|
|
 |
/ Home / Analysis Tools / Differential Analysis
Search for differentially expressed genes from replicated experiments
Version 1.6
This tutorial can be download
as a PDF file (994.4 Kb).
Once you have your normalized data file, open it with Excel.
You can filter out weak intensity spots (eliminate the weakest intensities in
both channels) keep spot with ratio greater than 1 or lower than –1. Remember
we are working with log2(ratio) so log2(2)=1. This method called “fold
change” is the one used at the beginning of microarray analysis and is
still useful if you do not have enough replicates to apply statistical treatments.
The “fold change” method lack accuracy regarding
the significant threshold to be fixed. That’s the reason why it is useful
to apply a statistical method able to take into account intensity variations
and most of all, the variability among experiments.
Significance Analysis of Microarrays (SAM):
SAM is an Excel macro freely available for academics on the
web. The use of SAM in Excel spreadsheet makes this tool easier to use for most
of microarray users. Using SAM implies several modifications in your data file:
- The ratio or intensity values in the Excel sheet must not contain any comas
but only points as decimal separator.
- The header line depends on the type of analysis you want to perform. You
can refer to SAM manual for more information. So you must duplicate your header
if you don’t want to loose the experiment information (see image below).
- Two annotation columns are available. SAM always references its calculation
to the line number in the departure sheet.
- Before launching the macro, it is necessary to select the data precisely
because SAM rejects lines with too much missing values (such as empty lines).
- When the SAM macro is launched in the tool bar (“SAM”), a setting
window appears. For further informations on the various options you can choose,
the best is to refer to the SAM manual. However, the first important things
to do is to indicate if the data source has been transformed in log2 or not,
then, as data bootstrapping uses a random generator, you need to initialize
it several times by creating a various number of seeds.
- Once all the chosen iterations have been done, SAM displays a plot representing
each gene thanks to its score in the real distribution compared to the random
distributions. Therefore, the differentially expressed genes are the ones
moving away from the 45° slope line.
- First, display the delta table. This table indicates for each delta value,
the number of putative differentially expressed genes, the significant genes,
and the number of false positive genes estimated using the False Discovery
Rate (FDR). The user fixes the delta value according to the number of false
positive or significant genes he wants to obtain.
- To choose the delta value, get back to the SAM plot sheet and display the
“SAM plot controller” by clicking on the SAM macro button.
- The SAM Plot Controller window lets you fix the delta value you want: “Manually
Enter Delta”. Then if you select the “List Significant Genes”
button, SAM displays the list of differentially expressed genes in the “SAM
output” sheet according to the delta value you chose.
- This sheet summarizes the selected parameters and gives you the list of
induced and repressed genes.
Useful links:
- SAM (Significance Analysis of Microarray), Excel macro
allowing to search for differentially expressed genes using a bootstrapping
method. Website: http://www-stat.stanford.edu/~tibs/SAM/
|
