COMPREHENSIVE INFORMATIC ANALYSES OF AML GENOMES AND EPIGENOMES

摘要 十年的基因组研究揭示了急性髓细胞白血病基因组中出现的突变图景 髓系白血病(AML)。这些突变中的一个子集已被确定为AML启动事件 在造血干细胞/祖细胞(HSPC)中创造生长优势。这导致了早期阶段的 疾病，但这些突变的作用机制知之甚少。这项研究计划将 研究原发肿瘤和小鼠模型中的关键AML启动事件 全基因组转录和表观基因组数据(RNA-seq，亚硫酸氢盐-seq，atac-seq，单细胞RNA-seq， 等人)。我的角色将是将这些庞大、复杂的数据集转化为结论和可测试的假设 关于这些特定基因突变引发AML的确切分子机制。这样做将会 需要开发新的算法和统计模型，这两个领域都是我所从事的生物信息学 精通。然后，我们将利用这些知识来开发新的疗法。 第二个关键问题集中在50%的急性髓系白血病患者在治疗后首次完全缓解 化疗，但最终复发。我们之前的工作表明，基因组测序通常可以 在这些患者的化疗后活组织检查中识别与白血病相关的持续性突变，甚至 当它们处于形态缓解期时。这种突变清除的缺乏与患上 复发，但不是绝对的预测：少数持续突变的患者经历了长时间的 缓解，以及其他清除了所有突变的人很快复发。剩下的两个问题很重要 临床意义是：1)从更多的治疗后时间点对活检进行排序是否能更好地确定 突变清除的轨迹，并提供额外的预测有用的信息？2)可能超深 测序是否检测到了导致复发的残留病，从而促使更早的干预？ 共同领导了最初的突变清除研究，并编写了跟踪肿瘤克隆进化的工具 通过治疗，我拥有了所需的专业知识来支持这些研究，并进行分析、想象和解释 这些数据。我们正在开始临床试验，使用清除法将患者分配到或多或少的攻击性 对上述问题的处理和提供更清晰的信息将是翻译这些发现的关键部分 变成了强大的临床测试。 我的跨学科技能将对AML生物学的深刻理解与统计学敏锐性和 设计新算法方面的专业知识。我的训练、经验和成功的科学贡献记录 使我成为唯一适合推动这些项目的信息学和分析方面向前发展的人。

Abstract A decade of genomic studies has revealed the landscape of mutations appearing in the genomes of Acute Myeloid Leukemias (AML). A subset of these mutations have been identified as AML-initiating events that create growth advantages in a hematopoietic stem/progenitor cell (HSPC). This leads to the early stages of disease, but the mechanisms by which these mutations act is poorly understood. This research program will study key AML-initiating events in both primary tumors and mouse models by generating comprehensive whole-genome transcriptomic and epigenomic data (RNA-seq, bisulfite-seq, ATAC-seq, single-cell RNA-seq, et al). My role will be to translate these large, complex data sets into conclusions and testable hypotheses about the precise molecular mechanisms by which these specific genetic mutations initiate AML. Doing so will require the development of new algorithms and statistical models, both areas of bioinformatics in which I am proficient. We will then leverage this knowledge to develop novel therapies. A second key question focuses on the 50% of AML patients who initially experience complete remission after chemotherapy, but ultimately relapse. Our previous work has shown that genome sequencing can often identify persistent leukemia-associated mutations in post-chemotherapy biopsies from these patients, even when they are in morphological remission. This lack of mutation clearance was strongly associated with risk of relapse, but was not absolutely predictive: a few patients with persistent mutations experienced long remissions, and others who cleared all mutations quickly relapsed. Two remaining questions with important clinical implications are: 1) Can sequencing of biopsies from additional post-treatment timepoints better define the trajectory of mutation clearance and offer additional prognostically useful information? 2) Could ultra-deep sequencing have detected the residual disease responsible for relapse, thus prompting earlier interventions? Having co-led the original mutation clearance study, and authored tools for tracking a tumor's clonal evolution through therapy, I have the required expertise to deign these studies, and analyze, visualize, and interpret these data. We are beginning clinical trials that use clearance to assign patients to more or less aggressive treatments and providing more clarity on the above questions will be a key part of translating these findings into robust clinical tests. My interdisciplinary skillset couples a deep understanding of the biology of AML with statistical acumen and expertise in designing new algorithms. My training, experience, and record of successful scientific contributions make me uniquely suited to drive the informatics and analysis aspects of these projects forward.