Diagnosis of Microarray Data Quality

Genedata's strategy for detecting errors in gene expression measurements is outlined in this section. Although microarray technology is astonishingly precise and reliable, its routine application in toxicology and pharmacology requires process control that takes all errors into account that may have a significant impact on the results.

The accuracy of microarray measurements depends on a high signal-to-noise ratio. Although frequently expressed genes are relatively easy to measure, rare transcripts can no longer be accurately measured if their expression falls within the range of technical noise. As many genes involved in regulatory systems are expressed at relatively low levels, data quality assessment is essential for generating meaningful data.

Errors can occur at each step of a hybridization experiment, for example, during extraction of the genetic material, microarray hybridization, or quantification of the scanned image [12, 13]. Data quality control must include quality measures that cover the entire process - from sample extraction to the digital quantification of expression signals. Some examples of potential errors and their corresponding quality measures are discussed below and demonstrated in Figure 9.2.

0 0

Post a comment