AF: EAGER: Bayesian Factor Modeling of Context-Specific Gene Regulation

细胞对环境变化的反应是通过调节细胞中的分子来调控基因表达的复杂调控，其中最重要的是转录因子(TF)和微RNA(MiRNAs)。理解TF和miRNA调控如何定义细胞状态，如细胞存活、细胞增殖和细胞死亡，以及最终包括各种疾病在内的表型，是计算系统生物学家面临的主要挑战。智力价值这个迫切的项目将开发和验证一个新的计算模型，称为半参数贝叶斯因子调控模型(SB-FARM)，用于Tf和miRNA基因表达的调控。与其他现有模型相比，SB-FARM的优势在于它整合了模型中现有的基因调控知识，并能够在未测量的条件或背景下通过转录因子和miRNAs发现特定的调控。研究人员将检查建模细节以及从一组基因表达数据重建模型的算法。他们将应用SB-FARM来研究大肠杆菌和人类癌症中的特定背景法规。研究人员的长期目标是开发信号处理和统计学习方法，以便在系统层面上理解不同生物过程背后的基因调控网络，并将其应用于更好地理解生物多样性和疾病发展的基因组基础。SB-FILE的一般结构允许整合基因调控的其他方面。SB-FARM的成功将对基因调控研究产生长期影响，并有望显著推进统计信号处理和贝叶斯学习。广泛影响这项研究是高度跨学科、交叉的科学和工程。它将为基因组学、信号处理和计算生物学领域的高级跨学科学习和教育提供一个环境。督导计划亦会积极让研究生和本科生参与研究活动。特别是，他们将利用UTSA和UTHSCSA的少数族裔院校地位来招收和参与本研究的少数族裔学生。开发的计算方法将被应用到公共可用的软件中，以帮助计算生物学研究人员调查特定背景的基因调控。计算方法和工具将促进信号处理和机器学习的研究，并最终导致新的理论和方法的发展。

Response of cells to their changing environment is governed by intricate regulations of gene expression by regulating molecules in cells including, most importantly, transcription factors (TFs) and microRNAs (miRNAs). Understanding how TF and miRNA regulations define cellular states such as cell survival, cell proliferation, and cell death, and eventually phenotypes including various diseases is a major challenge facing computational systems biologists. Intellectual Merit This EAGER project will develop and validate a novel computational model called Semi-parametric Bayesian FActor Regulatory Model (SB-FARM) for TF and miRNA regulation of gene expression. The advantages of SB-FARM over other existing models are that it integrates existing knowledge about gene regulations in the model and enables the discovery of specific regulations by TFs and miRNAs under unmeasured conditions or contexts. The investigators will examine the modeling details as well as algorithms for reconstructing the model from a set of gene expression data. They will apply SB-FARM to study the context-specific regulations in E-coli and human cancer. The long term goal of the investigators is to develop signal processing and statistical learning methods for the system-level understanding of gene regulatory networks underlying different biological processes and apply them to better understand the genomic basis for diversity in organisms and development of diseases. The SB-FARM has a general structure that permits integration of additional aspects of gene regulation. The success of the SB-FARM will have long lasting impact on gene regulation research and is expected to also significantly advance statistical signal processing and Bayesian learning.Broader Impact This research is highly interdisciplinary, cross-cutting science and engineering. It will provide an environment for advanced interdisciplinary learning and education in the area of genomics signal processing and computational biology. The PIs will also actively involve graduate and undergraduate students in research activities. Particularly, they will utilize the minority institution status of UTSA and UTHSCSA to recruit and involve minority students in this research. The developed computational methods will be implemented into publically available software to aid computational biology researchers to investigate context specific gene regulations. The computational methods and tools will enhance the signal processing and machine learning research and ultimately lead to development of new theory and methods.