Genomic approaches in dissecting complex biological pathways.
Genomic approaches in dissecting complex biological pathways.
复制标题
DOI：
10.1517/phgs.5.2.163.27488
复制
发表时间：
2004-11
期刊：
影响因子：
2.1
通讯作者：
Ning Sun;Hongyu Zhao
中科院分区：
文献类型：
--
作者：
Ning Sun;Hongyu Zhao
基因组研究的进展提供了许多类型的大规模数据，其中包含各种生物途径的丰富信息。人们已经做出了大量的努力，使用这些基因组数据定性或定量地模拟生物途径。讨论了网络的无标度性和网络模体等一般性质，并应用各种网络模型对路径进行了重构。然而，在这些分析中缺乏对先验知识和不同基因组数据的系统整合。本文从生物系统的复杂性、生物通路的全局和局部特性出发，对通路重构进行了综述。我们回顾了主要的方法，包括聚类方法，无标度网络模型，贝叶斯网络模型，布尔网络模型，微分方程系统和数据集成方法。我们专注于每种方法在建模生物学途径的困难，并强调不同的模型捕捉生物学途径或基因组数据的不同方面。“噪声”大规模基因组数据要求数学模型和计算方法既鲁棒又可识别。此外，我们认为，理想的模型应该有能力纳入各种数据类型，这些模型需要通过与经验数据的严格比较进行评估。
Advances in genomic research have provided many types of large-scale data that contain rich information on various biological pathways. Intensive efforts have been made to qualitatively or quantitatively model biological pathways using these genomic data. Some general network properties, such as the scale-free property and network motifs, have been discussed and various network models have been applied to reconstruct pathways. However, there is a lack of systematic integration of prior knowledge and different genomic data in these analyses. In this review, we discuss pathway reconstruction under the consideration of the complexity embedded in the biological system, and the global and local properties of biological pathways. We review major methodologies, including clustering methods, scale-free networks models, Bayesian networks models, Boolean networks models, systems of differential equations, and data integration methods. We focus on the difficulty of each methodology in modeling biological pathways, and emphasize that different models capture different aspects of biological pathways or genomic data. The 'noisy' large-scale genomic data require the mathematical models and computational methods to be both robust and identifiable. In addition, we believe that ideal models should have the capability of incorporating various data types and these models need to be assessed through rigorous comparisons with empirical data.