Statistical Methods for Intra-Tumor Heterogeneity Studies Using Sequencing Data

肿瘤异质性是指观察到肿瘤细胞可以显示不同的表型和形态学特征，如基因表达、代谢、细胞形态学增殖、运动性和转移潜力。这种现象发生在肿瘤之间（肿瘤间异质性）和肿瘤内（肿瘤内异质性）。了解肿瘤的异质性是肿瘤个体化诊断和治疗的前提。本项目的目的是开发统计方法，使用下一代测序数据研究肿瘤内异质性。这些方法将解决最近癌症基因组学研究中出现的重要统计，计算和生物学挑战。这些应用将进一步加深我们对肿瘤异质性机制及其临床后果的理解。这项研究将提供机会，吸引和培养不同的未来科学家在计算癌症基因组学的前沿工作。该项目还将为本科生和研究生提供研究培训机会。将开发、分发和支持用户友好的开放源码软件，以实施研究方法，使基因组学和统计学界受益。该项目将开发具有坚实统计基础的克隆性分析方法，并针对不同测序技术的数据特征进行定制。一个新的随机过程将开发模型克隆扩张。使用基于可能性的方法，PI将通过几种新策略构建克隆历史：1）将相位信息与癌症基因组学研究以及由联盟（如1000 Genome Project）分析的丰富种系变体资源相结合;（2）整合单细胞RNA测序和大量组织DNA测序数据;（3）调整局部序列的特征。这些分析还可能为联合分析肿瘤异质性和丰富的表观基因组特征（如DNA可及性和甲基化）开辟新的机会。所有目标的成功实现将大大提高癌症基因组学研究中亚克隆鉴定的能力。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Tumor heterogeneity refers to the observation that tumor cells can display distinct phenotypic and morphological characteristics, such as gene expression, metabolism, cellular morphology proliferation, motility, and metastatic potential. This phenomenon occurs both between tumors (inter-tumor heterogeneity) and within tumors (intra-tumor heterogeneity). Understanding tumor heterogeneity is a prerequisite for personalized tumor diagnosis and treatment. The objective of this project is to develop statistical methods to study the intra-tumor heterogeneity using next generation sequencing data. These methods will address important statistical, computational, and biological challenges that arise from recent cancer genomics studies. The applications will further our understanding of mechanisms underlying tumor heterogeneity as well as its clinical consequences. The research will provide opportunities to attract and nurture diverse future scientists to work at the frontiers of computational cancer genomics. The project will also provide research training opportunities for undergraduate and graduate students. User-friendly open-source software implementing the research methods will be developed, distributed, and supported to benefit the genomics and statistics community.This project will develop clonality analysis methods with a firm statistical footing and tailored for the characteristics of data from different sequencing technologies. A new stochastic process will be developed to model clonal expansion. Using likelihood-based approaches, the PI will construct clonal history empowered by several novel strategies: 1) incorporating phase information bridges cancer genomics studies with rich germline variant resources profiled by consortium such as the 1000 Genome Project; (2) integrating single cell RNA sequencing and bulk tissue DNA sequencing data; (3) adjusting for characteristics of local sequences. These analyses can also potentially open up new opportunities for joint analysis of tumor heterogeneity and a rich list of epigenomic features such as DNA accessibility and methylation. Successful achievement of all aims will dramatically increase the power of subclone identification in cancer genomics studies.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.