P-value for enrichment of ChIP-Seq peaks within a genomic annotation

Question

[User 7754]

Hi,

I am trying to find a p-value for enrichment of ChIP-Seq peaks from a bed file within a chromatin state from the UCSC stored in another bed file. I overlapped the two bed files using BedTools and now I have these values:

The number of ChiP-Seq peaks that overlap a chromatin state (more than one peak can overlap a chromatin state, but a specific peak belongs to only one chromatin state)= 10450

The total number of ChiP-Seq peaks = 100561

The total coverage of a chromatin state = 204652

How can I get a p-value of enrichment for the CHiP-Seq peak within the chromatin state?

Thank you

[Fidel]

This biostar answer is quite comprehensive. Personally, I had used the bedtools fisher, you'll need your list of peaks, your list of chromatin states and a file with the chromosome sizes.

Answer 1

[dally]

Have you thought about the phyper function in R?

[User 7754]

Hi and thanks for the reply!

I have, but I am not sure what I would use for the totals including the non-overlapping peaks and annotations, would this be correct?

phyper(10450, 100561, 100561+204652, 204652-100561)

[dally]

Probability of gene list overlap

See if this helps you, I'd give you specifics but I'm not comfortable enough with my knowledge of this function to do that.

[User 7754]

Thanks this helps, but I am still confused about what our k would be, if we consider the number of ChiP-Seq peaks:

q=10450-1, m=100561, n=204652, k=204652?

Also would it not be important to consider that more than one ChIP peak can overlap each chromatin state?