Computational Methods for Inferring Single-cell DNA Methylation and its Spatial Landscape

项目总结/摘要 候选人侯文平博士是一位应用数学家和计算生物学家， 约翰霍普金斯彭博公共卫生学院生物统计学博士后研究员。她的长期 职业目标是通过发展计算和统计来改善疾病治疗的临床实践 使用多组学数据破译空间和时间基因调控程序的方法， 这些方法对发育过程和疾病的影响。她提出的研究题为计算 推断单细胞DNA甲基化及其空间景观的方法结合了先进的空间 转录组学技术与计算方法来推断空间DNA甲基化景观， 能够准确评估治疗中的表观基因组空间变异性和表观基因组靶点， 疾病侯博士在香港大学完成了数学博士学位，她专注于 推断和控制基因调控网络（GRNs）。她的第一个博士后研究（2017-2019）与博士。 约翰霍普金斯大学（JHU）的Aravinda Chakravarti和Suchi Saria代表了 从理论到计算基因组学的GRNs。她的第二个博士后研究（2019年至今）与博士。 JHSPH的Hongkai Ji和Stephanie Hicks进一步提供了重建和 使用单细胞数据预测空间转录组和表观基因组景观。侯博士的指导团队 包括Hongkai Ji（初级导师），博士，开发计算和统计工具的专家， 分析单细胞基因组数据，包括重建和预测时间和空间转录组 和表观基因组景观;斯蒂芬妮希克斯（共同导师），博士，在开发统计方法， 解决单细胞中的技术变异性和空间转录组学;和Andrew Feinberg（共同导师），博士， 癌症表观遗传学领域的创始人之一，他领导了第一个由NIH资助的表观基因组中心。这 提供了解决统计，表观基因组学和空间交叉领域重大挑战的机会 转录组学与先进的实验技术。她的科学顾问是卡斯珀汉森博士， Gregory Hager和Xiaobin Wang在计算表观基因组学，深度学习， 临床翻译和疾病预防。利用知识、实验和计算 来自所有导师和顾问以及通过JHSPH，JHU和约翰霍普金斯医学的资源，侯博士 将接受密集的培训，指导和职业发展，以实现本报告中提出的目标。 并在奖励期间取得了富有成效的成果。目标1将开发预测DNA的方法 使用批量基因表达的甲基化景观。目标2将开发预测单细胞DNA的方法 甲基化和差异DNA甲基化。Aim 3将重建组织空间DNA甲基化景观， 单细胞水平，在空间背景下生成评估数据集，并执行跨研究评估 使用来自DNA元素百科全书（ENCODE）、人类细胞图谱和Recount 2的数据。

Project Summary/Abstract The candidate, Wenpin Hou, PhD, is an applied mathematician and computational biologist serving as a postdoctoral fellow of Biostatistics at Johns Hopkins Bloomberg School of Public Health (JHSPH). Her long-term career goal is to improve clinical practice in therapy for diseases by developing computational and statistical methods to decipher spatial and temporal gene regulatory programs using multi-omics data and implementing these methods into developmental process and diseases. The research she proposes entitled Computational Methods for Inferring Single-cell DNA Methylation and its Spatial Landscape combines advanced spatial transcriptomics techniques with computational methods to infer spatial DNA methylation landscape, which enables the accurate evaluation of the epigenomic spatial variability and epigenomic targets in therapy for diseases. Dr. Hou completed her PhD in Mathematics at The University of Hong Kong where she focused on inferring and controlling gene regulatory networks (GRNs). Her first postdoctoral research (2017-2019) with Drs. Aravinda Chakravarti and Suchi Saria at Johns Hopkins University (JHU) represented the first shift in focus of GRNs from theoretical to computational genomics. Her second postdoctoral research (2019-present) with Drs. Hongkai Ji and Stephanie Hicks at JHSPH has further provided complementary training in reconstructing and predicting spatial transcriptomic and epigenomic landscape using single-cell data. Dr. Hou's mentoring team consists of Hongkai Ji (primary mentor), PhD, an expert in developing computational and statistical tools for analyzing single-cell genomic data, including reconstructing and predicting temporal and spatial transcriptomic and epigenomic landscape; Stephanie Hicks (co-mentor), PhD, an expert in developing statistical methods to address technical variability and spatial transcriptomics in single cells; and Andrew Feinberg (co-mentor), PhD, one of the founders of the field of cancer epigenetics who directs the first NIH funded Epigenome Center. This offers the opportunity to tackle significant challenges in the intersection of statistics, epigenomics and spatial transcriptomics with advanced experimental techniques. Her scientific advisors are Drs. Kasper Hansen, Gregory Hager and Xiaobin Wang who have leading expertise in computational epigenomics, deep learning, clinical translation and disease prevention, respectively. Leveraging the intellectual, experimental, and computing resources from all mentors and advisors as well as through JHSPH, JHU and Johns Hopkins Medicine, Dr. Hou will receive intensive training, mentoring and career development to achieve the goals proposed in this application and have productive outcomes during the award period. Aim 1 will develop methods to predict DNA methylation landscape using bulk gene expression. Aim 2 will develop methods to predict single-cell DNA methylation and differential DNA methylation. Aim 3 will reconstruct tissue-spatial DNA methylation landscape at the single-cell level, generate evaluation datasets in spatial context, and perform across-study assessments using data from The Encyclopedia of DNA Elements (ENCODE), Human Cell Atlas, and Recount2.