基于转录组数据的药物重定位

发布时间：2018-04-20 05:11

本文选题：转录组学 + 共表达基因集富集分析　；参考：《国防科学技术大学》2016年博士论文

【摘要】：虽然计算药物重定位分析已经鉴定了大量重定位药物,但是它们的临床试验鲜有进展。究其原因,除知识产权问题外,整合的计算药物重定位数据库匮乏阻碍了实验药物开发人员有效地获取并转化计算分析结果;现有的基于转录组药物重定位方法分析疾病环境下药物作用方式的“黑箱”属性是计算重定位药物在预临床和临床试验阶段失败的重要原因;基因排序效果欠佳是基于转录组药物重定位假阳性结果出现的关键原因之一。这些问题反映了构建开放的药物重定位数据库、鉴定疾病核心靶基因、探索疾病环境下的药物作用方式和发展高效基因排序方法在药物重定位中的重要性。本文提出了共表达基因集富集分析方法(co-expressed gene set enrichment analysis,cogena),实现并发布了Bioconductor软件包。该方法结合基因共表达信息和多种不同功能的基因集进行富集分析,可以更深层次地多角度理解基因的功能。首先将cogena应用于帕金森综合症的共表达通路分析、靶标发现及基于靶标的药物重定位中,鉴定了39个帕金森综合症核心靶基因,得到了靶向其中13个基因的药物,包括具有神经保护作用的quercetin、cholic acid和oprozomib。第二,提出了如果一个药物基因表达谱与疾病的关键共表达基因群表达谱反相关,那么该药可能用于治疗该疾病的假说,基于该假说设计了cogena的药物重定位和药物作用方式分析流程,并将其应用于一种疾病——银屑病的研究,计算得到了2个具有不同作用方式的已用于治疗银屑病的药物methotrexate和ciclosporin,预测了若干治疗银屑病的候选药物及其作用方式,例如细胞周期阻断剂resveratrol、ciclopirox、etoposide和trifluridine。另一方面将该分析流程应用于一种药物——methotrexate的药物重定位中,鉴定了与cogena的银屑病分析结果高度一致的细胞周期阻断剂药物,得到了methotrexate已经用于治疗的6种疾病,预测了该药新的适应症。第三,整合7275个药物和疾病基因表达谱,基于参数的基因集富集分析方法计算它们之间的相似性,构建了可视化的交互式数据库和网络服务器D2Dpage,从多个角度验证了D2Dpage在药物重定位方面的用途和效果。在银屑病的药物重定位分析中,计算得到了银屑病抗体类药物etanercept,鉴定了超药品说明书使用的银屑病药物pimecrolimus。D2Dpage是目前唯一基于基因表达谱的包含大量药物和疾病的药物重定位数据库。第四,结合上述分析开展了银屑病的转化系统生物学研究,通过整合分析银屑病共表达基因簇、抗体类药物基因表达谱和自身免疫病基因表达谱之间的相关性及对抗体类药物转录组进行cogena药物重定位分析,计算验证了已上市药物etanercept和ixekizumab治疗银屑病的有效性,分析发现未上市药物brodalumab和guselkumab可以用于治疗银屑病,基于抗体类药物转录组计算得到了methotrexate和超药品说明书使用的银屑病药物isotretinoin,也预测了若干治疗银屑病的候选药物,这些候选药物与基于银屑病转录组的cogena药物重定位分析结果高度一致,发现了鲁棒地存在于银屑病、抗体类药物和多种自身免疫病转录组的关键共表达基因聚类簇,鉴定了32个银屑病核心靶基因,并基于靶基因进行药物重定位分析,预测了若干银屑病候选药物,包括靶向TOP2A的etoposide。最后,针对RNA-Seq转录组数据,实现了判别非负矩阵分解DNMF的R软件包,提出了基于DNMF的基因排序模型,证明了基于DNMF的基因排序分析方法优于当前通用的方法。
[Abstract]:Although a large number of repositioning drugs have been identified, their clinical trials have rarely progressed. The reason, in addition to intellectual property rights, is that the lack of integrated computational drug repositioning databases impedes the effective acquisition and conversion of experimental drug developers by experimental drug developers; the existing transcriptional drugs are based on the existing drugs. The "black box" attribute of drug action in the disease environment is an important reason to calculate the failure of repositioning drugs in pre clinical and clinical trials, and the poor gene sequencing results are one of the key reasons for the emergence of false positive results of relocation of drugs in the transcriptional group. These problems reflect the construction of an open drug weight. Locating the database, identifying the core target genes of the disease, exploring the importance of drug action in the disease environment and developing the efficient gene sequencing method in the relocation of drugs. This paper proposes a co expressed gene collection and enrichment analysis method (co-expressed gene set enrichment analysis, cogena), and implements and releases the Bioconductor software package. Method combined with gene co expression information and a variety of different functions of gene set to enrich and analyze, we can understand the function of gene in a deeper and multi angle. First, cogena is applied to the analysis of co expression pathway of Parkinson syndrome, target discovery and target based drug relocation, identified 39 core targets of Parkinson syndrome. A drug that targets 13 genes, including neuroprotective quercetin, cholic acid and