Analyze normalized illumina microarray data

Question

[niu.shengyong]

Hi there, is there any way to use normalized signal density table from illumina microarray to do analysis? My goal is to get differential expressed genes. I've looked into limma, but not sure how to use such processed data.

Example:

#ID_REF = 
#VALUE = Normalized signal intensity
ID_REF  VALUE
ILMN_1762337    -11.53652
ILMN_2055271    62.16199
ILMN_1736007    79.71209
ILMN_2383229    13.5637
ILMN_1806310    10.5413
ILMN_1779670    5.988304
ILMN_1653355    26.11047
ILMN_1717783    -21.43603
ILMN_1705025    -13.99969

[ATpoint]

Start from the point right after normalization in the limma manual.

[niu.shengyong]

Thanks for your help. I've looked into the limma manual, but still don't know how to use this as input by limma package. Could you give me a hint where to start from the codes? Thanks!

[Benn]

I have no experience with Illumina arrays, but if you have a matrix with normalized expression data and you are sure the normalization is good for limma analysis. You'll have to make an Elist object of it first (assuming it is single channel technology). You can do this with NormalizeBetweenArray function, using method = "none".

E_list <- normalizeBetweenArrays(your_matrix, method = "none")

Then start with the design and fit, like ATpoint is advising.

[adam.faranda]

I'm not sure that this "normalized" data is good for Limma Analyses yet. The "background normalization" method in BeadStudio can introduce negative values[1], which seems to be the case in the OP's data set. I found the paper below to be very helpful when trying to analyze illumina data using the "lumi" package.

[https://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-11-349][1]

[niu.shengyong]

Thank you! There are many published microarray data in GEO are in this normalized format. I'm curious how did you guys work on that? Thanks