Inconsistency between biological replicates

Nicholas_ · 04-06-2011, 03:36 AM

Hello,

So i am relatively new to the whole chip-seq analysis.

I have 2 different biological replicates for the same protein and i am using 2 controls as well.

For my peak analysis i am using the new version of Cis Genome, with a cutoff of 5, a bin size of 33 and no poisson.

And here is where i have the problem. I have uploaded my initial before peak calling BED files in the Genome Browser in order to visualise them and see how they worked. My BED files look very similar before peak calling but when i perform peak calling the consistency is very very low...around 50%

What could go wrong? Is it my settings? I am trying to figure it out for weeks but i can't. Maybe i am too stupid! Please help!!! Any suggestions would be apprecited!

Thnx
Nick

tir_al · 04-06-2011, 04:18 AM

Have you tried looking at the duplication rate of your libraries? Although the distributions might look the same in the browser, the high duplication rate in one sample can really mess up the peak calling.

You can check the duplication rates very easily using the excellent FastQC app.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Biological replicates with cuffdiff, cummeRbund	turnersd	Bioinformatics	15	11-19-2012 06:59 AM
using Cuffdiff with biological replicates	Jane M	RNA Sequencing	0	09-01-2011 01:42 AM
overdispersion and biological replicates	shilez	Bioinformatics	3	08-29-2011 08:43 AM
overdispersion and biological replicates	shilez	RNA Sequencing	0	08-25-2011 07:10 PM
cuffdiff with biological replicates	PFS	Bioinformatics	1	06-14-2011 07:51 PM

04-06-2011, 03:36 AM	#1
Nicholas_ Junior Member Location: Athens Join Date: Apr 2011 Posts: 2	Inconsistency between biological replicates Hello, So i am relatively new to the whole chip-seq analysis. I have 2 different biological replicates for the same protein and i am using 2 controls as well. For my peak analysis i am using the new version of Cis Genome, with a cutoff of 5, a bin size of 33 and no poisson. And here is where i have the problem. I have uploaded my initial before peak calling BED files in the Genome Browser in order to visualise them and see how they worked. My BED files look very similar before peak calling but when i perform peak calling the consistency is very very low...around 50% What could go wrong? Is it my settings? I am trying to figure it out for weeks but i can't. Maybe i am too stupid! Please help!!! Any suggestions would be apprecited! Thnx Nick

04-06-2011, 04:18 AM	#2
tir_al Member Location: Croatia Join Date: Sep 2010 Posts: 22	Have you tried looking at the duplication rate of your libraries? Although the distributions might look the same in the browser, the high duplication rate in one sample can really mess up the peak calling. You can check the duplication rates very easily using the excellent FastQC app.