After background correction, normalization, and fitting a linear model, down-regulated genes were recovered.
Background subtraction, data normalization, and fitting to the EXAFS spectra were performed using the software packages Athena and Artemis.
2 Robust Microarray Analysis (RMA) was used for background correction of.CEL DTPs (n = 1,215), followed by quantile normalization, and fitting of a multichip linear model to each probe set; all techniques were implemented using the "affy" library from Bioconductor.
Briefly, fluorescence data processing encompassed the filtering of spots marked as unreliable by the scanning software or weak (when compared to control elements) and loess normalization before fitting a two-stage, mixed-model analysis.
The analysis of the microarray data consisted of the following steps: 1) within-array and between-array normalizations; 2) fitting the data to a linear model; and 3) computing differential gene expression.
A commonly used normalization method is the intensity dependent LOESS normalization that fits a locally weighted polynomial regression to the average of the red and green intensities, that is, the LOESS curve [ 2, 4].
Finally, some normalization methods fit the location and the shape of the original distribution (e.g. quantile normalization), or adjust data in a intensity-dependent way (e.g. variance stabilization, LOESS).
For this, we used an ANOVA normalization, i.e. fitting the "normalization" ANOVA model, that estimates non-biological variation due to Dye [Fixed] and Subarray [Random] and saving the residuals [ 66].
Intensity-based LOWESS normalization instead fits a smooth regression line through all M value points.
Then it uses a quantile normalization to fit each β-distribution of the Infinium II profile to the corresponding β-distribution of the Infinium I profile.
