rcModelMedianPolish of the package preprocessCore

[Julie77]

Hello,

I want to imitate RMA on illumina and agilent chips. There is no problem with background correction, quantile normalization and log2 tranformation.

But I don't know what I should do exacty with the median polish.

I use an example matrix which look like this:
> matrix
[,1] [,2] [,3] [,4] [,5]
[1,] 18 11 8 21 4
[2,] 13 7 5 16 7
[3,] 15 6 7 16 6
[4,] 19 15 12 18 5

Than I do the median polish
> rcModelMedianPolish(matrix)
$Estimates
[1] 16.25 9.75 8.00 18.00 5.00 1.25 -2.75 -1.25 2.75

$Weights
NULL

$Residuals
[,1] [,2] [,3] [,4] [,5]
[1,] 0.5 0.0 -1.25 1.75 -2.25
[2,] -0.5 0.0 -0.25 0.75 4.75
[3,] 0.0 -2.5 0.25 -0.75 2.25
[4,] 0.0 2.5 1.25 -2.75 -2.75

$StdErrors
NULL

>

so what should I do next? What are my expression values. I have read that I must substract the Residuals from my original data

> MedianPolish<-rcModelMedianPolish(matrix)
> matrix-MedianPolish$Residuals
[,1] [,2] [,3] [,4] [,5]
[1,] 17.5 11.0 9.25 19.25 6.25
[2,] 13.5 7.0 5.25 15.25 2.25
[3,] 15.0 8.5 6.75 16.75 3.75
[4,] 19.0 12.5 10.75 20.75 7.75


But somewhere else I've read that the first 5 values of the Estimates are my expression values. But in this case I have for example 4 different genes. So this sounds not correct.

So what is the right procedure?