FET: Small: Accurate and Scalable Methods for Analysis of Complex Genomic Populations

活细胞和病毒中的遗传物质经历了突变，导致了不同基因组群体的出现。在人类中，无论是遗传的还是在一个人的一生中获得的，基因突变都会导致遗传紊乱，使个人容易患上复杂的疾病，从而影响个人的健康。例如，癌症的发生和发展在一定程度上是由随着时间积累的体细胞突变推动的，这些突变创造了一个或多个肿瘤细胞群体。基因变异也发生在病毒中，它们导致出现丰富的种群，其光谱反映了特定突变为种群中存在的菌株提供的增殖优势。因此，推断基因组群体的组成和研究进化可以揭示关于疾病遗传特征的有价值的信息，并通常为医学和制药研究提供方向。这一研究工作将通过在教育创新、增强多样性、参与社区和向更广泛的公众传播成果方面的协调努力，使该领域和社会更广泛地受益。该项目的目标沿着三个综合研究方向：第一，一个方向是实现对以累积的点突变和插入/缺失为特征的不同基因组序列混合的准确和可扩展的分析。作为这项研究的一部分，研究人员将引入深度学习和社区检测范式来解决单倍型组装和重建病毒种群的问题，依靠领域知识在这些环境中为新算法的设计提供信息。第二，另一个方向是设计算法框架，用于重建以结构变化为特征的序列混合，例如拷贝数的变化。利用随机几何和混合整数优化的思想，研究人员将提供准确发现此类混合物组成的新方法，重点是绘制癌细胞中的肿瘤内异质基因组图谱。第三，最后的方向涉及发展研究基因组混合物的动力学的方法，前两个推进剂在静态环境下探索。特别是，最后一项研究旨在追踪感染网络中病毒种群的演变，并发现肿瘤细胞克隆成分中的体细胞突变之间的祖先关系。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Genetic material in living cells and viruses experiences mutations which lead to emergence of diverse genomic populations. In humans, whether inherited or acquired over a lifetime of an individual, genetic mutations impact the individual’s health by causing genetic disorders and rendering the individual predisposed to complex diseases. The onset and progression of cancer, for example, is in part driven by somatic mutations that accumulate over time, creating one or more populations of tumor cells. Genetic variations also occur in viruses where they lead to emergence of rich populations whose spectrum is reflective of the proliferative advantage that particular mutations provide to strains present in the population. Therefore, inferring the composition and studying evolution of genomic populations reveal valuable information about genetic signatures of diseases and generally suggest directions for medical and pharmaceutical research. This research effort will benefit the field and society more broadly through coordinated efforts in innovating in education, enhancing diversity, engaging the community, and disseminating results to a wider public.The aims of this project lie along three integrated research directions: First, one direction is enabling accurate and scalable analysis of diverse mixtures of genomic sequences characterized by accumulated point mutations and insertions/deletions. As part of this research thrust, investigators will introduce deep learning and community detection paradigms to the problems of haplotype assembly and reconstructing viral populations, relying on domain knowledge to inform the design of novel algorithms in those settings. Second, another direction is designing algorithmic frameworks for the reconstruction of the mixtures of sequences characterized by structural variations, such as the variations of copy numbers. Drawing upon ideas from stochastic geometry and mixed integer optimization, investigators will provide novel methods for accurate discovery of the composition of such mixtures, with a focus on mapping intra-tumor heterogeneous genomic landscapes in cancer cells. Third, the final direction involves development of methods for studying dynamics of genomic mixtures that the first two thrusts explore in static settings. In particular, the last research thrust aims to enable tracking the evolution of viral populations in infection networks and the discovery of ancestral relationships between somatic mutations in clonal components of tumor cells.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.