GINI: from ISH images to gene interaction networks.
GINI: from ISH images to gene interaction networks.
复制标题
DOI：
10.1371/journal.pcbi.1003227
复制
发表时间：
2013
影响因子：
4.3
通讯作者：
Xing EP
中科院分区：
文献类型：
--
作者：
Puniyani K;Xing EP
要准确推断基因之间的分子和功能相互作用，尤其是在果蝇等多细胞生物中，通常不仅需要对基因表达量之间的相关性进行统计分析，还需要对它们的时空模式之间的相关性进行统计分析。原位杂交技术是一种有效的基因表达显微成像技术，可用于大规模的mRNA时空分布研究。然而，从这些数据中发现基因相互作用的分析工具仍然是一个开放的挑战，由于各种原因，包括难以从图像中提取基因活动的规范表示，以及从这些表示中推断统计上有意义的网络。在本文中，我们提出了GINI，一个机器学习系统推断基因相互作用网络的果蝇胚胎ISH图像。GINI建立在由我们最近开发的系统实现的ISH图像中基因表达空间模式的计算机视觉启发的向量空间表示上;以及一种新的多实例内核算法，该算法学习稀疏马尔可夫网络模型，其中，每个基因（即，节点）由矢量值空间模式表示，而不是如在诸如高斯图形模型的常规方法中那样由标量值基因强度表示。通过捕获基因表达的空间相似性概念，同时通过多实例内核适当考虑每个基因存在多个图像，GINI可以很好地从图像数据中推断出统计上合理且具有生物学意义的基因相互作用网络。使用合成数据和一个小的手动策划的数据集，我们证明了我们的方法在网络建设的有效性。此外，我们报告了一个大的公开收集的果蝇胚胎ISH图像从伯克利果蝇基因组计划，GINI使新的和有趣的预测基因相互作用的结果。GINI软件可在http://sailing.cs.cmu.edu/Drosophila_ISH_images/上获得。随着用于分子丰度分析的高通量技术变得越来越便宜和容易获得，基于有充分根据的统计学原理从这些数据对基因相互作用网络进行计算推断对于促进对各种生物系统中的调节机制的理解是必要的。基因网络的逆向工程传统上依赖于全基因组微阵列数据的分析，在这里，我们提出了一种新的方法，GINI，从ISH图像推断基因网络，从而使网络推理的基因表达的空间特征的探索。我们的方法生成一个马尔可夫网络，它封装了全球有意义的向量值基因空间模式的遗传依赖。换句话说，我们在使用更丰富的表达数据形式以及使用原则性统计方法对这种新形式的数据进行合理的网络推理方面都取得了进展。我们的研究结果表明，分析基因表达的空间分布，使我们能够捕捉到的信息不能从微阵列数据。这种分析在分析果蝇胚胎发育中涉及的基因以揭示决定成年果蝇14个节段发育的特定空间模式方面尤其重要。
Accurate inference of molecular and functional interactions among genes, especially in multicellular organisms such as Drosophila, often requires statistical analysis of correlations not only between the magnitudes of gene expressions, but also between their temporal-spatial patterns. The ISH (in-situ-hybridization)-based gene expression micro-imaging technology offers an effective approach to perform large-scale spatial-temporal profiling of whole-body mRNA abundance. However, analytical tools for discovering gene interactions from such data remain an open challenge due to various reasons, including difficulties in extracting canonical representations of gene activities from images, and in inference of statistically meaningful networks from such representations. In this paper, we present GINI, a machine learning system for inferring gene interaction networks from Drosophila embryonic ISH images. GINI builds on a computer-vision-inspired vector-space representation of the spatial pattern of gene expression in ISH images, enabled by our recently developed system; and a new multi-instance-kernel algorithm that learns a sparse Markov network model, in which, every gene (i.e., node) in the network is represented by a vector-valued spatial pattern rather than a scalar-valued gene intensity as in conventional approaches such as a Gaussian graphical model. By capturing the notion of spatial similarity of gene expression, and at the same time properly taking into account the presence of multiple images per gene via multi-instance kernels, GINI is well-positioned to infer statistically sound, and biologically meaningful gene interaction networks from image data. Using both synthetic data and a small manually curated data set, we demonstrate the effectiveness of our approach in network building. Furthermore, we report results on a large publicly available collection of Drosophila embryonic ISH images from the Berkeley Drosophila Genome Project, where GINI makes novel and interesting predictions of gene interactions. Software for GINI is available at http://sailing.cs.cmu.edu/Drosophila_ISH_images/ As high-throughput technologies for molecular abundance profiling are becoming more inexpensive and accessible, computational inference of gene interaction networks from such data based on well-founded statistical principles is imperative to advance the understanding of regulatory mechanisms in various biological systems. Reverse engineering of gene networks has traditionally relied on analysis of whole-genome microarray data; here we present a new method, GINI, to infer gene networks from ISH images, thereby enabling exploration of spatial characteristics of gene expressions for network inference. Our method generates a Markov network, which encapsulates globally meaningful statistical-dependencies from vector-valued gene spatial patterns. In other words, we advance the state-of-art in both the usage of richer forms of expression data, and the employment of principled statistical methodology for sound network inference on such new form of data. Our results show that analyzing the spatial distribution of gene expression enables us to capture information not available from microarray data. Such an analysis is especially important in analyzing genes involved in embryonic development of Drosophila to reveal specific spatial patterning that determines the development of the 14 segments of the adult fly.
登录
查看更多内容
DOI：
10.1090/s0002-9947-1950-0051437-7
发表时间：
1950-01-01
影响因子：
1.3
作者：
ARONSZAJN, N
通讯作者：
ARONSZAJN, N
影响因子：
30.8
作者：
Janssens, Hilde;Hou, Shuling;Reinitz, John
通讯作者：
Reinitz, John
DOI：
10.1073/pnas.0901910106
发表时间：
2009-07-21
影响因子：
11.1
作者：
Ahmed, Amr;Xing, Eric P.
通讯作者：
Xing, Eric P.
影响因子：
64.8
作者：
通讯作者：
--
影响因子：
1.9
作者：
DEMPSTER, AP
通讯作者：
DEMPSTER, AP