many more ChIP-seq peaks called in 1 condition than the other

Question

[Pei]

Hello everyone!

We do ChIP-seq across 2 conditions each with 2 replicates.

The number of called narrow peaks (MACS2 2.2.7.1, FDR < 0.1) are:

cond_1-1: 8,915

cond_1-2: 10,099

cond_2-1: 1,446

cond_2-2: 801

At first, I suspected that there are many small peaks detected in condition_1.

However, I observed the opposite as the size of called narrow peaks tend to be larger in condition_1.

(1) ~24% of peaks are longer than 600 bps in condition_1 while that number is ~13% for condition_2.

(2)the median peak size is ~400 bps in condition_1 but only ~330 bps in condition_2.

So, usually what could be the reason to call many more peaks in one sample than the other? the library sizes are similar across samples.

Thanks in advance!

Pei

Answer 1

[rfran010]

Either its biological, where your factor's binding is greatly decreased in condition two. Or it's technical where the IP did not work as well for samples in condition 2. If all samples were not prepared in parallel, I would be more wary of the latter.

I think it would be important to check other QC aspects of the libraries, like fragment sizes and replicate correlation. Also always good to visualize the signal on a genome browser.

[Pei]

Got it. Thanks!