Partition clustering identifies four distinct patterns of gene expression in genome-wide analysis of Drosophila immune response
The following tutorial walks through the identification of biological themes associated with gene clusters identified in an immune response time series.
Before you begin, we recommend that you review the analysis summary. The top button of the four on the right will download it if you haven't already.
When you are ready to begin this hands-on tutorial, click the third button down on the right, labeled "Log in to the dataset with tutorial."
1. Select Projects in the menu bar on the left side of the screen labeled Control Panel. You will see “Projects” under the heading marked “Analysis.”
2. Select the magnifying glass icon next to “Immune Response::Filtered” from the list of projects.
Note: Immune Response::Filtered is a sub-project of Immune Response, which contains all genes on the Drosophila Genome I array. The filtered sub-project contains 624 genes that were differentially regulated with at least a 1.5 fold change across the time series and statistically significant (5% false discovery rate based on ANOVA followed by correction for multiple testing using the method of Benjamini and Hochberg).
3. The Project Summary section lists the time points examined in this project. The immune response time series examines gene expression in uninfected flies (0hr) and at 1.5, 3, 6, and 12 hours after infection.
Select the Cluster link from the Project Analysis section.
4. Partition clustering will be used to separate the 624 genes into groups based on expression pattern. Set the Clusters pull-down menu to 4 in the PAM section and click Search.
5. Each line graph summarizes the gene expression pattern for that cluster (each gene in a cluster has an expression pattern more similar to pattern shown in the graph than to the pattern associated with each of the other clusters). The number of genes in each cluster is listed below the graphs. Silhouette widths measure how the genes in each group are clustered and can be used to select the best number of clusters for a set of genes. See the help documents for this page for more information about silhouettes.
6. To view page-specific help documents for this, or any page, select the question mark icon located at the upper right page corner.
7. Select the cluster with 37 genes showing dramatic up-regulation during the time series.
8. The resulting heat map shows the expression pattern of the 37 genes in that cluster.
9. To view a summary of the Gene Ontologies associated with the genes in this list, click on the Ontology link to bring up the Ontology Report (go to step 12 in this tutorial).
10. To view data and a gene summary for any gene in the list, click the Gene Name. This will bring up a data summary and a One-Click Gene Summary™ (OCGS) for the gene. The One-Click Gene Summary provides a synopsis of current UniGene and Entrez Gene (formerly known as LocusLink) information for the gene.
11. Go back to the gene list by clicking the “Back” button in your browser.
12. Select the Ontology link at the top of the screen to view a summary of the Gene Ontology terms associated with the genes in the list. See the online help system for information about the other reports.
13. The Ontology Report lists the Gene Ontology terms associated with the gene list. See the help documents for this page for more information about the Ontology Report.
14. Click on Z-score report in the upper right corner of the Ontology Report window.
15. The z-score report lists the biological process ontologies that are significantly over or under-represented. Genes involved in the defense response are greatly enriched in this cluster.
Z-score reports can be generated for each of the clusters. The biological themes associated with each cluster are very different.
Z-score reports can be generated for the Molecular Function and Cellular Component ontologies as well.
Only a few specific aspects of the data set have been explored here. Feel free to examine the data further on your own.