Gene Array Facility Protocols

Aminoallyl probe labeling (PDF 16kb)

How To Create Hyperlinks Protocol (PPT 4.0Mb)

A Short Explanation of Microarray Technology We are Using

DNA microarrays provide a method of comparing large populations of genes between distinct samples. RNA, either total or poly (A+) selected, is extracted from two sources (i.e. knockout versus wildtype, drug-treated versus untreated, diseased versus healthy tissue or cells in culture) and run on a gel to establish quality. High quality RNA is an essential component of a good microarray experiment. We currently recommend that our users use the QIAGEN RNeasy® mini kit for RNA extraction combined with their on-column RNase-free DNase treatment. We recommend following the protocol exactly, including use of the QIAshredder® mini-columns to ensure adequate cell disruption. For our own experiments with testis, brain, and other tissues, we transfer the frozen tissue to a QIAshredder® mini-column with the RLT + BME buffer and then sonicate the sample in the QIAshredder® column itself. An example of high quality RNA extracted from tissue shown below.

The mRNAs to be compared are labeled with two different fluorescent cyanine dyes (Cy 3 or Cy5) using an indirect labeling method. Amino allyl-dUTP is incorporated during reverse transcription and then post-coupled with a NHS-esther mono-reactive cyanine dye. This indirect labeling method allows us to avoid one common form of false positive resulting from the fact that the reverse transcriptase enzyme differentially utilizes the Cy5 and Cy3 dyes. The two labeled cDNA samples are then combined and hybridized on a single microarray slide overnight using a standard formamide-based RNA hybridization protocol (see “Protocols” toggle on website). The following morning the slide is washed, dried, and scanned using the Axon 4000a scanner. The GenePix® software on the Axon 4000a scanner, after simultaneous laser scanning dual excitation/emission capturing the Cy5 and Cy3 channels, creates a merged 16-bit TIFF image, finds the spots, performs a background subtraction/signal to noise analysis, and then generates a ".gpr" file containing a variety of parameters concerning each individual spot. The core ".gpr" file, which can be examined in Excel format, is then imported directly into the Silicon Genetics GeneSpring® software. A normal experiment involves slides corresponding to each individual biological sample and a dye-swap for each biological sample (a technical replicate), along with multiple biological samples in each treatment group (biological replicates). Thus, multiple individual “.gpr” files must be imported into GeneSpring® to create a single experiment. After the “.gpr” files are imported into GeneSpring®, all slides are normalized using a LOWESS (Locally Weighted Regression and Smoothing) function to address intensity-dependent dye biases. Once all the slides are entered and normalized, the slides corresponding to each treatment group (including dye-swaps) can be merged, the background-subtracted data for each spot log-transformed, spots not above the background threshold removed, and the data visualized and analyzed for statistical significance. It is from the experiments created in GeneSpring® that the Excel spreadsheet data report derives, which includes the normalized ratio of treatment to control, the treatment average for each spot, the control average for each spot, the T-test p value of treatment and control values for each spot, and the annotation (Plate ID, Genbank, LocusLink, Gene Name, Gene Symbol, and Gene Description) for each spot. If ANOVA has been performed on two or more treatment groups, then the p value displayed represents the intra-group versus inter-group variability p value (F test).

What You Need to Provide for a Microarray Experiment

One of the most important components of a successful microarray experiment is high quality RNA. It is very important that the RNA is not degraded and not contaminated with genomic DNA. We are happy to provide a copy of our RNA isolation protocol.

We need at least 20µg of each RNA type (i.e. an experimental and a control) per slide when using total RNA, or 2µg of each RNA type for poly (A+) selected mRNA. In order to improve the statistical power of our microarray experiments, we have increased the number of slides run for each treatment/condition, including running several dye-reversals, which helps us to eliminate some false positives. For experiments using mouse or human tissue, we generally perform arrays with a minimum of four to five individuals per treatment and control group. The tissue from each individual sample is also dye-reversed (a technical replicate). This greatly improves the quality of the results we obtain, but also requires that you provide us with at least 40ug total RNA from each individual tissue sample. If the samples are from a tissue source that is scarce (i.e. pituitary, hypothalamus), the RNA can be amplified using one of several commercially available kits (Arcturus, Ambion) with well-established protocols. It is a well-established fact that spotted cDNA microarrays have large slide-to-slide variation, which has to do with spot effects, probe labeling effects, and several other sources of non-biological variation. The best way for us to deal with this non-biological variation and to accurately assess biological variation is to do multiple slides and multiple experiments.

We also would like to address biological variation within the experimental system. If a researcher is examining the knockout of a particular gene on a particular mouse genetic background, we generally run the individual knockout animal tissues versus pooled wildtype tissue. By pooling the wildtype animals, we essentially normalize all the slides and thereby greatly reduce the slide-to-slide variation inherent in spotted arrays. We also generally run a pooled wildtype versus pooled wildtype slide to give us an estimate of the slide-to-slide variation due to non-biological parameters. Please feel free to contact us if you have questions about setting up your experiment. Below is an example of a pooled wild type vs. pooled wild type control experiment viewed with Gene Spring software.

*Funded by the Specialized Cooperative Centers Program in Reproductive and Infertility Research NICHD

This page last updated: Thursday, 29-Mar-2007 14:35:40 PDT          Contact the webmaster