Similar protocols

Pipeline publication

[…] , GSE15653, GSE29718, GSE48964, GSE9624) and T2D (GSE18732, GSE13760, GSE20966, GSE23343, GSE25724, GSE38396, and GSE38642). All of these datasets were curated and reported in the GEO Datasets (GDS). Each dataset was required to have at least three samples for both case and control groups. And the samples from these patients who suffered both obesity and diabetes were excluded., The preprocessing of microarray data was conducted by the RMA [–] integrative method, and the statistical analysis of gene differential expression was computed by the linear models and empirical Bayes methods []. And then the P values of each gene were obtained., From the gene network of obesity and T2D, we used the jActiveModules [] and multiple gene expression profiles to find the active gene modules showing significant changes in expression in disease/normal conditions. The jActiveModules (Version 1.8) is a widely used method for identifying active modules integrating multiple gene differential expression datasets. In the algorithm of jActiveModules [], the P values of each gene in a subnetwork in a single condition are transformed into one standard normal z-score by the binomial order statistic. The highest score obtained in multiple experiments is recorded as the final score for a subnetwork. Higher z-score represents more significant expression changes. Here the top 5 scoring modules of obesity and T2D were enumerated separately by jActiveModules with default parameters in Cytoscape []., In order to further identify the subclusters with tight topology structures, we decomposed the active modules and the disease seed genes into several subclusters. As a result, ten and seven subclusters were identified by the MCODE method [] for obesity and T2D, respectively. The workflow was illustrated in ., Through collecting the protein interactions and transcriptional regulation data of the known genes of human obesity and T2D and their interacting neighbors from HPRD, TRANSFAC, and KEGG pathways, we compiled a multi-level biological network of human obesity and T2D called NOT2D (gene network of obesity and type 2 diabetes) (). As shown in , the majority of links in the obesity network were obtained from KEGG database while HPRD and KEGG databases contributed almost equally to the T2D network. Very few interactions are reported by two or more data sources. Finally, there are 606 nodes and 2907 edges in the o […]

Pipeline specifications

Software tools jActiveModules, Cytoscape, MCODE