CAREER: Computational modeling and analysis of gene expression patterns from microscopy image data

杜克大学获得了教师早期职业发展计划（Career）的资助，用于开发生物图像数据分析和解释的综合框架。长期以来，图像在分子和发育生物学中被用作记录实验结果的手段，但越来越多地被视为定量数据，而不仅仅是定性描述。随着显微镜技术的进步，以及可视化生物分子的手段，越来越多的可用数据使图像成为计算生物学的一种新的数据类型，具有新的和令人兴奋的挑战和可能性。特别是，显微镜允许在高分辨率和活的有机体中测量基因表达模式。从图像中提取、表示和比较空间和时间表达模式的算法仍处于早期阶段，通常是针对特定场景量身定制的。该项目的主要贡献在于一个原则性的概率框架，该框架利用自上而下的生成策略从图像中提取样本，从显微镜数据中模拟基因表达模式，并将图像表达数据与其他基因组数据集成以了解基因调控。与从事动植物模型系统研究的生物学家密切合作将确保所开发的方法广泛适用，并将允许对特定模型预测进行有针对性的验证。这一提议的跨学科性质跨越了研究和教育。与研究计划相一致，计算生物学的研究生课程将扩大到包括案例研究模块，结合方法背景和实际例子来分析主要研究数据。主题将包括基因组注释、基因调控和图像分析。为了增加这一努力的影响，PI将与正在进行的NSF iPlant计划密切合作，并在高中、本科和研究生阶段的教育工作者研讨会上授课。PI还将继续参与由杜克大学系统生物学中心牵头的本科生和研究生系统生物学课程的开发和教学。中心正在进行的国际努力包括开发一个平台，允许开放共享教学资源。

Duke University is awarded a grant from the Faculty Early Career Development program (CAREER) to develop an integrated framework for the analysis and interpretation of biological image data. Images have been long used in molecular and developmental biology as a means to document the outcome of experiments, but are increasingly seen as quantitative data and not just qualitative descriptions. With recent advances in microscopy technology, as well as means to visualize biological molecules, the growing amount of available data has turned images into a new data type for computational biology with new and exciting challenges and possibilities. In particular, microscopy allows for measuring gene expression patterns at high resolution and in living organisms. Algorithms to extract, represent, and compare spatial and temporal expression patterns from images are still in early stages, and are often tailored to a particular scenario. The key contribution of this project lies in a principled probabilistic framework which utilizes top-down generative strategies to extract samples from images, model gene expression patterns from microscopy data, and integrate image expression data with other genomic data to understand gene regulation. Close collaborations with biologists working on animal and plant model systems will ensure that the developed methods are widely applicable, and will allow for the targeted validation of specific model predictions. The interdiscplinary nature of this proposal reaches across both research and education. In concert with the research program, a graduate course in computational biology will be expanded to include case study modules combining methodological background with hands-on examples to analyze primary research data. Topics will include genome annotation, gene regulation, and image analysis. To increase the impact of this effort, the PI will closely collaborate with the ongoing NSF iPlant initiative and teach at workshops for educators at the high school, undergraduate, and graduate level. The PI will also continue to participate in development and teaching of systems biology curricula for undergraduates and graduates spearheaded by the Duke Center for Systems Biology. Ongoing international efforts by the Center include the development of a platform to allow for an open sharing of teaching resources.