The BeadStudio software package is included with every Illumina BeadStation and BeadLab System and provides a tool for analyzing gene expression data. After scanned microarray images of Sentrix® BeadChips collected from the Illumina BeadArray Reader, image data files are directly downloaded into BeadStudio for data visualization and analysis. BeadStudio executes two types of data analysis: gene analysis, or quantifying gene expression signal levels, and differential analysis, or determining if gene expression levels are different between two experimental groups. The gene analysis tool produces output files containing probe and gene lists, associated hybridization intensities (normalized or raw), and control summary reports. With the Illumina Bead technology, a single hybridization of RNA from one cell line sample to an array produces approximately 30 intensity values for each bead type, i.e. transcript. Gene signals are ranked relative to the distribution of signals of the negative controls and detection scores are calculated.
In addition, data visualization tools can create more sophisticated plotting analyses such as scatter plots, perform cluster analysis and produce resultant dendrograms, as well as output control summary graphs. Normalization of the expression signal is required for both gene expression and differential expression analysis. When performing gene expression analysis, only the method and error model need to be chosen to perform normalization (from several available); for differential expression analysis, a so-called reference group also needs to be selected from one of the pre-set groups of samples, which then will be compared to the so-called experimental group. Useful information contained in the files includes target ID, which identifies the transcript name; minimum intensity of the transcript in the group; average intensity of the transcript in the group; minimum intensity of the transcript in the group; number of arrays, or samples, in the group; array standard deviation associated with sample-to-sample variability within the group (undefined when the group contains a single sample).
Once gene analysis or differential expression analysis has been completed, generating scatter plots is the easiest method of data visualization. One of the samples (or group of samples) is plotted on X axis, the other, on the Y axis. A variety of options is available to detect and select specific genes and transcripts: Selection can be made based on their detection level, level of expression, or statistical differences between groups or samples, such as differential score and fold change. Just by just clicking on the name of specific gene, you can immediately access gene ID, symbol, accession number, definition, and ontology from the GenBank database as well as sample specific information on detection level, signal, and number of beads.
The clustering analysis method can be used to identify genes which show similar patterns of gene expression across a series of samples, or to identify which samples are most similar based on the expression levels of genes within them. Four clustering metrics for calculating dissimilarities (correlation, absolute correlation, Euclidean, and Manhattan) are available in BeadStudio.
I used BeadStudio software to perform data visualization, clustering, and differential analysis of the three Sentrix® Human-6 Expression BeadChips, each containing 6 samples from patients and controls. Overall, data quality was excellent with great reproducibility for duplicated samples. Correlation between samples duplicated on the same or on different chips was extremely high. I performed clustering and pair-wise differential expression analysis comparing reference group of controls to the condition group of affected. The comparisons were done using a T-test with the assumption of equal variance error model. Interesting results were obtained and will be published in a peer-reviewed scientific journal.
A number of functions, such as generation of heat plot from expression signals and multiple group comparisons, were not available as a part of BeadStudio. A new version containing these options will be a good choice for analysis and visualization of the expression data generated on the Illumina’s platform.