Chip-Seq Analysis

[suludana]

Hi

I am analyzing my first data of Chip-Seq (with the antibody H4K16Ac) using Partek.

In the initial step of detecting peaks I obtained a lot of peaks for some of the samples (100.000 – 200.000) and of course, then, I obtained large lists of associated genes. How can I do to filter these data? The control (not IP) is not enough to reduce substantially these values.

Can I filter by the number of reads in each peak? In this case which would be the optimal value?

Thanks!

[RNA-seq]

You can filter based on
1. sample ID--only keep peaks from ChIP sample
2. total number of reads or normalized read count--low numbers should be filtered out. The filter criteria is based on your data.
3. binomial P-value sample Vs control--low P-value peaks should be kept.

[suludana]

Hi,
thanks for your post.
The true is that I am completly lost in this analysis.
I have 5 samples and 2 controls (no IP samples). 3 of the samples goes with one of the controls and the other 2 goes with the second no IP sample.
I obtained the first list of peaks, for the 7 samples. After I applied the second filter you refer (normalized read count).
After that I want to filter using the controls, so I go to "Create a list of enriched regions". I select the sample I want to filter (for example the sample 3, with 490 peaks) and I select "Specify new criteria", there I select Sample ID vs S1 (S1 is the control of S3). Than I obtained a list of 630 peaks.
I don't understand that. I filtered and I obtained more peaks?
What mistake am I committing?
Thank you very much for your help.