1. DNA Microarray Data Acquisition and Analysis - Introduction to Stanford Microarray Database
AFGC
Damares C Monte
Carnegie Institution of Washington,
Plant Biology Department
Stanford, CA

2. DNA Microarray Technology
AFGC
Cy5: ~650 nm
Cy3: ~550 nm
Walk through exact method of hybridizing and making probe for microarrays. Discuss the CY Dyes used and current scanning technology.
No differential expression
Induced
Repressed

3. Data Analysis Acquisition - QC Input/Storage/Retrieval
Analysis/Pattern Recognition
Visualization
Interpretation/Annotation
Publication/Repository
SMD
Biologist

4. Fluorescence Intensities Extraction
To extract data from a microarray by accurately identifying the location of each of the spots.
Data extracted on: Fluorescence intensities, background intensities, fluorescence intensities ratios.

5. Load Images
Adjust Gain (brightness of image) and Normalization (balance between images in red and green)

6. Create Grids Create new grid Number of grids to create
Numbers of rows and columns
Spot width and height
Column and Row spacing
Resolution (Xres and Yres)
Tip spacing
Create grid

8. What gets flagged? SMD Overlay Cy3 Cy5 Dust speck
Saturated in both channels
Doughnut
Saturated on outside & in the middle doughnut
Saturated in one channel
Streak
Dust/Background/saturation/something coming into spot

11. Mean Intensity plots
Broad distribution pattern with curved or flattened arms. Common to have gap in the center. (Crab Claw)
Concentrated cluster of spots in a linear or fan pattern with clearly distinguished outliers

13. Median signal to background (How strong the signal is compared to background)
Mean of median background (Determines how high is the background)
Median signal to noise (Confidence to quantify peak signal to background)
(F635Med - B635Med)/B635SD

14. Data Analysis Flow New Scan Gene Pix Clustering
Quality Control
Clustering
Stanford Microarray Database - SMD
Data Selection
Quality Control
Complete Data Table (cdt)
XML
Download

20. Comparison plot
This plot can be effectively used to get a measure of the overall consistency between two duplicated or reversibly duplicated experiments. It compares the log2(RAT2N) values of the two experiments and calculates the regression line, which optimally should be close to one. Future functions will include filters and plotting of multiple experiments.

23. Why do we use log space? Exponential distribution of intensities
Most genes at low level
Very few high level
Log scale approx. normal distribution ??

24. Frequency of intensity levels This plot allows the user to plot the frequency of spots within certain intensity intervals for the red and green channel respectively. It is possible to use normalized or non-normalized intensity values and linear or logarithmic scales. This graph will make it easier to determine background levels for filtering when using other analysis tools. The two examples below give a good view of what normalization does to the data.

25. Frequency of intensity levels

26. Distribution of ratios
This plot draws the distribution of the logarithm (base 2) of the red/green ratios [log2(RAT2N)]. It gives an immediate overview of the range of expression and the normality of the distribution.

34. Clustering and Image Generation Page

36. Spot Data Page

40. Things to be Considered When Analyzing Microarray Data
1. Clones corresponding to spots with intensities of at least 350 in both channels and
ratios of 2.0 / 0.5 are worth further investigation.
2. For genes of interest, check that the spot(s) is acceptable.
3. Check for consistency among replicate values.
4. Validate important results obtained from the microarray data by an independent test or method.
@pilot

48. Acknowledgements Shauna Somerville Katrina Ramonell Lorne Rose
Bi-Huei Ho
Sue Thayer
Shu-Hsing Wu

49. Acknowledgements Shauna Somerville
Bench team
Katrina Ramonell
Lorne Rose
Bi-Huei Ho
Sue Thayer
Shu-Hsing Wu
Bioinformatics team
David Finkelstein
Jeremy Gollub
Fredrik Sterky
Rob Ewing
Lalitha Subramanian
Mira Kaloper
Todd Richmond
Mike Cherry