Transcript PPT - MIT

Data Analysis
DNA Microarrays
Dr. Rebecca Fry
Raw Data File
ID
Cy3 Spot Mean
Cy3 Background
Cy5 mean
Cy5 Background
Data Analysis

Step 1-Subtract Background Intensity from
each spot
Generate Excel Formula
Paste into all cells
Data Analysis

Step 2-Caculate average cy3 and cy5
intensities for control spots to determine
normalization factor
Sort File Based on Clone ID to locate controls
File is now alphabetized
Calculate average signals for cy3 and cy5
Calculate average signals for cy3 and cy5
cy3=32
cy5=9
cy3/cy5=
3.55

Normalization factor=
cy3/cy5=#
#=normalization factor
Our data=3.55
Data Analysis

Step 3-Normalize data using cy5 adjustment
factor
Paste into all cells
Data Analysis


Step 4-Filter Genes showing no expression in
both channels
Filter Genes with ½ intensity (backgound
subtracted, normalized) of the average
control intensity in both channels
Generate Excel Formula
Paste into all cells
Filter genes that are not present in cy3 or cy5
Data Analysis

Step 5-Calculate Ratios of Gene expression
change between final intensity value
(background subtracted, normalized)
Oligofectamine control (cy3) and knockdown
(cy5)

Cy3/cy5=ratio=fold change
Why we filter!
Data Analysis

Step 6-Generate Summaries




Number of Genes with differential gene
expression fold change above a cutoff (1.5)
Number of Genes expressed on array (need not
be differential)
How do you explain number of genes expressed
(biology/technology)?
How could you test this hypothesis?