Gene expression + RNAPol II ChIPSeq

[hunakos]

Hello Everybody,

I'm new in this field, so I might have a very naive / simple question:

Can I study gene expression level changes by comparing two or more RNA POL II Chipseq datasets? For example: treated vs non-treated?
(I know that the RNA-seq would be the best tool, but i have onlw ChIPseq datasets...)

If yes, could anybody suggest me a tool to compare datasets, and to find candidate genes that shows alteration in expression levels?

It would be a great help.

Regards

[mudshark]

I would not recommend to use the RNA Pol ChiPSeq data for many reasons:

a) the precision of ChIPSeq for quantification of extended feature enrichments (as expected for RNA Pol II) is very bad given the current data analysis procedures

b) you need very deep sequencing to get good RNA Pol signals on the body of weakly expressed genes (I hope you kind of normalize to input as the input reads are a function of gene expression as well, i.e. more expression, more control reads).

c) the Pol II binding - RNA level correlation depends on which isoform of Pol (S2ph, S5ph, all Pol) you measure and where you actually calculate the enrichment (promoter, body). Promoter peaks might not well indicate the true RNA production (stalled Polymerases)

d) you will need a serious amount of replicates, do you have those?

I strongly doubt that you will get a robust expression estimate. Maybe you will be able to detect dramatic expression changes, but I guess most of the differentially expressed genes you will miss. if you just want some candidates you might, however, try.

why not do an - watch out - expression microarray
seriously, they work, they are highly validated, analysis procedures are calibrated using reference data sets and if you do not need alternative splicing and stuff they are a true option.