Development of Bayesian Graphical Models for Next-Generation Genetic Studies

在过去的二十年里，图形模型一直是用于分析复杂结构化高维数据的最有效的统计工具之一（Friedman，2004）。图形模型提供了一个概率框架，用于进行推理和表示我们对这些复杂结构化数据的知识。在生物学研究中，特别是在基因组学、蛋白质组学、代谢组学、转录组学等新兴的“组学”学科中，数据往往来自复杂的高通量实验和复杂的实验设计。图形模型可以表示这些复杂的生物学问题，导致相对简单和易于处理的计算算法，以获得感兴趣的数量。尽管其明显的优势，使用图形模型下的频率或贝叶斯框架仍然相对较少的生物学研究。在“组学”学科中，这可以部分地解释为在将图形模型拟合到高维数据方面存在计算困难，以及缺乏贝叶斯图形模型的一般理论框架。在这项提案中，我们计划开发一个基于贝叶斯图形模型的通用统计框架，用于分析复杂的遗传数据。我们将通过几个可能对生物学和公共卫生领域产生重大影响的应用程序来证明我们方法的相关性，包括：**1）使用全基因组关联研究（GWAS）和下一代测序（NGS）数据进行复杂人类疾病的基因发现 *2）关于复杂生物系统的推断 *3）为各种研究设计开发贝叶斯图形模型 * 尽管遗传学和生物信息学中出现了复杂的统计模型，但理论与应用之间仍然存在重要差距。这项研究提案可以为遗传发现的方法发展带来新的见解，并改善生物学和统计学之间的界面。 **

Graphical models have been one of the most efficient statistical tools used in the last twenty years for the analysis of complex structured high-dimensional data (Friedman, 2004). Graphical models provide a probabilistic framework for making inference and representing the knowledge that we have about these complex structured data. In biological research and more particularly in the emerging `omics disciplines such as genomics, proteomics, metabolomics, transcriptomics, data are often generated from complex high-throughput experiments and from complex experimental designs. Graphical models can represent these complex biological problems, leading to relatively simple and tractable computational algorithms to obtain the quantities of interest. Despite its obvious advantages the use of graphical models under either the frequentist or Bayesian framework remains relatively rare in biological research. In `omics disciplines, this could be partly explained by the computational difficulties in fitting graphical models to high-dimensional data and also the lack of a general theoretical framework for Bayesian graphical models.******In this proposal, we plan to develop a general statistical framework based on Bayesian graphical models for the analysis of complex genetic data. We will demonstrate the relevance of our approach through several applications that could have a major impact in biology and public health areas, including:******1) Gene Discovery in Complex Human Diseases using Genome-wide Association Studies (GWAS) and Next Generation Sequencing (NGS) data***2) Inference about Complex Biological Systems***3) Development of Bayesian graphical models for various study designs***Despite the explosion of complex statistical models appearing in genetics and bioinformatics, there still remains an important gap between theory and its applications. This research proposal could bring new insights into method developments for genetic discoveries and improve the interface between biology and statistics.  **