当前位置: X-MOL 学术IEEE/ACM Trans. Comput. Biol. Bioinform. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Common Neighbors Extension of the Sticky Model for PPI Networks Evaluated by Global and Local Graphlet Similarity
IEEE/ACM Transactions on Computational Biology and Bioinformatics ( IF 3.6 ) Pub Date : 2020-08-18 , DOI: 10.1109/tcbb.2020.3017374
Sridevi Maharaj , Taotao Qian , Zarin Ohiba , Wayne Hayes

The structure of protein-protein interaction (PPI) networks has been studied for over a decade. Many theoretical models have been proposed to model PPI network structure, but continuing noise and incompleteness in these networks make conclusions about their structure difficult. Using newer, larger networks from Sept. 2018 BioGRID and Jan. 2019 IID, we show the joint distribution of degree products and common neighbors has a greater impact on PPI edge connectivity than their individual distributions, and introduce two new models (CN and STICKY-CN) for PPI networks employing these features. Since graphlet-based measures are believed to be among the most discerning and sensitive network comparison tools available, we assess their overall global and local fits to PPI networks using Graphlet Kernel (GK). We fit 10 theoretical models to nine BioGRID networks and twelve Integrated Interactive Database (IID) networks and find: (1) STICKY and STICKY-CN are the overall globally best fitting models according to GK, (2) Hyperbolic Geometric Graph model is a better fit than any STICKY-based model on 4 species, (3) though STICKY-CN provides a better local fit than the STICKY model, the CN model provides the greatest local fit over most species. We conclude that the inclusion of CN into STICKY-CN makes it the best overall fit for PPI networks as it is a good fit locally and globally.

中文翻译:

通过全局和局部 Graphlet 相似性评估的 PPI 网络粘性模型的公共邻居扩展

蛋白质-蛋白质相互作用 (PPI) 网络的结构已经研究了十多年。已经提出了许多理论模型来对 PPI 网络结构进行建模,但是这些网络中持续存在的噪声和不完整性使得对其结构的结论变得困难。使用 2018 年 9 月 BioGRID 和 2019 年 1 月 IID 的更新、更大的网络,我们展示了度积和共同邻居的联合分布对 PPI 边缘连接的影响比它们的个体分布更大,并引入了两个新模型(CN 和 STICKY- CN) 用于采用这些功能的 PPI 网络。由于基于 Graphlet 的度量被认为是最有洞察力和最敏感的网络比较工具之一,我们使用 Graphlet Kernel (GK) 评估它们对 PPI 网络的整体全局和局部拟合。我们将 10 个理论模型拟合到 9 个 BioGRID 网络和 12 个集成交互式数据库 (IID) 网络,并发现:(1)STICKY 和 STICKY-CN 是根据 GK 的整体全球最佳拟合模型,(2)双曲几何图形模型更好比任何基于 STICKY 的模型对 4 个物种的拟合,(3)尽管 STICKY-CN 提供了比 STICKY 模型更好的局部拟合,但 CN 模型对大多数物种提供了最大的局部拟合。我们得出结论,将 CN 包含在 STICKY-CN 中使其成为 PPI 网络的最佳整体,因为它非常适合本地和全球。(3) 尽管 STICKY-CN 提供了比 STICKY 模型更好的局部拟合,但 CN 模型对大多数物种提供了最大的局部拟合。我们得出结论,将 CN 包含在 STICKY-CN 中使其成为 PPI 网络的最佳整体,因为它非常适合本地和全球。(3) 尽管 STICKY-CN 提供了比 STICKY 模型更好的局部拟合,但 CN 模型对大多数物种提供了最大的局部拟合。我们得出结论,将 CN 包含在 STICKY-CN 中使其成为 PPI 网络的最佳整体,因为它非常适合本地和全球。
更新日期:2020-08-18
down
wechat
bug