Integrated Clinical and Transcriptomic Profiling to Characterize Disease Phenotype

项目摘要 外显子组和全基因组测序正在成为癌症中越来越常规的方法[1]， 疾病[2]和罕见病诊断。[3]尽管他们取得了成功，但我们完全解释临床相关性的能力 个人基因组变异的研究仍然存在重大差距[4-6]。考虑到这一点，最关键的需要是更多 将遗传变异与疾病病因联系起来的基因型-表型数据。这项建议的目的是 改善遗传变异的临床解释;特别是通过开发综合方法， 预测遗传变异对临床表型的影响。这一建议涉及的假设，支持 通过初步数据，将患者转录组学数据与基因型和临床数据相结合（而不是 每一个单独的）提供了一个更好的机制理解疾病的自然史，从最初的表现， 进展 特定的目标被设计成使得每个目标独立地添加大量的功能基因组信息， 在先前可用的患者遗传数据之上，以进一步解析临床表型。要求1 建立了一个全面和广泛共享的患者转录组（和遗传）变异数据集， 体细胞获得性肿瘤患者的多种癌症、心血管和血栓/出血表型 骨髓增生性肿瘤（MPN）和选择其他罕见的遗传性血液病（HBD）。目标2有条不紊 确定MPN和HBD临床相关亚组之间的差异RNA表达和加工 患者目标3将这些元素结合在一起-并应用两个集成的贝叶斯和机器学习 方法，RIVER[24]（RNA信息变体对调控的影响）和LASSO[25]（最小绝对收缩和 选择算子），以解决罕见变异的功能和临床相关性;并识别最常见的特征 预测疾病风险或进展。 这些目标的完成将为如何整合转录组数据提供新的科学知识 改善其他遗传（和罕见）疾病的临床基因组分析。此外，该项目将使 首席研究员，发展基因组医学信息学和数据科学方面的专业知识 补充了她目前在生物物理学、生物化学和转化血液学方面的背景。组合 在斯坦福大学通过课程、研讨会、一对一咨询等方式进行额外的信息学培训 从项目导师，并与更广泛的统计，生物信息学和基因组学社区的互动，这 该项目将为首席研究员在基因组医学领域开展独立的学术生涯做好准备。

PROJECT SUMMARY Exome and whole-genome sequencing are becoming increasingly routine approaches in cancer[1], common disease[2]and rare disease diagnosis.[3] Despite their success, our ability to fully interpret the clinical relevance of personal genome variation remains a significant gap[4-6]. Considering this, the most crucial need is more genotype-phenotype data that link genetic variation with disease causation. The objective of this proposal is to improve the clinical interpretation of genetic variation; in particular, by developing integrative approaches that predict the effect of genetic variation on clinical phenotype. This proposal addresses the hypothesis, supported by preliminary data, that combining patient transcriptomic data with genotypic and clinical data (as opposed to each alone) offers a better mechanistic understanding of disease natural history, from initial presentation to progression. The specific aims are designed such that each independently add substantial functional genomic information, over and above previously available patient genetic data, to further resolve the clinical phenotype. Aim 1 establishes a comprehensive and widely-shared dataset of patient transcriptomic (and genetic) variation across multiple cancer, cardiovascular and thrombosis/bleeding phenotypes, in patients with somatically-acquired myeloproliferative neoplasms (MPN) and select other rare heritable blood diseases (HBD). Aim 2 methodically determines differential RNA expression and processing between clinically-relevant subgroups of MPN and HBD patients. Aim 3 brings these elements together – and applies two integrative Bayesian and machine learning approaches, RIVER[24] (RNA-informed variant effect on regulation) and LASSO[25] (Least Absolute Shrinkage and Selection Operator), to resolve the functional and clinical relevance of rare variants; and identify signatures most predictive of disease risk or progression. Completion of these aims will contribute new scientific knowledge on how integrating transcriptomic data improves clinical genomic analyses in other genetic (and rare) diseases. In addition, this project will enable the Principal Investigator to develop expertise in the informatics and data science aspects of genomic medicine that complement her current background in biophysics, biochemistry and translational hematology. Combined with additional informatics training at Stanford University through coursework, seminars, one-on-one advising from project mentors, and interactions with the wider statistics, bioinformatics and genomics communities, this project will prepare the Principal Investigator to launch an independent academic career in genomic medicine.