GenePix Pro 3.0, Cy3:pmt voltage 532nm,550V,laser power 100%, Cy5:635nm,pmt voltage 600V,laser power 100%
Description
Epigenomic mapping
Data processing
Data were normalized as ChIP-chip data in Turck e al.(2007). This method is based on the properties of dye-swaps to remove technical biases. To be specific, let Yij be the signal of the sample labeled with the dye j on the array i. Given that the second array is a technical replicate of the first one, the distribution of Y21 (respectively Y22) should be close to that of Y12 (respectively Y11). In practice, the relationship between Y21 and Y12 is linear but it is not the identity function. The parameters of the two linear models are estimated by Y21=a+bY12+N(0,sigma2) and Y22=c+dY11+N(0,sigma2),and these estimates are used to define the normalized IP and INPUT values of the second array relative to the first one: Y21=(Y21-a)/b and Y22=(Y22-c)/d.For each sample and for each tile, the values of the two arrays of the dye-swap are then averaged.To analyze data, we use ChIPmix, a method proposed by Martin-Magniette et al.(2008) that we have adapted to study several biological replicates simultaneously. The method investigates the relationship between IP and Input by a mixture model of regressions. For a probe, available observation are the two measurements IP and INPUT for the two biological replicates. These latter are assumed to be independent by definition. The (unknown) status of a probe is characterized through a label Z which is 1 if the probe is enriched and 0 if it is normal (not enriched).