Discovery and investigation of cell type abundance biomarkers using single-cell and bulk genomic data

拟议的研究计划将开发第一个软件管道，以发现基于细胞类型丰度的生物标志物（使用单细胞基因组学技术），并在大规模基因组研究中验证其与临床结果（或其他表型）的相关性。背景：组织微环境由各种细胞类型的动态混合组成。细胞类型丰度的变化可以改变器官的状况和功能。因此，基于细胞类型丰度的生物标志物可作为多种疾病的药物靶点。我们发现和研究细胞类型丰度生物标志物的新方法包括三个步骤。在步骤1中，我们使用单细胞rna测序（scRNA-seq）来鉴定特定类型组织中的所有细胞类型并推断其丰度。发现的生物标志物是在健康人与患病人之间存在显著差异的特定细胞类型的丰度。尽管新的scRNA-seq技术使我们能够在单细胞水平上研究组织，但迫切需要统计方法和软件来从数据中提取信息并精确推断。验证丰度-生物标志物的临床相关性需要大量样本。由于单细胞技术的限制，使用scRNA-seq进行如此大规模的研究是不切实际的。因此，第二步是开发数字细胞术方法，在scRNA-seq和散装RNA-seq之间建立信息桥梁，从而通过使用更实惠的散装RNA-seq来推断细胞类型丰度。最后，第三步是进行大规模的基因组研究，并对数据进行分析，评估生物标志物的临床相关性。我的研究目标是为了支持上面讨论的三个步骤。首先，我将开发新的统计方法和软件来注释细胞类型，并使用scRNA-seq数据推断它们的丰度。其次，我将基于scRNA-seq数据开发新的数字细胞术方法。第三，我将开发新的统计方法来支持大规模基因组研究的设计和分析，包括三个研究领域：临床试验的设计，因果推理和多任务预测。我将根据公平、多元化和包容性的指导方针，通过培训高素质人才（HQP）来实施该提案。研究成果和影响：我的主要成果是统计方法和软件管道，其他研究人员可以在他们的基因组研究中使用。这些方法将用于发现基于肺泡巨噬细胞亚型丰度的生物标志物，这些标志物将用作慢性阻塞性肺疾病（COPD）的免疫治疗药物靶点。在加拿大，慢性阻塞性肺病是住院治疗的第一大原因，每年治疗200万慢性阻塞性肺病患者的直接支出超过15亿美元。目前没有任何治疗方法可以逆转慢性阻塞性肺病的进展。我们的研究可能会导致首个治愈慢性阻塞性肺病的疗法。

The proposed research program will develop the first software pipeline to discover biomarkers based on cell type abundance (using single-cell genomics technology) and to validate their relevance to clinical outcomes (or other phenotypes) in large-scale genomic studies. Background: The tissue microenvironment consists of a dynamic mixture of various cell types. The changes in cell type abundance can modify an organ's conditions and functions. Hence, the biomarkers based on cell type abundance can be used as drug targets for various diseases. Our novel procedure of discovering and investigating the cell type abundance biomarkers involves three steps. In step 1, we use single-cell RNA-sequencing (scRNA-seq) to identify all cell types in a specific type of tissue and infer their abundance. The discovered biomarkers are the abundance of specific cell types that vary significantly between healthy and diseased persons. Although the new scRNA-seq technology enables us to study tissues at the single-cell level, statistical methods and software are urgently needed to extract information from the data and make inferences precise. Validating the abundance-biomarkers' clinical relevance requires many samples. Conducting such large-scale studies using scRNA-seq is not practical due to the limitations of single-cell technology. So, the step 2 is developing digital cytometry methods to bridge information between scRNA-seq and bulk RNA-seq, which enables the inference of cell type abundance by using more affordable bulk RNA-seq. At last, the step 3 is to conduct a large-scale genomic study and analyze the data to evaluate the biomarkers' clinical relevance. Objectives: My research objectives are motivated by supporting the three steps discussed above respectively. First, I will develop novel statistical methods and software to annotate cell types and infer their abundance using scRNA-seq data. Second, I will develop novel digital cytometry methods based on the scRNA-seq data. Third, I will develop novel statistical methods to support the design and analysis of large-scale genomic studies, including three research areas: the design of clinical trials, causal inference, and multitask predictions. I will conduct the proposal by training highly qualified personnel (HQP) under equity, diversity, and inclusion guidelines. Research deliverables and impacts: My primary deliverables are the statistical methods and the software pipeline, which other researchers can use in their genomic studies. Those methods will be applied to discover biomarkers based on the abundance of alveolar macrophages subtypes, which will be used as immunotherapy drug targets for chronic obstructive pulmonary disease (COPD). In Canada, COPD is the top 1 reason for hospitalization, which costs over $1.5 billion/yr in direct expenditures to treat two million COPD patients. No current therapy can reverse disease's progression of COPD. Our research may lead to the first therapy that cures COPD.