CAREER: Statistical and Computational Tools for the Analysis of High Dimensional Genetic Data

项目编号：DMS-0239427PI: Chiara sabattitititle：职业：用于高维遗传数据分析的统计和计算工具摘要本项目将创建新的统计模型和计算工具，用于分析高维空间的数据，正如遗传学领域所产生的那样。特别是研究者和她的同事将(a)开发基因组序列模型，旨在建立结合位点的总数、它们的位置和它们之间的相互作用；(b)对基因阵列数据进行去噪处理，建立各种基因表达之间的依赖关系模型，并确定引起表达变化的不同化学信号的数量；(c)建立“单倍型块”的概念模型，并定义以基因定位为目的识别它们的程序，并为同一背景下的多次比较制定适当的纠正程序。该项目阐明了模型选择、多重比较、高维函数估计等主题之间的关系，并对贝叶斯模型、最小描述长度原则和错误发现率之间的联系有了更深的理解。该研究还将开发一套新的计算工具，该工具基于马尔可夫链蒙特卡罗采样和各种不同尺度上的目标分布的表示。概述的研究有助于解决有关基因的作用和表达的一些基本问题，从而通过发现与疾病有关的基因、发展遗传疗法和在工业规模上对感兴趣的蛋白质的过度生产进行工程设计，改善一般福利。通过公开基因组和基因表达分析的算法，以及升级计算基础设施，该计划扩大了对服务不足社区的科学调查的参与，并加强了研究的一般基础设施。拟议组织的跨学科讲习班、研究活动和课程确保了成果的广泛传播，以增进科学理解。为高中和大学教师组织跨学科背景下的教学统计研讨会，其方向是将研究与教育结合起来，促进教学、培训和学习。

Proposal ID: DMS-0239427PI: Chiara SabattiTitle: CAREER: Statistical and computational tools for the analysis of high dimensional genetic dataAbstractThis project will enable the creation of novel statistical models and computational tools for the analysis of data in high dimensional spaces, as the one generated in the field of genetics. In particular the investigator and her colleagues will (a) develop models for genomic sequences that aim at establishing the total number of binding sites, their location and their interaction with each other; (b) pursue de-noising of gene array data, modeling of the dependence between the expression of various genes, and the identification of the number of different chemical signals originating change in expression; (c) model the notion of ``haplotype blocks'' and define the procedures to identify them with the purpose of gene mapping, and develop appropriate procedures of correction for multiple comparison in the same context. The project illustrates relations between the topics of model selection, multiple comparison, high-dimensional function estimation and leads to deeper understanding of connections between Bayesian models, minimum description length principle, and false discovery rates. The proposed research will additionally develop a new set of computational tools that are based on Markov Chain Monte Carlo sampling and representation of the objective distribution on a variety of different scales. The outlined research helps to tackle some fundamental questions regarding the role and the expression of genes, thus leading to improvements of the general welfare, trough the discovery of genes related to diseases, the development of genetic therapies, and the engineering of the over-production of protein of interests on industrial scale. By making the algorithms for genome and gene expression analysis publicly available, and upgrading the computing infrastructure, the project broadens the participation to scientific investigation of under-served community and enhance the general infrastructure for research. The proposed organization of interdisciplinary workshops, research activities and courses assures a broad dissemination of the results to enhance scientific understanding. The organizations of seminars on teaching statistics in interdisciplinary settings for high-school and college instructors goes in the direction of integrating research and education, promoting teaching, training, and learning.