Database query, GEO Profiles database
The GPr (GEO Profiles requests) program does elaborate requests on the GEO Profiles database in order to select studies that display a significant difference between two (or more) conditions for the input list genes.
The process is decomposed in three main steps.
In the first step, it retrieves studies that use all genes indicated in the “Gene list” text field. The organism of interest can be specified or not, with the help of the “Organism” drop-down list. We recommend the use of the “Organism” field, because it accelerates the computing time by limiting the number of studies that will be further analyzed.
In the second step, for each retrieved study, a p-value is computed with an ANOVA statistic test. This p-value indicates if there is at least one significant difference between the averages of two conditions. Each green box of the GEO Profiles graphic represents a condition. If a condition contains only one value, then the ANOVA test cannot be applied and no p-value is associated (represented by a '-' character in the output). In this case, we consider that this dataset has a significant difference, and let the user to check himself this dataset. The significativity threshold can be adjusted with the “Significativity threshold for the p-value” text field. We recommend the use of one of the following values (the less stringent to the more stringent): 0.1, 0.05 or 0.01. If we set the threshold to 1, then all studies are retrieved whatever their p-value. More than one dataset can be associated to a gene, in this case, we consider that this gene is significant if there is at least one dataset that is significant.
In the third step, we select the studies that have at least x percentage of significant genes among the listed genes. This value can be adjusted with the “Show only studies that have a minimum percentage of significant genes in the gene list” text field. If this value is set to 1, then only studies that have a significant p-value in each gene are shown. If it is set to 0, then all studies are shown, even if they have not any significant p-value.
This tool has no specific warning, however keep in mind that some queries can take some time. To decrease the computation time, don’t use the “All” option for the “Organism” drop-down list.
No input file is required to use this tool. Nevertheless, it uses some options that are described below:
The output is in html format. It summarizes first all the input parameters, and then lists the retained studies with these criteria. The studies are displayed in a table where each row represents a study and where:
Download the result web page on your computer to have a better view of the results, which is different in the galaxy viewing module.
July 3rd, 2014.