search

biocore-ntnu / epic2

Ultraperformant reimplementation of SICER
https://doi.org/10.1093/bioinformatics/btz232
MIT License
55 stars 9 forks source link

use epic2 for ATAC-seq? #22

Open YichaoOU opened 5 years ago

YichaoOU commented 5 years ago

Since epic2 is optimized for diffuse peaks (correct me if it is not), is it OK to run epic2 for narrow peaks, like regular TF ChIP-seq? I'm trying to find differential peaks for ATAC-seq, as I understand, ATAC-seq peaks, could be a mixture of narrow and broad peaks. Can I still use epic2?

Thanks, Yichao

YichaoOU commented 5 years ago

Even though ATAC-seq is paired-end, epic2-df doesn't report any problem. Epic-bw does.

endrebak commented 5 years ago

I would not use epic2 if you want to see stuff at a resolution lower than 50bp. So it might work for ATAC-seq, I dunno if you expect to find extremely short regions not covered by a nucleosome in your data? You could also use epic2 and macs2 together and take the unions of the enriched regions

On Friday, June 7, 2019, Yichao Li notifications@github.com wrote:

Even though ATAC-seq is paired-end, epic2-df doesn't report any problem. Epic-bw does.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/biocore-ntnu/epic2/issues/22?email_source=notifications&email_token=AEHURUSUY5PHFMWKSXAIH7DPZGAHLA5CNFSM4HVJ2W42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXEH6NY#issuecomment-499679031, or mute the thread https://github.com/notifications/unsubscribe-auth/AEHURUQIQKJXEP6T2F345MDPZGAHLANCNFSM4HVJ2W4Q .

YichaoOU commented 5 years ago

BTW, when I ran epic2-df for ATAC-seq, and I tried different parameters, it always gave me a diff region of 2kb. Is there any parameter for the length of diff peaks?

Thanks, Yichao

endrebak commented 5 years ago

The lengths should differ naturally. The default setting is to use bins of 200 (but multiple enriched bins near each other will be clustered). Are you sure the size of the regions was not 200 instead of 2000? And perhaps you should try a bin size of 50?

Is the data you use public? Could you upload it to a file sharing service? Would love to explore this with you

On Tuesday, June 11, 2019, Yichao Li notifications@github.com wrote:

BTW, when I ran epic2-df for ATAC-seq, and I tried different parameters, it always gave me a diff region of 2kb. Is there any parameter for the length of diff peaks?

Thanks, Yichao

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/biocore-ntnu/epic2/issues/22?email_source=notifications&email_token=AEHURUUQURUZSHVJZORDLL3PZ7VWXA5CNFSM4HVJ2W42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXOCYWY#issuecomment-500968539, or mute the thread https://github.com/notifications/unsubscribe-auth/AEHURUXITQFKMSMCSLQNSGLPZ7VWXANCNFSM4HVJ2W4Q .

YichaoOU commented 5 years ago

That data is not public. I will try some public data and see if I have the same problem.

Thanks!