Computational Methods for Sequence Alignment, Genotyping, and Diploid Genome Assembly

项目总结/摘要 大规模测序正在彻底改变生物学研究和临床实践。在过去的几十年里，项目 例如1,000个基因组计划，TCGA，GTEx和GEUVADIS已经产生了数百万亿个 阅读。最近完成的英国100 K WGS项目激励了许多其他国家开发其 拥有10万个WGS项目。通量的提高和测序成本的降低使得更多的 对癌症、遗传性疾病和人类生物学的其他领域进行彻底和深入的研究。先进 测序比对和计算方法在进行这些研究中发挥了重要作用。 分析。近年来，我们的实验室为这些全球规模和前所未有的努力做出了贡献， 开发了几种广泛使用的用于分析NGS测序读数的生物信息学工具：TopHat 2和 用于比对RNA-seq读数的HISAT，用于鉴定基因融合的TopHat-Fusion， 宏基因组测序读数，用于人类基因组规模的图形比对的HISAT 2，以及HISAT- HLA基因分型和装配。 该提案解决了序列比对、基因分型和 二倍体基因组组装首先，我们计划研究和开发各种索引策略。几乎所有 比对程序依赖于一种类型的索引来将读段与参照进行比对。对准精度和速度 将通过纳入其他类型的索引来进一步增强。第二，我们将开发基因分型， 二倍体基因组组装算法。随着测序成本的不断下降， 为了临床目的对他们自己的基因组进行测序。我们将进一步开发我们的初始版本， HISAT基因型转化为一套全面的工具，可以基因型和组装一个人的整个二倍体 一天之内就能在桌面上完成一个基因组。第三，我们将继续维护和改进HISAT 2，并开发新的 更灵活的校准器。我们建议通过开发几种通用的比对程序来统一广泛使用的比对程序。 校准程序的功能（输入处理、索引、校准和报告）作为模块， 应用程序编程接口（API），公开这些模块，使生物信息学工程师， 使用API开发自己的索引和对齐算法，这些算法是为最佳分析而定制的 自己的数据集。我们计划展示新的序列比对器SARTOR（Sequence 比对库优化参考引导分析），通过有效处理不同类型的读取 (WGS、WES、RNA-seq、ChIP-seq、BS-seq等）通过不同的测序技术（短，长， 链接读取）。成功完成后，拟议的软件系统将促进个性化医疗 通过利用定制的个人基因组，其关键功能包括差异基因表达 分析和体细胞突变鉴定。该计划还将允许研究人员进行公正， 在大规模NGS实验中进行准确、快速的分析。

Project Summary/Abstract Massive sequencing is revolutionizing biological research and clinical practice. Over the past decades, projects such as the 1,000 Genomes Project, TCGA, GTEx, and GEUVADIS have generated hundreds of trillions of reads. The recent completion of the UK’s 100K WGS project has inspired many other nations to develop their own 100K WGS projects. The improvements in throughput and reduced costs of sequencing have enabled more thorough and deeper studies of cancer, genetic disorders, and other areas of human biology. Advanced sequencing alignment and computational methodologies have played a major role in conducting these analyses. In recent years, our lab has contributed to these global scale and unprecedented endeavors by developing several widely used bioinformatics tools for analyzing NGS sequencing reads: TopHat2 and HISAT for aligning RNA-seq reads, TopHat-Fusion for identifying gene fusions, Centrifuge for classifying metagenomics sequencing reads, HISAT2 for graph alignment at the human genome scale, and HISAT- genotype for HLA gene typing and assembly. This proposal addresses several key challenges in the areas of sequence alignment, genotyping, and diploid genome assembly. First, we plan to research and develop various indexing strategies. Virtually all alignment programs rely on one type of index for aligning reads to a reference. Alignment accuracy and speed will be further enhanced by incorporating additional types of indexes. Second, we will develop genotyping and diploid genome assembly algorithms. As sequencing costs continue to decline, it will become routine for people to have their own genomes sequenced for clinical purposes. We will further develop our initial version of HISAT-genotype into a comprehensive suite of tools that can genotype and assemble a person’s whole diploid genome in one day on a desktop. Third, we will continue to maintain and improve HISAT2, and develop a new more versatile aligner. We propose to unify widely used alignment programs by developing several common functions of alignment programs (input processing, indexing, aligning, and reporting) as modules and provide application programming interfaces (APIs) that expose those modules, enabling bioinformatics engineers to use the APIs for developing their own indexes and alignment algorithms that are customized for best analyzing their own data sets. We plan to demonstrate the usability of the new sequence aligner, SARTOR (Sequence Alignment Repertoire To Optimize Reference-guided analysis), by effectively handling different types of reads (WGS, WES, RNA-seq, ChIP-seq, BS-seq, etc.,) produced by different sequencing technologies (short, long, and linked reads). Upon successful completion, the proposed software systems will promote personalized medicine by drawing upon customized personal genomes, with key functionalities including differential gene expression analysis and somatic mutation identification. The programs will also allow researchers to perform unbiased, accurate, and rapid analyses in large-scale NGS experiments.