how to run ROSE to find super enhnacers-

Question

[Chava]

Hello,

I'm pretty new to bioinformatics, and I'm having troubles when trying to run the ROSE package to find SEs. On the website, it's not clear which files I should bring and what specific file format is required. If someone has experience running it, I would greatly appreciate any help or comments.

Thank you very much for your assistance!

Answer 1

[Trivas]

This seems pretty straight forward to me - http://younglab.wi.mit.edu/super_enhancer_code.html

You need:

1. .bam file of your mapped reads
2. .gff file of known enhancers

These need to be UCSC/RefSeq notation, not Ensembl. This is easy enough to convert between.

[jared.andrews07]

I'll add that the reference data for ROSE hasn't been updated since like 2012, and it really shows. I've found it worth the effort to munge a more recent annotation into the refflat format and slap it into the annotation directory/alter source code as necessary.

Additionally, it has some hidden parameters that aren't accessible without altering source code that lead to unclear behaviors during loci stitching (e.g. a region that spans 3 TSS elements won't be stitched together). This results in certain loci looking like super enhancers without being identified as such. Again, not much you can do without dorking with the source code, but just be aware of it.

[Chava]

How can I obtain the annotation file? I have a peak file of enhancers; can I simply use it as a GFF file?