Structural and Functional Biology-based analysis of non-oncogene cancer dependencies

哥伦比亚癌症系统治疗中心(CAST)已经成功地为 以患者为中心的癌症治疗策略的优先顺序和评估。通过这一成本延长，我们 寻求完成因新冠肺炎而推迟的与原班人马目标相关的剩余工作 大流行。具体来说，我们会： 评估胰腺导管内致癌前体细胞群的肿瘤补充潜能 腺癌(PDA)：在PDA中，我们已经推断并实验验证了三种疾病的存在 转录上截然不同的细胞状态。从这些研究中出现的最关键的假设，由 RNA速度研究，是致癌前体(OP)群体-不同于其他两个，代表 是所有PDA的主要组成部分，以前未被检测到-代表了关键的肿瘤维持 通过补充形态亚群和谱系亚群来补充水库。为了证实这一假设， 就我们对这种肿瘤的理解而言，这将是范式的转变，我们正计划基于条形码 我们已经证明由OP和形态发生细胞组成的细胞系中的谱系追踪分析 (KP4)或OP和系细胞(CAPAN1)。 增强癌症稳态蛋白质数据库的结构信息注释： 作为主要的可交付成果，CABD是一个全面的、功能上和结构上都有注释的清单， 主调节蛋白(MR)，通过系统地解剖来自 公开可用的存储库。在Neo4j数据库中以图形数据库的形式实现了该图形数据库 系统，并已用来自PrePPI的标准癌症特征进行了部分注释。我们现在 建议将全部CAST结构信息资源编入该数据库，以便创建 该版本将包含派生自已在 CAST项目1和3的背景。 完全开发一个高度可扩展的单细胞DNA和RNA联合测序平台： CAST项目2的主要技术目标是开发同时单人可扩展技术 细胞DNA和RNA序列。正如先前报道的那样，我们开发了一种新型的非对称微孔阵列 同时用两个珠子同时共封装数千个单个细胞的技术-一个 含有用于直接对基因组DNA进行条形码编码的寡核苷酸，以及用于mRNA捕获和条形码的寡核苷酸。 我们已经证明，这种装置有助于同时固相捕获信使核糖核酸和溶液相 从单个细胞捕获gDNA。在这里，我们建议完成我们的DNA条形码的开发 并优化了我们的片上库构建的性能，用于全面演示 技术 从一组基准数据集中的scRNA-Seq图谱推断拷贝数变化：我们开发了 毕加索(从Scrna-Seq观测中的拷贝数变化推断系统发育)来推断拷贝 单细胞RNA测序数据的数量变化。毕加索被证明比现状更好 最先进的方法，如蜜獾，在模拟数据上，那里的地面真相答案是已知的。我们进一步 在一个胰腺癌铸型数据集上展示了准确性。基于这些初步数据，我们建议 系统地将毕加索应用于多种癌症类型的公开可用的scRNA-seq数据集，重点是 在人类肿瘤图谱网络上。

The Columbia Center for Cancer Systems Therapeutics (CaST) has successfully developed methodologies for the patient-centric prioritization and evaluation of cancer therapeutic strategies. Through this cost extension we seek to complete residual work related to the original CaST aims that was delayed due to the COVID-19 pandemic. Specifically, we will: Assess the tumor replenishing potential of the oncogenic precursor cell population in pancreatic ductal adenocarcinoma (PDA): In PDA we have inferred and experimentally validated the existence of three transcriptionally distinct cell states. The most critical hypothesis emerging from these studies, as supported by RNA velocity studies, is that the oncogenic precursor (OP) population – which unlike the other two, represents a dominant component of all PDAs and was previously undetected – represents the key tumor maintenance reservoir by replenishing both the morphogenic and the lineage sub-populations. To confirm this hypothesis, which would be paradigm shifting in terms of our understanding of this tumor, we are planning barcode-based lineage tracing assays in cell lines that we have already shown to comprise either OP and morphogenic cells (KP4) or OP and lineage cells (CAPAN1). Enhance the structure-informed annotation of the Cancer Homeostasis Proteins Database (CHoPD): One of the key CaST deliverables, the CHoPD is a comprehensive, functionally and structurally annotated inventory of master regulator (MR) proteins, compiled by systematically dissecting tumor checkpoints across samples from publicly available repositories. The CHoPD has been implemented as a graph database in the Neo4j database system and has been partially annotated with the standard cancer hallmarks as derived from PrePPI. We now propose to incorporate the full repertoire of CaST structure-informed resources into this database, so as to create a version of CHoPD that will contain extensive annotation derived from tools that have been developed in the context of CaST Projects 1 and 3. Complete development of a highly scalable platform for joint single-cell DNA and RNA sequencing: One of the major technical objectives of CaST Project 2 is the development of scalable technology for simultaneous single cell DNA- and RNA-seq. As previously reported, we have developed a novel asymmetric microwell array technology for simultaneous co-encapsulation of thousands of individual cells with exactly two beads – one harboring oligonucleotides for direct barcoding of genomic DNA and the other for mRNA capture and barcoding. We have demonstrated that this device facilitates simultaneous solid-phase capture of mRNA and solution-phase capture of gDNA from individual cells. Here we propose to complete the development of our DNA barcoding beads and optimize the performance of our on-chip library construction for a full-scale demonstration of the technology. Infer copy number alterations from scRNA-Seq profiles in a collection of benchmark datasets: We developed PICASSO (Phylogenetic Inference from Copy number Alterations in Scrna-Seq Observations) to infer copy number changes from single-cell RNA sequencing data. PICASSO was demonstrated to outperform state-of- the-art methods, such as HoneyBadger, on simulated data, where ground truth answers are known. We further demonstrated accuracy on a CaST pancreatic cancer dataset. Based on these preliminary data, we propose to systematically apply PICASSO to publicly available scRNA-seq datasets across multiple cancer types, focusing on the Human Tumor Atlas Network.