Preprocessing and differential expression with CARMAWeb
From LipidomicsWiki
(University of Graz P07 )
CARMAweb (Comprehensive R based Microarray Analysis web service) is an intuitive graphical user interface for the normalization and analysis of microarray data. Created by Johannes Rainer at the Graz University of Technology (Graz, Austria), CARMAweb currently supports microarray technology platforms such as Affymetrix GeneChips, two-color microarrays, and ABI microarrays. For details regarding CARMAweb, please see the official userguide.
CARMAWeb can be accessed via the following URL http://CARMAWeb.genome.tugraz.at
Contents |
Registration & Login
To register as a new user, please follow the "Create account" link in the left navigation panel. Otherwise, follow the "login" link to access your data directory and past data analyses.
Data Upload
After logging in, there are new menu options on the top, horizontal bar, as well as in the left navigation panel. To upload data, follow the "Data Directory" link found in the left navigation panel.
In most cases, the analysis will require upload of multiple files. To make this as easy as possible, CARMAweb offers a multiple file upload function. Click the link entitled "Upload Multiple Files" to start the Java file selection applet. In the resulting window, click on "Add Files" and use ctrl-click to select all relevant files before clicking "Upload" to send those files to the server.
When the file upload has completed, a click on "Data Directory" allows to check that all files are ready for analysis. This screen can be accessed at any time, and allows to download or delete files from the CARMAweb server.
Preprocessing & Normalization
- Click on "New Analysis" (found in the horizontal navigational bar)
- In the main content area, click on "Perform an Affymetrix GeneChip Analysis".
- Select all files to use in your analysis. *Only select those arrays corresponding to the conditions to be compared. When all are added, click "Next".
- Select "GCRMA" as the preprocessing and normalization method.
- After selecting the normalization type, select the plots you would like to see reported with your analysis results. This may be helpful in quality control of your arrays.
Differential Expression Analysis
- If using technical replicates, users should define their replicates here and go on to perform a fold change analysis. Since we are working with biological replicates, we select "skip", and move forward to "Define differentially expressed genes using test statistics".
- We must now define our biological replicates. Choose '2' groups from the drop down. Using the provided interface, assign your CEL files to their respective group (based on the condition). As of now, only two groups are allowed, so only once comparison at a time can be performed. For those CEL files not being considered, select the 'Skip' option.
- On the same screen, click the 'test statistics' tab, and select 'moderated t-statistics (limma)' as the test statistic.
- Click on the 'tables & plots' tab. Here you can manage the requested output from CARMAweb. If interested in differential expression between conditions, it is important to check the boxes to include M and A values, as well as annotation data. The M value represents the log of fold-change between conditions, so 2^M = fold change.
- Enter an analysis name, and click 'Analysis'.
- Click 'Start' to begin your analysis. While the analysis is running, you will see something similar to the image below. Upon analysis completion, all results will be saved under "Analysis/(Analysis Name)".
Accessing Results
- Click the name of your experiment in the left navigational panel.
- You will see a list of files produced by CARMAweb, each providing different information about your analysis.
- The default settings place all results in the file named "PValues.txt". Please note the importance of opting to include annotation, M (log fold change), and A (average expression) values in your output. Without these values, your results may be somewhat meaningless.
- Results may also be downloaded in .zip format by clicking the link found at the bottom of the results page.
Citing CARMAWeb
In the event that this work leads to publication, please cite CARMAweb in the following manner:
Rainer J, Sanchez-Cabo F, Stocker G, Sturn A and Trajanoski Z. CARMAweb: comprehensive R- and Bioconductor-based web service for microarray data analysis. Nucleic Acids Research, 2006;34(Web Server Issue):W498-503.