oprozomib. second, suggests that the drug may be used to treat the disease if the gene expression profile of a drug is anti related to the gene group expression of the key co expression of the disease, and based on this hypothesis, the cogena is designed. The process of drug relocation and drug action analysis is applied to the study of a disease, psoriasis, and 2 drugs, methotrexate and ciclosporin, which have different roles in the treatment of psoriasis, have been calculated, and a number of candidates for the treatment of psoriasis and their modes of action, such as cell cycle interruption, are predicted. The agent resveratrol, ciclopirox, etoposide and trifluridine., on the other hand, applied the analysis process to a drug relocation of a drug, methotrexate, identified the cell cycle blocker, which was highly consistent with the results of cogena's psoriasis analysis, and obtained 6 diseases that had been used in the treatment of methotrexate and predicted the new drug. Third, third, integrate the gene expression profiles of 7275 drugs and diseases, and calculate the similarity between them based on the genetic enrichment analysis method based on the parameter. A visual interactive database and a network server D2Dpage are constructed to verify the use and effect of D2Dpage in the relocation of drugs from various angles. In the repositioning analysis, the psoriasis antibody drug Etanercept was calculated and the psoriasis drug pimecrolimus.D2Dpage used in the super drug instruction was identified as the only drug relocation database containing a large number of drugs and diseases based on the gene expression spectrum. Fourth, the transformation system of psoriasis was carried out in combination with the above analysis. According to the analysis of the correlation between the gene expression profiles of the psoriasis, the correlation between the antibody drug gene expression profiles and the gene expression profiles of the autoimmune disease, and the cogena relocation analysis of the antibody drug transcriptional group, the effectiveness of etanercept and ixekizumab in the treatment of psoriasis has been verified. Listed drugs brodalumab and guselkumab can be used for the treatment of psoriasis. The psoriatic drug isotretinoin used in methotrexate and super drug instructions is calculated based on the antibody drug transcriptome, and a number of candidates for the treatment of psoriasis are also predicted. These candidates are relocated to the cogena based drugs based on the psoriasis transcriptional group. The results are highly consistent. It is found that the key co expression gene cluster of psoriasis, antibody drugs and various autoimmune diseases transcriptional groups has been identified, and 32 core target genes of psoriasis are identified, and the drug relocation analysis based on the target gene is used to predict some of the candidate drugs for silver chip disease, including the final etoposide. target of the target TOP2A. According to the data of RNA-Seq transcriptional group, a R software package to distinguish DNMF from non negative matrix decomposition is implemented, and a gene sequencing model based on DNMF is proposed. It is proved that the method of gene sequencing analysis based on DNMF is superior to the current general method.

【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：R96

【相似文献】