Comments From the Maintainer: I believe this dataset is representative of many small lab experiments. Many of the arrays have small artifacts, but in general the arrays have about equivalent quality relative to each other. Groups of 3 consecutively labelled arrays correspond to treatment groups (1,2,3 is the first treatment, etc). There is an interesting pattern in the NUSE plot which can be explained as a temporal effect. This experiment was carried out over a period of 9 months, with arrays 1, 4, 7, 10 hybridized first, followed by the group 2,5,8, and 10 and finally the remaining arrays. Also arrays 3, 6 and 12 were scanned after the scanner had been tuned to different settings. This highlights how experimenters should be mindful about how they carry out their experiments and good experimental design is essential. Ben Bolstad |