Long peaks

khb · 02-14-2011, 01:30 AM

Hi
I've done a chip-seq experiment, and used the program MACS. When I look at the peaks in UCSC genome browser, the peaks are very long, see pdf file for example from mitochondrial chromosome. Is this peak likely too bee a real peak?
We believe that it will not bind direct to a DNA but likely as a part of a complex.
Thanks for all replies

csoong · 02-14-2011, 08:20 AM

Never had any chip-seq experience. But are the peaks directly related to coverage? since chip-seq only selects a portion of the genome having high coverage (hence seeing higher peaks) is expected, right?

By the way, what were you expecting? And please forgive me if I totally misunderstood your question.

danielr · 02-14-2011, 12:23 PM

If it's on the mitochondrial chromosome, assume it's not real. You pick up things there because it doesn't have the same copy number as the nuclear chromosomes.

epigen · 02-15-2011, 06:04 AM

The length of your peaks depends on the value for the --bw parameter you give MACS. It should be the same as the fragment size. The higher the value, the longer the peaks.

mudshark · 02-16-2011, 01:36 AM

looks weird.

what was your experimental layout, did you use a control/input sample?

is your chip target supposed to enter the mitochondria, is it a mitochondrial protein, is your result plausible?

did you try any other peak caller, SISSR, Quest?

mapper · 02-16-2011, 02:36 AM

I have also seen very large peaks with MACS ...you can either use peak splitter program...or you can try sissr which will give very sharp peaks....

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Peaks Analysis	Joe Petrosino	Epigenetics	1	02-01-2012 01:09 PM
Comparing input peaks to IP peaks	biznatch	Epigenetics	2	09-24-2011 12:38 PM
No peaks	chipmonk	Genomic Resequencing	4	09-07-2011 04:44 AM
Bioanalyzer Peaks	Wilkes	Sample Prep / Library Generation	1	09-06-2011 12:32 PM
MACS peaks	VeenaV	Bioinformatics	2	11-10-2010 09:46 AM

02-14-2011, 08:20 AM	#2
csoong Member Location: Connecticut Join Date: Jun 2009 Posts: 74	Never had any chip-seq experience. But are the peaks directly related to coverage? since chip-seq only selects a portion of the genome having high coverage (hence seeing higher peaks) is expected, right? By the way, what were you expecting? And please forgive me if I totally misunderstood your question.

02-14-2011, 12:23 PM	#3
danielr Member Location: Sweden Join Date: Sep 2009 Posts: 11	If it's on the mitochondrial chromosome, assume it's not real. You pick up things there because it doesn't have the same copy number as the nuclear chromosomes.

02-15-2011, 06:04 AM	#4
epigen Senior Member Location: Germany Join Date: May 2010 Posts: 101	The length of your peaks depends on the value for the --bw parameter you give MACS. It should be the same as the fragment size. The higher the value, the longer the peaks.

02-16-2011, 01:36 AM	#5
mudshark Senior Member Location: Munich Join Date: Jan 2009 Posts: 138	looks weird. what was your experimental layout, did you use a control/input sample? is your chip target supposed to enter the mitochondria, is it a mitochondrial protein, is your result plausible? did you try any other peak caller, SISSR, Quest?

02-16-2011, 02:36 AM	#6
mapper Member Location: India Join Date: Nov 2010 Posts: 17	I have also seen very large peaks with MACS ...you can either use peak splitter program...or you can try sissr which will give very sharp peaks....