Statistical Methods for Characterizing Tumor Heterogeneity at the Single Cell Level

点击翻译按钮获取中文摘要

Chronic lymphocytic leukemia (CLL) is a cancer that exhibits genetic and transcriptional heterogeneity along with a highly variable disease course among patients that remains poorly understood. Previous research has highlighted vast inter- and intra-patient genetic heterogeneity, with subclonal evolution commonly occurring in treatment settings leading to therapeutic resistance and relapse in many cases. In addition, our understanding of the role of co-existing non-cancer cells in the tumor-microenvironment remains limited. Therefore, characterization of these subclonal populations and their corresponding microenvironment will be paramount to enabling precision medicine and synergistic treatment combinations that target subclonal drivers and eliminate aggressive subpopulations thereby improving clinical outcome. In order to accurately dissect the genetic landscape and reconstruct the underlying subclonal architecture in CLL, measurements must be made on the single cell level. In the F99-phase of this proposed research, Jean Fan will continue developing statistical methods and computational software to analyze single cell RNA-seq data derived from CLL patient samples. Specifically, Jean will develop methods to identify aspects of genetic heterogeneity, such as the presence of small single nucleotide mutations and regions of copy number variation, in single cells. Jean will then reconstruct the genetic subclonal architecture and characterize the gene expression profiles of identified subclonal populations. In the K00-phase of this proposed research, Jean will characterize heterogeneity in the tumor-microenvironment and develop methods to assess potential reciprocal interactions between subclones and their microenvironment over time in response to therapy. The proposed work will yield innovative statistical methods to enable the identification and characterization of subpopulations in cancer and yield open-source software that can be tailored and applied to diverse cancer types. Ultimately, application of these developed methods to CLL will provide a better understanding of CLL development and progression.