RPKM normalization between samples

[bt7]

I have 8 RNA-seq datasets comprising 4 biological replicates for 2 treatments (control and condition). From what I understand, the samples were run in a 4-plex format such that each Illumina run contained one biological rep from each of the 2 treatments. So, none of the biological reps were run together. The reads have already been mapped to an EST database and I now have RPKM values for all of the genes in each of the 8 samples.

1. RPKM values are normalized for comparison of genes within a sample, but can RPKM values be compared between samples (treatment vs. control, in my case) within an Illumina run? And can RPKM values be compared between samples (biological replicates, in my case) in separate Illumina runs? If not, what is the best method of normalization?

2. How do I go about determining which genes are differentially expressed between the control and treatment, using the RPKM values?

Any info for someone new to RNA-seq analysis would be much appreciated.

Thanks!

[d17]

Take a look at the DESeq or edgeR packages. Both of these will take raw counts (not RPKM) and perform statistical tests for differential expression between control and condition.