The microarray menus provide facilities for the design and analysis of microarray experiments. Microarray experiments are used in genetics to screen RNA samples against libraries of DNA/RNA samples/fragments.
Two channel microarrays
A microarray experiment consists of a series of slides that have been printed with spots of DNA from a library. The DNA printed onto the slides or chips is known as the probes. The spots can be printed using a robot or ink jet printer to transfer DNA from the library (often held in 96 or 384 well plates) to the slide. Alternatively the probes can be laid down on the slide chemically, building up the probes base by base. RNA or DNA from two sources, known as the targets, (e.g. individuals, cell lines, cells at different times or under different conditions etc.) are then bound to 2 fluorescing dyes (Cy5 = red and Cy3 = green, note: to help remember the colour of the dyes Cy3 and Cy5, the words three and green share 3 letters). Two dyed targets are then added to a slide in equal amounts. The target RNA binds to the probe DNA on the slide, where there is a long sequence of complementary bases (i.e. the base A binds with T and C with G). This step is known as hybridization, and is normally carried out over a period of time under controlled conditions (e.g. temperature etc.) After the hybridization step, when the targets are left to bind to the probes, the unbound target DNA is washed from the slides. The amount of bound DNA for each target is read by scanning the slide, often with a confocal scanning laser which produces an image of the slide. The scanner uses two lasers tuned to the Cy3 and Cy5 wavelengths to produce a green and red image.
The following image shows a section of a microarray slide printed by one pin. The complete microarray slide can be seen here.
The two images are then input into an image analysis package (such as Spot, GenePix or ScanAnalyze) to produce a data file with a single level of intensity for each colour/channel and each spot. The image analysis routine should locate each spot, and then read an average value for the signal for each spot (the foreground value). In addition it may also provide background values for each spot to allow for background correction.
See the two channel microarray example to view how the menus can be used to analyse a two channel experiment.
One channel/Affymetrix microarrays
Affymetrix produce chips which have just a single target added to each chip. A single level of intensity is read from each spot or cell on the chip. Each gene is presented on the chip by multiple short sections of DNA (probes) which match the gene at various locations. The set of sections is known as a probe set. For each matching section of DNA (the perfect match (PM)), as a control, a corresponding section of DNA with a single base change near the middle of the sequence is also put on the chip (the mismatch (MM)). The pairs of PM/MM spots are then summarized over to obtain an average reading for the presence of each gene. The images below show an Affymetrix chip and its components.
Further help can be found on the Affymetrix GeneChips help page or by visiting the AffyMetrix website.
Affymetrix chips can be very large (for example they can have about 500,000 spots on them), so analysing a series of Affymetrix chips can be time consuming, and can require a large amount of RAM when the CEL files are processed within Genstat. As an alternative, Affymetrix supplied software can be used to summarize over the PM/MM pairs and store the results in summary files (CHP files), which can then be read into Genstat.
A CEL file provides data for each physical location, or cell, (defined by row and column coordinates) and intensity. In addition it contains the name of the CDF file needed for the location-probe-set mapping. The CDF files store the probe set related to each location on the array. The computation of summary expression values from the probe intensities requires a mapping/merging of the probes set (CDF) data to the location (CEL) data. When opening a series of CEL files in Genstat the corresponding CDF file name needs to be specified. CDF files can be downloaded from the Affymetrix website.
See the Affymetrix microarray example to view how the menus can be used to analyse an Affymetrix microarray experiment.
Menu Items
The microarray menu has been designed so that you can progress easily through each step of an analysis. For example, the first section is for designing an experiment. This is followed by a section for importing data and then calculating log-ratios. Further sections are available for exploring the data and then normalizing. The data is explored and then normalized. The Analyse section can then be used to estimate the effects and perform empirical Bayes estimation to add power from the parallelism of the many probes. The False Discovery rate can then be accessed to aid in deciding how many probes to follow up. The final two sections can be used to graphically visualize the results and to cluster probes and targets to look at groups or for structure in the experiment.
- Design
- Generate One Channel Design
- Generate Two Channel Design
- Check Two Channel Microarray Design
- Data
- Open Microarray Data Files
- Calculate
- Log-ratios
- Affymetrix Expression Values
- Explore
- Histograms
- Density Plot
- 2D Plots
- Spatial Plot
- Normalize
- One Channel
- Two Channel
- Analyse
- One Channel ANOVA
- One Channel Regression
- Estimate Two Channel Effects
- Robust Means Analysis
- Empirical Bayes Estimates
- False Discovery Rate using Mixture model
- False Discovery Rate using Bonferroni
- Display
- Volcano Plot
- Q-Q plot
- Cluster
- Probes/Genes
- Targets/Slides
- Two-way
See microarray procedures for details of the procedures that can be used to analyse microarray data, which are also used by these menus.