On the relevance of technical variation due to building pools in microarray experiments [Mouse]

The objective of the study was to assess the technical error due to blending of individual samples into pools in experimental data. The blending error variance component corresponds to random effects for inaccuracies, which were modeled on the logarithmic scale of normalized gene expression. It's estimation based on a linear mixed model, fitted for each transcript. Ovaries from 30 animals of line FL1 and of the control line DUKsi were collected at the metestrous stage. They were combined to 6 pools of 5 individuals each and processed for microarray analysis. Additionally, the 10 individuals used for the first two pools of each line were processed.