Novel Statistical methods for DNA Sequencing Data, and applications to Autism.

摘要 人类遗传学的主要问题之一是理解遗传原因。 潜在的复杂表型，包括自闭症等神经精神特征 谱系障碍和精神分裂症。尽管在过去几年里做了大量的工作 几十年来，大多数情况下对潜在的生物学机制知之甚少。 高通量、大规模并行基因组技术的最新进展 彻底改变了人类遗传学领域，并承诺将导致重要的科学 预付款。尽管在数据生成方面取得了这些进展，但分析仍然非常具有挑战性 并对这些数据进行解读。这项建议的主要重点是发展强大的 全基因组测序数据与RICH整合的统计方法 功能基因组学数据，目标是改进涉及到基因的发现 自闭症谱系障碍。我们建议整合来自多个不同来源的数据， 包括来自Encode、Roadmap和 心理编码，来自GTEx、心理编码和普通思维的eQTL数据 来自大型遗传变异数据库的数据，如ExAC和 GnomAD，以预测遗传变异在非编码遗传中的功能效应 以特定于组织和细胞类型的方式显示区域，并生成跨 人体中的大量组织和细胞类型，然后我们可以用它们来识别 全基因组测序研究中与自闭症的新关联。建议数 功能预测和功能图将在流行的 ANNOVAR数据库。我们还建议将这些函数预测用于 三大全基因组测序中近2万个全基因组的分析 自闭症的研究。我们相信拟议的研究是非常及时的，并具有 有可能大大改进对非编码遗传变异的分析，因此 提供对自闭症潜在风险的生物机制的新见解，以及更多 泛指其他神经精神疾病。

Summary One of the major problems in human genetics is understanding the genetic causes underlying complex phenotypes, including neuropsychiatric traits such as autism spectrum disorders and schizophrenia. Despite tremendous work over the past few decades, the underlying biological mechanisms are poorly understood in most cases. Recent advances in high-throughput, massively parallel genomic technologies have revolutionized the field of human genetics and promise to lead to important scientific advances. Despite this progress in data generation, it remains very challenging to analyze and interpret these data. The main focus of this proposal is the development of powerful statistical methods for the integration of whole-genome sequencing data with rich functional genomics data with the goal to improve the discovery of genes involved in autism spectrum disorders. We propose to integrate data from many different sources, including epigenetic data from projects such as ENCODE, Roadmap, and PsychENCODE, eQTL data from the GTEx, PsychENCODE and CommonMind consortia, data from large scale databases of genetic variation such as ExAC and gnomAD, in order to predict functional effects of genetic variants in non-coding genetic regions in a tissue and cell type specific manner, and generate functional maps across large number of tissues and cell types in the human body that we can then use to identify novel associations with autism in whole-genome sequencing studies. The proposed functional predictions and functional maps will be broadly available in the popular ANNOVAR database. We further propose to use these functional predictions in the analysis of almost 20,000 whole genomes from three large whole genome sequencing studies for autism. We believe that the proposed research is very timely and has the potential to substantially improve the analysis of non-coding genetic variation, and hence provide new insights into the biological mechanisms underlying risk to autism, and more broadly to other neuropsychiatric diseases.