[BioC] Data consisting of several groups, persons, and cells

[Seungwoo Hwang]

Dear all,

I have a question on how to analyze the following Affymetrix gene expression data. 
- There are three groups (each representing a haplotype).
- There are several persons in each group (3 persons in group 1, 5 persons in group 2, and 4 persons in group 3).
- There are 4 biological replicates in each person. The experimenter took a cell from a person and cultured it four times, in consideration of possible variation that may take place when cell is cultured. The experimenter expects the within-individual variation (i.e., variation among the four replicates) is small.

The ultimate goal of the analysis is to reveal transcriptional differences among the three groups.

The problem is that I am not sure how to take care of the four replicate data. If there were no four replicate data, I would have just done one-factor ANOVA.

For each person, should I just take an average of normalized intensities from four replicate data and proceed with one-factor ANOVA? If I can do that, what kind of analysis should I do first in order to check the within-individual variation is small enough to take an average? Or is there an alternative way other than taking an average?

Thanks,

Seungwoo
------------------------------------
Seungwoo Hwang, Ph.D.
Senior Research Scientist
Korean Bioinformation Center