Data Quality Assessment

Assessing the data quality of experiments is essential for building a reference database with predictive value. That microarrays can contain small defects is well known.

Additionally, small changes in the experimental procedure may lead to deviations in data quality. Changes in the scanner setting or misalignment of the scanner may lead to faulty quantification of chip images (see Section 9.4). If unnoticed, erroneous data from such experiments can affect the quality of the toxicological reference compendium and reduce its robustness. A critical aspect of data quality assessment (and by the same token, of all the ensuing analysis steps) is the guaranteed reproducibility of all applied methods. Judgments based on visual inspection of microarrays are not sufficient and should be avoided in favour of well characterized automated computational methods [8].

0 0

Post a comment