量疾病的相似度,并最终将鉴定疾病间相似度的方法实现为另一个R工具包--DSviaDRM。另一方面,我们比较了两种鉴定位点特异性表达的方法,为今后的位点特异性表达的研究提供了支持。
[Abstract]:How to excavate the transcriptional regulation mechanism from the expression profile data of disease tissues and normal tissues and compare the difference between them, namely differential regulation, has become one of the urgent problems that people want to reveal.Based on the results of differential co-expression analysis, this paper identifies the regulatory relationship of differential co-expression analysis results, and displays the differential regulation information in all directions. Finally, the importance of differentially regulated genes is ranked.The higher importance of differentially regulated genes, which are more capable of capturing differentially expressed genes or pairs of genes, was successfully implemented as R Toolkit -DCGLv2 (this work corresponds to Chapter 2 of this paper).Today, people are beginning to pay attention to the intricate relationship between diseases, because it not only helps to understand the full picture of disease spectrum, but also provides a new perspective to study the etiology and pathogenesis of diseases.And the development of drugs and the exploration of treatment strategies.Using the above developed differential regulation analysis tool DCGLv2 for disease tissues and normal tissues, we identified 1326 pairs of significant disease-related relationships in 108 diseases.It was also found that the disease correlation obtained from differential co-expression attributes was more consistent with known molecular biological findings than the disease correlation obtained from differential expression attributes.At the same time, we also analyzed many diseases from the same tissue and the same disease from different tissues. It was found that the similarity between diseases was affected by both the disease types and the diseased tissues.In addition, we use a sub-disease network to illustrate how to discover the common maladjustment mechanism in disease relationship pairs, and try to prove that common maladjustment mechanism is the common cause of disease. (this work corresponds to the contents of Chapter 3 of this paper.In order to facilitate researchers to use native data to mine the correlation between diseases, the above method of identifying disease similarity (i.e. disease correlation) is implemented as DSviaDRM Toolkit.The toolkit consists of five functions to identify disease-related relationships, DCEAV DCpath way DS.com DCGL and comDCGLplotl. It is the first tool to identify disease-related relationships based on the similarity of dysfunctional mechanisms between gene-transcriptional data.It provides tools for clinicians and biomedical researchers to study diseases (this work corresponds to Chapter 4 of this paper).In addition, in the field of gene expression, locus specific expression has participated in many important biological mechanisms, and has been paid more and more attention.In this paper, two site-specific expression methods based on second-generation sequencing data were introduced, and the advantages and disadvantages of the two methods were compared. It was found that mREF was more efficient under ideal conditions.This work provides methodological support for the identification of site-specific expression in the future (this work corresponds to Chapter 5 of this paper).In general, we developed a R Toolkit-DCGLv2, which can accurately mine differential regulation information, and calculate a large number of disease differential control information by DCGL v2, then compare the similarity of differential regulation information between diseases to measure disease similarity.Finally, the method of identifying the similarity between diseases is realized as another R toolkit, DSvia DRM.On the other hand, we compared two methods of locus specific expression, which provided support for the future research of locus specific expression.
【学位授予单位】：华东理工大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：R3416;Q811.4

【相似文献】