Representing structural haplotypes and complex genetic variation in pan-genome graphs

项目摘要 人类泛基因组测序的初始阶段已经导致了一个 各种各样的基因组。同时，一个复杂的计算生态系统 开发了一种方法来组织泛基因组，一种图形数据结构， 反映了全球人口的多样性，以及所需的序列分析方法， 遗传学家使用泛基因组，以改善他们的研究是如何进行相对于一个 单一参考基因组。泛基因组揭示了人类遗传的重要因素 变化量特别地，存在相当数量的序列多样性和新颖性。 泛基因组中的重复DNA序列。因为最初的方法 用于分析泛基因组变异的基因组模型是针对相对简单的变异而创建的 在重复DNA之外，有必要开发新的方法来发现，基因分型， 组织和验证基因组重复区域的变异。本分析的范围 从几百个碱基长的短的重复DNA序列， 包括基因。我们将专门开发方法来发现罕见的变异， 可变数目串联重复序列，并进行旁系同源特异性发现， 基因的拷贝数变异。这些方法将被开发来分析短读 对数据进行排序，以便大规模数据集（如TOPMed生成的数据集）可以 这些方法的优点是改善其群组中的变体发现。我们还将 开发方法，以改善重复或重排序列的代表性， 泛基因组的图示。这将通过对 重复序列的进化关系，同时构建图形，并验证 使用来自单细胞测序技术的公共数据集的组装组织， 链测序我们所有的开发都将与其他成员合作进行， 人类泛基因组参考联盟。我们将分享变异发现的方法， 其他研究人员正在研究大群体。最后，泛基因组的任何改进 将与生产和其他小组协调发布图表， 标准化的泛基因组图，供公众中的其他研究人员进行基础研究。

Project Summary The initial phase of sequencing the human pangenome has resulted in the assembly of a diverse collection of genomes. In parallel, an ecosystem of sophisticated computational methods were developed to organize the pangenome a graphical data structure that efficiently reflects the diversity of a global population, as well as sequence analysis methods required for geneticists to use the pangenome to improve how their studies are performed relative to a single reference genome. The pangenome revealed important factors about human genetic variation. In particular, there is a considerable amount of sequence diversity and novel sequences in the pangenome that arise from repetitive DNA. Because the initial methods developed to analyze variation in the pangenome were created for relatively simplistic variation outside of repetitive DNA, it is necessary to develop novel methods to discover, genotype, and organize, and validate variation in repetitive regions of the genome. The scope of this analysis spans short repeated DNA sequences that are hundreds of bases long, to entire regions that encompass genes. We will specifically develop methods to discover rare variation in variable-number tandem repeat sequences, and perform paralog-specific discovery of copy-number variation of genes. These methods will be developed to analyze short-read sequencing data so that large scale datasets such as those generated by TOPMed can take advantage of these methods to improve variant discovery in their cohorts. We will additionally develop methods to improve the representation of repetitive or rearranged sequences in the graphical representation of the pangenome. This will be accomplished by modeling the evolutionary relationships of repetitive sequences while building the graph, and validating assembly organization using public datasets from the single-cell sequencing technique, Strand-Seq. All of our development will be performed collaboratively with other members of the Human Pangenome Reference Consortium. We will share methods for variant discovery with other researchers who are studying large cohorts. Finally, any improvements in the pangenome graph will be released in coordination with production and other groups so that there is a standardized pangenome graph for other researchers in the public to base research from.