The Affymetrix GeneChip® Laboratory is a Purdue Cancer Center Shared Resource.
How to contact us:
Array Coordinator: Ann Feil.
Ann's email: afeil@purdue.edu
phone: 67368 (Affy Facility) or 66329 (High Throughput Laboratory)
Genomics Core Director: Phillip SanMiguel
phone: 66328
Phillip's email: pmiguel@purdue.edu
Please contact Ann Feil prior to bringing RNA samples for GeneChip® Analysis.
As of January, 2008 we are using new Enzo kits for cDNA/cRNA synthesis. These Enzo kits are incompatible with certain Qiagen kits that you may be using for sample prep. So if you are thinking about using Qiagen kits then please consult with us. To be more precise, it is the 'polyA RNA carrier' that is used in extracting total RNA from small samples (such as in the Qiagen RNeasy Micro Kit) that is incompatiable with the VanGelder/Eberwine methods such as we use from Affymetrix and Enzo.
Results
Access to the results below will require your LTL (low throughput) password. These results are just the raw data files. You will need to do the analysis via the methods outlined below. As of Apr-2005 all files are being bundled together into the CAB format that was used by the Affy GCOS software [note that GCOS is no longer available; we include it as a historical reference.] This format can be read by GCOS, WindowsXP and many un-zip programs.
We no longer provide a tab-delimited text file automatically. We can do so upon request.
Click on your last name in order to retrieve your results
| Customers | LTL | FL | Genotyping | Affy | All services |
|---|---|---|---|---|---|
| Genomics personnel only | LTL | FL | Genotyping | Affy | All services |
| Abernathy, Brian | |
| Bhunia, Arun K. | |
| Carpita, Nicholas C. | |
| Chapple, Clinton C. | |
| Kim, Chang | |
| Konieczny, Stephen F. | |
| Kuang, Shihuan | |
| Lelievre, Sophie | |
| M. Sedlak, | |
| Mendrysa, Susan | |
| Muir, Bill | |
| Murdock, Larry | |
| Ogas, Joseph P. | |
| Salt, David E. | |
| Sherman, Louis A. | |
| Szymanski, Daniel B. | |
| Woloshuk, Charles P. | |
| Xu, Jin-Rong |
Analysis
The first step -- which is image analysis in order to get the results from the raw image data into a spreadsheet compatible format -- should be done via Affymetrix's Expression Console (this has taken the place of their GCOS software) or via 'Bioconductor/R'.
The important statistical steps can be done via 'Bioconductor/R' or 'SAS microarray solutions.' This software is available at the Genomics facilty. Get hold of our system administrator Doug Yatcilla if you need an account on our computers.
Another option is to use the Purdue Statistical Consulting Service (SCS) which can provide personalized help in your analysis. Bruce Craig is one person at the SCS who knows a lot about microarrays. Please note that we do not automatically forward your samples and CAB files to the SCS. You are responsible for doing so.
As a historical reference the following paragraph is left on this web page. Note
that Affy now provides the
Expression Console instead of GCOS.
Yet another option is for your lab to do the statistical analysis via the
Affymetrix GCOS software yourself.
Here are some instructions on how to
install the software on your own PC.
You will need to
install your genome library
which involves yet another step.
No matter which of the statistical methods above that you use make sure that your data is all scaled and normalized using the same parameters. It is probably obvious that this should be done. However we have seen erroneous results because of failure to do the same scaling & normalization either in the conversion of the image to spreadsheet compatible data (via GCOS) and/or in the statistical analysis stage. Also note that the use of Excel to do the analysis is discouraged. Excel is made for business applications and can potentially cause problems with statistical applications such as the ones needed by Affy chips.
To reiterate the above point. We can generate .CHP files that have been target value normalized -- which is the default setting for the Affy software. The reason we do not, by default, is that when we first set the service up, Purdue statisticians told us not to do any normalization or scaling. Generally statisticians want to do their own normalization/scaling and dislike the idea of getting pre-normalized data.
We continue to operate this way just to be consistent. We believe that most of our customers use the upstream (non-processed, non-gene-summarized) .CEL files for their analysis. For them the normalization issue is moot, because it happens during the conversion from .CEL to .CHP data formats. That said, if you would like us to generate target value normalized .CHP data, just say the word and we will. It is an issue infrequently enough, that we sometimes to forget about it. But it just involves our clicking a box and re-generating the .CHP files.
| Genomics Home | ||||
| Affymetrix microarray home | ||||
 | Purdue University | Purdue Genomics Core Facility | ||
| Room S045 WSLR (Whistler Hall) | ||||
| Copyright © 2006, Purdue University | 170 S. University St. | |||
| All Rights Reserved. | West Lafayette, IN 47907-1057 | |||