基于新一代测序数据的基因调控网络数学模型与方法的研究

The emerging next generation sequencing (NGS) technologies have been widely employed in biological and biomedical studies. With the huge amount of data being generated, the processing and analysis of NGS data raises lots of mathematical and computational challenges. As new mathematical and computational methods have been successfully developed to handle these difficulties in data processing, we are now moving to the next stage of the downstream analysis with the goal to deepen our understanding of the biology, especially the gene regulatory networks (GRNs) based on the NGS data. Currently, however, most existing methods of GRNs were developed based on the microarray data rather than the NGS data, making them not applicable to the NGS-based studies. To fill this gap, our project will develop new mathematical models and methods, especially the probabilistic and statistical models and methods, for the NGS-based GRN studies. Specifically, the project will study the following three key scientific questions, (1) the model based bias adjustment methods for the NGS data; (2) the causal inference of the GRNs based on the modeling of allele-specific expression; and (3) the study of the dynamic GRNs by considering RNA degradation. This project will not only provide new methodologies for the studies of systems biology and bioinformatics, but also contribute to our understanding of gene regulatory mechanisms of human diseases.

新一代测序技术在现代生物学和医学研究中得到广泛应用。随着前期数据处理的数学与计算方法的基本完善，如何进一步分析和挖掘新一代测序数据，深入研究基因调控网络，成为了当前研究的热点与难点。针对新一代测序数据的自身问题及其提供的基因调控的关键信息，本项目将致力于发展基于新一代测序数据的基因调控网络的数学模型和方法，特别是概率统计的模型和方法，着重解决三个方面的关键问题：（1）基于模型的新一代测序数据系统偏差的校正，（2）基于等位基因特异性表达机制与模型的基因调控网络的因果推断，以及（3）基于RNA降解模型的基因调控动态网络的研究等。本项目将为系统生物学和生物信息学领域提供新的研究方法，并将推进我们对人类复杂疾病的基因调控机制的深入理解与研究。

新一代测序技术在现代生物学和医学研究中得到广泛应用。如何有效的分析与整合以新一代测序数据为主的生物医学数据，深入研究基因调控网络，特别是复杂疾病调控的机制与网络,成为了当前研究的热点与难点。本项目重点研究和发展了基于新一代测序数据的基因调控网络的数学模型和方法，并且着重解决了三个方面的关键科学问题：（1）发展新一代测序数据的定量模型与算法，有效校正数据中的系统性偏差;（2）基于新一代测序数据的关键信息发展基因调控网络的构建与推断方法;以及（3）基于新一代测序数据的基因调控动态网络进行探索性研究等。具体而言,本项目取得了如下成果: (1)建立了新一代测序数据的碱基辨识的定量模型及空间混杂校正的自适应解和算法;(2) 发展了RNA-Seq定量模型与统计算法,提升了转录组估计的精确度;(3) 提出了基于时空型新一代测序数据的动态基因调控网络的新的分析框架;(4) 开展了基于多模态数据驱动的脑疾病的调控机制研究,建立了从影像到基因跨尺度的机制关联;以及(5)系统比较了新一代测序与生物芯片的拷贝变异检测功效,发现了复杂疾病遗传变异的基因调控机制与网络。本项目的研究不但为系统生物学和生物信息学领域提供了新的研究方法，而且深化了我们对诸如脑疾病等人类复杂疾病调控机制的认识与研究。