Normalizing to housekeeper genes with microarray data

Question

[moonkinwen]

Hi, I want to analyze some gene expressions in human cancer sample, so I downloaded the raw CEL files from NCBI GEO from different GEO series(ie. GSE60697 and GSE9348). then i used RMA normalization for each series. and i found differences between samples from different series. like this:

          X1 1007_s_at 1053_at   117_at   121_at 1255_g_at  1294_at  1316_at  1320_at 1405_i_at  1431_at
1  GSM358347  11.00181 8.44362 7.158441 9.215471  4.769039 7.664939 6.415506 6.180747  8.607589 4.830004
2 GSM1581701  10.01269 7.81094 5.572275 7.395991  2.884339 7.597989 5.883037 4.473780  7.194491 3.380647

I think every value in the second line is smaller then the first line. if i want to analyze expressions from different series in a "pooled" dataset, should i re-normalize it to housekeeper genes?

[WouterDeCoster]

Isn't this microarray data? Since you write in your title and use the tag 'RNA-seq' I'm rather confused.

[moonkinwen]

sorry, it should be a microarray data

Answer 1

[Lluís R.]

Yes, you should be looking for batch effect and normalize/estimate the batch effect.