This chapter describes the application of the BeadArrayTM technology for gene expression profiling. It introduces the BeadArrayTM technology, shows possible approaches for data analysis, and demonstrates to the reader how the technology performs in comparison to alternative microarray platforms. With this technique, high quality results can be achieved so that many researchers consider employing it for their projects. It can be expected that the technology will gain much importance in the future. The author hopes that this résumé will introduce researchers to this novel way of performing gene expression experiments, thus giving them a profound base for judging which technology to employ.
The Beadarraytm Technology
In the first phase of microarray technology probes were spotted by robots at known locations onto a dedicated substrate. Probes were then hybridized with a radioactive or fluorescent labeled target. Thus, the abundance of hybridized material was transformed into a signal which could be read by a scanner. Many error sources influenced the results achieved by this technique: twisted spotting needles, needles transferring different DNA volumes, labeling differences for, e.g. red and green channel, dust on the substrates, systematic local background changes, bad signal-to-noise rates in the scanned image, etc. . A great improvement in quality was achieved by synthesizing oligonucleotides using photolithographic processes known from the semiconductor industry. As a logical consequence the product was called “chip” as a reference to the origin of its manufacturing technique.
An alternative approach claiming to reach similar quality benchmarks as the chips is the BeadArrayTM technology. It takes advantage from the ability of beads to be randomly assembled at very high densities. In the literature it has been described that beads of 300nm have been randomly assembled into 500nm wells (Michael et al., 1998). For the BeadArrayTM technology at present a size of 3μm is used for the silica beads. The beads are generated by joining oligonucleotides to their surfaces and are pooled in libraries. They are self-assembled into etched substrates.
With the beads randomly distributed over the array the problem of decoding each bead’s information content arises. The solution of this problem is described in (Gunderson, K. et al., 2004) and is an essential precondition for the employment of the technique.
Key Terms in this Chapter
Gene Chip: A gene chip contains a matrix of photolithographically assembled oligonucleotides representative for an organism’s genome on a dedicated substrate.
VSN (Variance Stabilizing Normalization): The VSN (Variance stabilizing normalization) transforms the data in such a way that the variance remains nearly constant over the whole intensity spectrum. Without this (or another) normalization a dependency between intensity and variance can be observed in may cases which deteriorates the analysis results.
Clustering: Clustering is an analysis method for grouping of probes and samples by similarity. Similar data sets fall into the same cluster while dissimliar data sets fall into different clusters. For the hierarchical clustering a hierarchy of clusters is determined. Thus, one large cluster comprising all data sets is stepwise subdivided into smaller clusters down to singletons which are clusters containing a single data set.
Microarray: A microarray is a collection of genetic substances, - mostly DNA(Deoxyribonucleic acid) - integrated at a large scale..Microarrays can be generated in high-throughput using robots transferring minimal amounts of probes onto substrates. The sample being tested is radioactively or fluorescently labeled and hybridized to the probes. A scanner scans the microarrays delivering signals for the labeled and hybridized substance. “Chips” (see Chips) and “bead arrays” (see bead array) can be regarded as subsets of microarrays.
BLAST: (Basic Local Alignment Search Tool) is an bioinformatics algorithm applied to align sequences versus each other.
Affymetrix: Affymetrix is a trademark of Affymetrix, Inc. .
Correlation Coefficient: The correlation coefficient tells how good two data sets correlate. The correlation coefficient ranges between –1 and 1. Good values are close to one, values near zero represent a random dependency, negative values a reciprocal dependency. The correlation coefficient calculated directly from the data is called Pearson correlation coefficient. The Spearman correlation coefficient first ranks the data before calculating the (Pearson) correlation coefficient. Thus, it is more robust against outliers than the Pearson correlation coefficient.
BeadArray: BeadArray is a trademark of Illumina, Inc. .
XML: XML is the extended markup language. It is a more general version of the Internet description language HTML (hypertext markup language) allowing the detailed description of documents.
Bead Array: A bead array is an array of randomly assembled beads covered with oligonucleotides representative for an organism’s genome.
Gene Expression: Gene expression is the transformation of a gene’s information by transcription and translation ( Lewin, 2002 )
Coefficient of Variation (cv): The coefficient of variation is the standard deviation divided by the mean. It is a measure for the reproducibility. For a low cv data has a good reproducibility, for a high coefficient of variation data contains much variation and thus measured values can only be reproduced very inprecisely.