结核分枝杆菌基因组重注释研究

发布时间：2018-08-20 14:32

【摘要】：结核病每年都对全世界人民的身体健康造成严重的危害,其中结核分枝杆菌(Mycobacterium tuberculosis)是结核病的病原菌。虽然关于结核分枝杆菌基因组学研究取得了很大的进展,在基因组公共数据库中已有结核分枝杆菌全基因组的注释信息,但随着时间的推移,越来越多新的基因功能信息被添加进数据库,这其中就可能包含当初对结核分枝杆菌进行注释时未被使用的具有序列相似性的基因。在基因组分析时,这些新添加的基因功能信息可能为一些假想的基因提供功能转移来源。同时一些原始注释中未包含的基因也可能通过与新添加的基因功能信息进行相似性比对而被发现。本文将针对以上问题,以最新的基因组公共数据库为标准,通过基因相似性比对以及基于从头预测的方法发现新基因等方法,对结核分枝杆菌基因组信息进行重新注释研究。本研究的方法可为其他物种的基因组重注释工作提供参考。本研究的主要内容为:1.以Z曲线理论方法为基础,从原始基因注释中筛选出从属于功能已知的蛋白质编码基因(第一类基因)作为正样本,并用第一类基因的随机洗牌序列生成负样本。以正负样本为训练集,通过基于五重交叉验证的Fisher模型确定假设基因(第二类基因)中非编码的部分,即原注释中错误注释的基因。2.使用Prodigal和Zcurve两款基因预测软件对结核分枝杆菌基因组进行基因预测,将基因预测结果与原始基因组注释进行对比,挑选出重叠率低的候选新基因进行Blast序列比对,应用设定的筛选参数挑选出满足条件的新基因,并为新基因添加具体的功能注释信息。3.在基因重注释过程中,需要研究人员进行手动筛选,当有大量的基因组需要重注释,尤其是从Blast结果中筛选出满足条件的新基因时,将会是一个非常繁重的工作。因此本研究还使用PHP开发了一套能够自动进行基因组重注释的Web工具,在实现重注释工具化减少人工手动筛选工作量的同时,也能大大提高基因重注释的效率。
[Abstract]:Tuberculosis causes serious harm to the health of people all over the world every year. Among them, Mycobacterium tuberculosis (Mycobacterium tuberculosis) is the pathogen of tuberculosis. Although much progress has been made in the study of the genomics of Mycobacterium tuberculosis, the annotated information on the entire genome of Mycobacterium tuberculosis is available in the genome public database, but over time, More and more new genetic functional information has been added to the database, which may contain sequence-like genes that were not used in the initial annotation of Mycobacterium tuberculosis. In genome analysis, these new gene function information may provide a functional transfer source for some hypothetical genes. At the same time, some genes not contained in the original annotation may be found by comparing with the newly added gene function information. In order to solve the above problems, we will reannotate the genome information of Mycobacterium tuberculosis by means of gene similarity comparison and new gene discovery based on ab initio prediction. The method of this study can be used as a reference for genome reannotation of other species. The main contents of this study are: 1: 1. Based on the Z curve theory, a protein encoding gene (the first type gene) with known function is selected from the original gene annotation as a positive sample, and a negative sample is generated by random shuffling sequence of the first type gene. Taking positive and negative samples as training set, the non-coding part of hypothetical gene (second type gene) is determined by Fisher model based on quintuple cross validation, that is, the wrong annotated gene. 2 in the original annotation. Prodigal and Zcurve were used to predict the genome of Mycobacterium tuberculosis. The results of gene prediction were compared with the original genome annotation, and the candidate genes with low overlap rate were selected for Blast sequence alignment. The new genes that meet the conditions were selected by using the selected screening parameters, and specific functional annotation information was added to the new genes. In the process of gene reannotation, it is necessary for researchers to carry out manual screening. When there are a large number of genomes that need to be re-annotated, especially when new genes are selected from Blast results to meet the requirements, it will be a very heavy task. Therefore, this study also developed a set of Web tools which can automatically reannotate genome by using PHP, which can reduce the manual screening workload and improve the efficiency of gene reannotation greatly.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R378.911

【相似文献】