ChIP-seq problem
1
0
Entering edit mode
17 months ago

Hello,

I have just uploaded two bw files to IGV, they are from the same experiment, only the second one was sequenced more. When I load them more I notice that the file sequenced more has peaks that are lower, why is that in your opinion?

N.B. they were processed in the same way so the steps from fastq files to bw files are the same but the result is very different.

I attach the photo to make you understand.

What could be the explanation?

Thanks for helping me

enter image description here

fastq Chip-seq • 1.1k views
ADD COMMENT
0
Entering edit mode

Hello.
What is the source of the BigWig files? Were they normalised? Do they contain duplicated reads/fragments? It could be the deeply sequenced sample contains more background (non-specific mapping).

ADD REPLY
0
Entering edit mode

I created bw files from bam files and filtered them so as to remove duplicates from both

ADD REPLY
0
Entering edit mode

I would use something like Deeptools bamCompare to generate bigWig files that are normalised by sequencing depth: https://deeptools.readthedocs.io/en/develop/content/tools/bamCompare.html

If the peaks are still shorter in the deeper sequenced sample then there is higher background (non-specific signal).

ADD REPLY
0
Entering edit mode

"from the same experiment"...are they two aliquots of sequencing from one library? Or are they replicates? (different IP samples but done within the same experiment). Replicates from ChIP seq (even within the same experiment) can vary quite a bit, such that sequencing depth would be beside the point.

ADD REPLY
0
Entering edit mode

They are two different replicates fro ChIP seq but in your opinion is normal?

ADD REPLY
0
Entering edit mode
17 months ago
seidel 11k

Independent IP trials from an experiment can behave differently. Signal to noise ratio in a data set is much more important than sequencing depth. If you look at figure 5 of Landt et al. (ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia Genome Res 2012), you'll see 4 replicate IPs of the same factor, 2 replicates each from two different experiments, and you'll see that the peaks are different heights, and each data set has unique characteristics in terms of IP enrichment. That paper also has a variety of recommendations for how to examine enrichment and compare replicates. Things like cross correlation plots, Fraction of Reads in Peaks (FrIP), and IDR analysis.

ADD COMMENT

Login before adding your answer.

Traffic: 1899 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6