Penga Jiajie, Wang Tao, Huc Jianping, Wang Yadong, Chen Jin
School of Computer Science, Northwestern Polytechnical University, Xi'an, P.R. China.
Department of Energy Plant Research Lab, Michigan State University, East Lansing, USA.
Curr Genomics. 2016 Oct;17(5):427-438. doi: 10.2174/1389202917666160726151048.
With the rapid accumulation of gene expression data, gene functional module identification has become a widely used approach in functional analysis. However, tools to identify organelle functional modules and analyze their relationships are still missing. We present a soft thresholding approach to construct networks of functional modules using gene expression datasets, in which nodes are strongly co-expressed genes that encode proteins residing in the same subcellular localization, and links represent strong inter-module connections. Our algorithm has three steps. First, we identify functional modules by analyzing gene expression data. Next, we use a self-adaptive approach to construct a mixed network of functional modules and genes. Finally, we link functional modules that are tightly connected in the mixed network. Analysis of experimental data from Arabidopsis demonstrates that our approach is effective in improving the interpretability of high-throughput transcriptomic data and inferring function of unknown genes.
随着基因表达数据的快速积累，基因功能模块识别已成为功能分析中广泛使用的方法。然而，用于识别细胞器功能模块并分析其关系的工具仍然缺失。我们提出了一种软阈值方法，使用基因表达数据集构建功能模块网络，其中节点是编码位于同一亚细胞定位的蛋白质的强共表达基因，链接表示强模块间连接。我们的算法有三个步骤。首先，我们通过分析基因表达数据识别功能模块。其次，我们使用自适应方法构建功能模块和基因的混合网络。最后，我们连接混合网络中紧密连接的功能模块。对拟南芥实验数据的分析表明我们的方法在提高高通量转录组数据的可解释性和推断未知基因功能方面是有效的。
