DESeq vs. PARTEK ANOVA

[aggp11]

I have RNA -seq data from the HiSeq with 2 normal and 2 diseased replicates (a total of 4 samples). I tried both DESeq and PARTEK GS to obtain differentially expressed genes. However, I see very different results.

I saw the following discussion:

http://seqanswers.com/forums/showthread.php?t=8500

and have read about both the tools and understand that both are workign differently. However, I am still not sure which one would be better to understand the expression changes.

For Partek, I used a 2-way anova (condition & replicates) on the RPKM values.

For DESeq I provide counts using HtSeq in the suggested order and perform the analysis following the documentation.

I would really appreciate any suggestions on this.

[dietmar13]

statistics -

my question is, why should a new technology override statistical principles:

two replicates, especially if unpaired, are never enough to find statistical significant differences...