H3k36me3 data: how to handle replicates?

[mrfox]

hi all,
I wonder if there is a decent way to handle replicates when studying H3K36Me3 chip seq data. Usually the peaks for this kind of histone modification is broad.

Assume I have 3 normal samples and 3 tumor samples to compare. Most of the software tools are designed to compare 1 normal vs 1 tumor. So what do you usually do to after identifying the differentially peaks between each pair of normal and tumor? In another word, how do you merge these differential peaks to give a h3k36me3 profile between the two groups?

Thanks.

[mudshark]

i would not work with peaks but rather with read summarization on gene bodies (which are the span of H3K36me3 enrichment). Then you can use DESeq or EdgeR/limma to do your testing for differential enrichment.