Define exact peak borders
1
0
Entering edit mode
4.3 years ago
ATpoint 55k

Hello,

I would like to ask if there is an implementation to define peak borders. To make it more clear, see this screenshot. I have some reference region from another assay (blue bar on top). Within this region, I check the signal of a second assay (blue peak profile). The question is now if there is a simple solution to define the peak borders (red bar) automatically. MACS does this nicely with the --nomodel option, but as it checkes the input data for significances etc. it does not output all the regions I am interested in. Is there a standalone solution for this task?

Peak Calling MACS • 819 views
ADD COMMENT
0
Entering edit mode

To clarify: you want the red bar to be the same length as the blue bar or as the blue peak region?

What about peak center +/- total fragment length?

More importantly....what is the purpose of defining the borders? what are you trying to achieve by doing this?

ADD REPLY
0
Entering edit mode

The read bar is what I printed in by hand, but I would like to get it automatically, based on the shown peak profile. I have a list of mutations. Many of them are within the blue-bar region, but outside of the actual peak signal (so the read bar). In order to exclude these, I would like to have exact borders, in order to exclude these passengers.

ADD REPLY
0
Entering edit mode

Why was this thread deleted?

ADD REPLY
0
Entering edit mode

Don't know?! Not on purpose

ADD REPLY
0
Entering edit mode

Odd, well, I restored it...

ADD REPLY
0
Entering edit mode
4.3 years ago
Charles Plessy ★ 2.7k

There are many alternative peak callers, but among them perhaps you can have a look at paraclu.

Since it outputs nested clusters, you can define by yourself a threshold that corresponds to granularity that suits the most your analyses. Also, you can run it in a way that all the data will be part of at least one cluster (unlike in the example above, where a small peak is left out).

ADD COMMENT
0
Entering edit mode

I will check it out, thank you. Cluster within cluster seems to be pretty much what I need.

ADD REPLY

Login before adding your answer.

Traffic: 2484 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