Before proceeding to normalize the arrays or any other higher analysis, it is important to take a look to the quality of the data. It is critical to do an adequate quality assessment to make sure the data is of high quality and is consistent and comparable for further analysis. In some cases, arrays are too bad to be corrected, even with normalization; these arrays should be removed from further analysis.

Chip pseudo-images are very useful for detecting artifacts on arrays that could pose potential quality problems.