RIP-seq analysis

[Rachelly]

Hi all,

I want to analyze data from a RIP-seq experiment (an IP of a protein that binds to mRNAs).
Can I use MACS for this (not forgetting to change the bandwidth, slocal and llocal accordingly)? Are there any problems or biases for using it for this assay (RNA instead of DNA, longer peaks, etc.)?

Is there another analysis tool for RIP-seq ?

Thanks,
Rachelly.

[Rachelly]

After more thinking - strand data is not preserved in an RNA protocol (unless a special process is done). So using MACS is not possible.
Another idea is to compare normalized counts for genome features (similar to RNAseq).

Any other thoughts?

[tboothby]

Hi Rachelly,

I am interested in doing some RIP-seq experiments. Do you have any advice from your experiences?

I know that RIP is supposed to produce a lot of background, did you encounter this problem. What sort of controls did you use?

Cheers,
T

[Dario1984]

You will also need to run an RNA-seq library and use that to account for high expressed genes, because the same genes usually show up at the top of every RIP-seq experiment, no matter which antibody you use. In other words, I'm confident you will get lots of reads for GAPDH.

There is a new tool that is going to be published soon, from what the author tells me. It's abstract A50.

[Rachelly]

Hi tboothby,

What I did in that project, is use an RNA-Seq analysis to compare the results of two experiments. I compared both FPKM values and raw counts of the two samples.

I think it was enough for that specific project, since we wanted to compare binding of whole mRNAs transcripts (rather than short binding sites, as it is in regular Chip-Seq).

I haven't had anymore experience with RIP-Seq since.
I'd be happy to hear other people's thoughts.

Cheers,
Rachelly.