Annotation and interpretation of loss-of-function polymorphisms in human genomes

描述（由申请人提供）：已知人类蛋白质编码基因中的功能丧失（LoF）突变在囊性纤维化和肌肉萎缩症等严重疾病中发挥重要作用，并且最近也被证明影响1型糖尿病和克罗恩病等复杂疾病的风险。作为1000基因组计划的一部分，我们最近对人类LoF变异进行了迄今为止最大的系统调查，该调查证实了LoF变异对人类疾病研究的价值，并确定了检测和解释这些变异的关键挑战。我们建议通过构建强大、准确的LoF变体注释、表征和高通量基因分型工具来克服这些挑战。首先，我们将开发一个集成的信息管道（注释of LoF Transcripts， ALoFT），用于识别和过滤所有类型的LoF变异，包括单核苷酸替换（snp），插入和删除。其次，我们将利用RNA测序实验和疾病突变数据库的数据，建立更准确的预测模型，预测遗传变异对基因表达和剪接的影响，以及它们引起疾病的概率。最后，我们将把ALoFT管道和上述预测模型应用于超过30,000个人类外显子组和基因组测序，这些测序是nih资助的其他项目的一部分，并将得到的注释作为可公开访问的LoF变体数据库（dbLoF）的基础。我们将使用我们的功能注释来开发加权关联测试，并将其应用于这些测序个体中新的疾病风险变异的发现。此外，我们将使用在这些样本中鉴定的LoF变异目录来设计一个定制的基因分型阵列，允许快速、经济地对人类队列中大多数常见的人类LoF变异进行查询，允许在单独资助的关联研究中评估这些变异的表型效应。这项研究将为发现和描述自然功能丧失变异以及探索它们与人类疾病风险的潜在关联提供强大的工具。

DESCRIPTION (provided by applicant): Loss-of-function (LoF) mutations in human protein-coding genes are known to play a major role in severe diseases such as cystic fibrosis and muscular dystrophy, and have also recently been shown to influence the risk of complex diseases such as type 1diabetes and Crohn's disease. We have recently conducted the largest systematic survey to date of human LoF variants, as part of the 1000 Genomes Project, which has confirmed the value of LoF variants for human disease studies and also identified key chalenges for the detection and interpretation of these variants. We propose to overcome these challenges by constructing robust, accurate tools for the annotation, characterization and high-throughput genotyping of LoF variants. Firstly, we will develop an integrated informatic pipeline (Annotation of LoF Transcripts, ALoFT) for the identification and filtering of all classes of LoF variant, including single nucleotide substitutions (SNPs), insertions and deletions. Secondly, we will exploit data from RNA sequencing experiments and disease mutation databases to create more accurate predictive models of the efects of genetic variants on gene expression and splicing, and of their probability of disease causation. Finally, we will apply the ALoFT pipeline and the predictive models described above to over 30,000 human exomes and genomes sequenced as part of other NIH-funded projects, using the resulting annotation as the basis for a publicly accessible database of validated LoF variants, dbLoF. We will use our functional annotation to develop a weighted association test and apply this to the discovery of novel disease risk variants in these sequenced individuals. In addition, we will use the catalogue of LoF variants identified in these samples to design a custom genotyping array permitting rapid, cost-effective interrogation of the majority of common human LoF variants in human cohorts, allowing the phenotypic effects of these variants to be assessed in separately funded association studies. This study will provide powerful tools for discovering and characterizing natural loss-of-function variants, and for exploring their potential association with human disease risk.