Development of Bioimage Informatics Tools for Spatial Expression Datasets

描述（由申请人提供）：探测基因功能的关键工具之一是用共聚焦或更高分辨率显微镜成像空间表达，产生特定基因何时何地表达的模式。由于这种空间表达模式的可用性增加以及高通量成像的重要性增加，因此需要用于管理和分析用于研究和系统生物学的空间基因表达数据集的通用工具。我们建议开发一套用于管理和分析大规模空间基因表达数据集的工具，并向公众提供。对于我们的大型果蝇空间表达数据集，我们已经开发了一个基于图像的虚拟表示，该虚拟表示对于可视化，数据挖掘和分析是实用的，同时足够紧凑以分发和存储大型表达数据集。此外，这种表示允许在没有专门的图像处理知识的情况下分析基于图像的数据。我们建议通过在流行的科学图像处理程序ImageJ中添加集成的数据管理和分析平台来增强这些工具。该ImageJ工作台将补充现有的网络界面，并将使生物学家能够执行额外的分析，生物信息学研究人员通过编写自己的自定义模块来扩展软件。我们将扩展虚拟胚胎表示的适用性，以适应在开发后期发现的复杂模式，并扩展显微镜自动化软件，以根据自动获取的图像生成虚拟表示。最后，我们计划扩展该软件的功能，以处理来自其他模式生物的图像，并开发跨物种比较的方法。该工具包将为基因表达数据与调控和基因序列的整合提供坚实的基础，以促进发现调控相互作用网络的研究。

DESCRIPTION (provided by applicant): One of the critical tools for probing gene function is imaging spatial expression with confocal or higher resolution microscopy, yielding a pattern of when and where a particular gene is expressed. As a result of the increased availability of such spatial expression patterns and the increased importance of high throughput imaging, there is a need for general purpose tools for managing and analyzing spatial gene expression datasets for research and systems biology. We propose to develop and make available to the general community a suite of tools for managing and analyzing large-scale spatial gene expression datasets. For our large Drosophila spatial expression dataset, we have developed an image based virtual representation that has been practical for visualization, data mining and analysis, while being compact enough to distribute and store large expression datasets. Moreover, this representation allows for analyzing image based data without specialized image processing knowledge. We propose to enhance these tools by adding an integrated data management and analysis platform to a popular scientific image processing program, ImageJ. This ImageJ workbench will be complementary to existing web-interfaces and will enable biologists to perform additional analyses and bioinformatics researchers to extend the software by writing their own custom modules. We will expand the applicability of our virtual embryo representations to complex patterns found at later stages of development and extend microscope automation software to generate virtual representations from automatically acquired images. Finally, we plan to extend the capabilities of the software to process images from other model organisms and develop methods for cross species comparisons. The proposed toolkit will provide a strong foundation for integration of gene expression data with regulatory and gene sequences to promote research for discovering networks of regulatory interactions.