CLIP/RIP-seq peak finding

[Mali Salmon]

A question to all users that run CLIP/RIP-seq and then try to find peaks in their data.
Since I couldn't find any specific tool for finding peaks in RNA-based data, I have decided to use the classical ChIP-seq tools (the logic is the same I think).
I align reads to the genome (I don't really care about junction reads), and then run one of the ChIP-seq tools such as MACS.
I would like to hear about the parameters people use when they run ChIP-seq tools on CLIP/RIP-seq data. For example many of the ChIP-seq tools use the "genome size" for random expectation calculations, and different genome sizes give different output.
Do you use the transcriptome size or genome size when trying to find significant peaks?
Any other issues that should be taken care of?
Thanks
Mali

[ETHANol]

I am by no means the expert on this topic but I have been reading about it a little here and there the last few days. From what I have read, using a ChIP-seq data analysis pipeline results in the identification of many false positives. I haven't seen any well documented workflows but you may want to check out this thread:

http://seqanswers.com/forums/showthread.php?t=14775

[Mali Salmon]

Thanks ETHANol for your reply. Can you explain more why ChIP-seq pipeline is not suitable for RNA-based data, what is the reason for the high false-positive rate?
I have RNA-IP sample and RNA-Input and would like to find enriched peaks. The idea is basically the same as with ChIP-seq.
Mali

[ETHANol]

I wish I could give you an answer but at this point I would probably just be polluting the forum with half-truths. Hopefully, someone with more experience on the topic will chime in.

[monique1]

We are interested in finding RNAs that interact with specific regions of chromatin. Do you know if there is a technique out there for this? Thanks.

[mudshark]

a) either you have a protein specifically associating with these loci, then use an IP followed by RNASeq

b) otherwise the only regions in chromatin that can be purified so far (i know many people that tried others and failed) are telomers
http://www.cell.com/abstract/S0092-8674(08)01520-1

c) or do you work in yeast? then there might be other options.