Mutational signature analysis: methods and applications to the clinic

项目摘要 突变标记分析是一种用于解释基因组中体细胞突变的最新分析方法。 它利用点突变的序列背景以及拷贝数和结构突变的大小和类型。 畸变将观察到的突变模式分解为不同的“签名”，其中一些已经被 与特定的生物过程有关。在这个项目中，我们将开发更强大和敏感的COM- 突变签名分析的推定方法，并将其应用于几个临床上重要的问题。 我们的初步工作已经引起了临床医生的极大关注，我们将分析来自几个 临床队列密切合作。在目标1中，我们将在以前工作的基础上设计一种方法， 它可以在基因组上识别癌症患者的同源重组缺陷。虽然 全基因组测序提供了几个优点，基因组测序仍然是一个关键的平台， 临床护理我们的方法将能够对基因组进行签名分析，其中只有少量的 可以观察到突变。我们将纳入更多的信息来源，并确定生物标志物， 患者分层。在目标2中，我们将研究其他类型的基因组不稳定性，如错配修复 缺陷、复制应激和APOBEC诱变。例如，尽管错配修复的患者 免疫缺陷通常对免疫疗法有更好的反应，但患者之间存在相当大的差异。我们 将使用突变特征来识别一部分反应更好的患者。在目标3中，我们将扩展我们的 从循环肿瘤DNA和单细胞RNA测序获得数据的方法， 建模在目标4中，我们将建立一个全基因组和全外显子组的通用分析框架 签名分析，将克服目前的方法的缺点。我们将使用这个新框架， 为社区建立一个高质量的参考目录。

PROJECT SUMMARY Mutational signature analysis is a recent analytical approach for interpreting somatic mutations in the genome. It utilizes the sequence context of point mutations as well as the size and type of copy number and structural aberrations to decompose the observed mutational patterns into distinct 'signatures', some of which have been associated with specific biological processes. In this project, we will develop more robust and sensitive com- putational methods for mutational signature analysis and apply them to several clinically important questions. Our initial work has attracted a great deal of attention from clinicians, and we will analyze data from several clinical cohorts in close collaboration. In Aim 1, we will build upon our previous work to devise a method that can identify homologous recombination deficiency in cancer patients profiled on gene panels. Although whole-genome sequencing offers several advantages, gene panel sequencing remains as a pivotal platform in clinical care. Our method will enable signature analysis for gene panels from which only a small number of mutations can be observed. We will incorporate additional sources of information and identify biomarkers for patient stratification. In Aim 2, we will investigate other types of genomic instability such as mismatch repair deficiency, replication stress, and APOBEC mutagenesis. For example, although patients with mismatch repair deficiency generally respond better to immunotherapy, there is a considerable variation across patients. We will use mutational signatures to identify a subset of patients that respond better. In Aim 3, we will extend our method to data from circulating tumor DNA and single cell RNA sequencing to enable signature-based predictive modelling. In Aim 4, we will develop a generalized analytical framework for whole-genome and whole-exome signature analysis that will overcome the shortcomings of current approaches. We will use this new framework to build a high-quality reference catalog for the community.