| Affymetrix Microarray Example |
| See Also |
These example data files can be found in the Examples\Microarrays folder of the GenStat installation.
To calculate expression values for these 9 slides, we first need to open the files. The CEL and CDF files can be opened individually using the Open dialog accessed via the File menu. However, there is a specific dialog available to open microarray data available on the Microarrays submenu of the Stats menu. To open this dialog select the Affymetrix CEL files option on the Data section of the Microarrays submenu on the Stats menu bar item as shown below:
To select the CEL files click on the browse button 
and select all the files
as shown below:
The CEL files will be opened in the order that the files appear in the list. The Up and Down buttons can be used to rearrange the order of the CEL files within the list.
Once the CEL files have been selected the corresponding CDF file can be selected by clicking
the browse button adjacent to the CDF file field and selecting
the file ATH1-121501B.CDF. Selecting this file will result in the completed dialog shown below.
Clicking the Open button will prompt you with the following dialog.
On this dialog we select the option for batch processing with the RMA method, and provide the filename, Expressions.gsh, to save the results. Note, this analysis can be very slow, as each CEL file contains over half a million observations. However, the results for the batch processing can be found in the GenStat spreadsheet file Hyb-Expressions.gsh located in the Examples\Microarrays folder.
Opening the CEL files or opening the file Hyb-Expression.gsh will produce a spreadsheet containing the following columns:
The data can be summarized using a single channel analysis of variance. Before the analysis can be performed the structure of the targets applied to the slides is required. This structure can be found in the file HybFiles.gsh. To open this file select the Open item on the File menu bar item, this will open the spreadsheet shown below.
Now, to summarize the data select Single Channel ANOVA from the Analyse section of the Microarrays submenu on the Stats menu bar item. The figure below shows the menu with the data names entered in the fields.
To estimate the difference between the standard treatment and the other two cell lines, we can specify a contrast, by clicking on the Contrast button. This opens the following menu:
On this menu we have selected the Target factor for the contrasts factor, set the number of contrasts to 2 and the comparisons contrast type. Clicking OK creates a blank matrix spreadsheet with 2 rows and 3 columns where the values for the contrasts can be entered. In this matrix (see below) we have entered 2 contrasts; the first compares line 1 vs the standard and the second line 2 vs the standard.
Returning to the single channel ANOVA menu, we now set additional options and specify the names of structures to save the results into. Clicking on the Options button opens the dialog shown below. Here we have left the options at their default settings.
The results from the analysis can be stored when the analysis is run. To store the results the names of the structures to save the results need to be supplied before running the analysis. To do this click on the Store button. This opens the dialog below where you can specify the items to be saved. The options at the bottom of the dialog can be used to control whether the results are to be displayed into spreadsheets.
Returning to the single channel ANOVA menu and clicking Run produces an analysis of variance for each probe
and displays the stored results in a spreadsheet (as shown below).
An alternative way to analyse this data would be to use the Robust Means Analysis menu. This menu can be accessed by selecting the from the Robust Means Analysis item from the Analyse section of the Microarrays submenu on the Stats menu bar item. The figure below shows the menu with the fields entered to perform the analysis.
Similar to the single channel analysis of variance menu, options can be set for this menu by clicking on the Options button. The figure below shows the default options set for a robust means analysis.
