Finding annotation of any region in human genome

Question

[hasani.iut6]

Hi all, I had some sequences with 25 lengths. I aligned them to the human genome and found the hits.

First I was interested to finding which of the hits are belong to any specific gene. So some the aligned sequences was belong to some specific genes and some others not. Now I'm interested to find annotation of the hits that not aligned to a gene region. Actually I want to know that is that hits belong to any specific regions such as: gene promoters, repeats, etc ?

Thanks all.

Mansoor

Answer 1

[GouthamAtla]

You can intersect them with the GTF file and see if they are part of a gene.

You can intersect them with the chromatin states from epigenome roadmap data to find out if its a regulatory element.

You can also upload into ucsc or washU browsers to check if they overlap with any regulatory elements from ENCODE or ROADMAP data.

If they are sequences relevant to any particular tissue, you can try to find out the regulatory regions of that particular tissue and intersect.

May be there are some tools to do it.

[hasani.iut6]

I've used GTF files for finding gene regions and I'm interested for finding annotation of hits that not happened in gene regions. Number of hits are a lots and I can't to use ucsc and browse them one by one. but I will considering chromatin states from epigenome roadmap so to find can I find something useful or not? thank you for your answering.

Answer 2

[Alex Reynolds]

Via BEDOPS:

$ bedops --element-of 1 <(bam2bed < hits.bam) <(gtf2bed < annotations.gtf) > hits_overlapping_annotations.bed
$ bedops --not-element-of 1 <(bam2bed < hits.bam) <(gtf2bed < annotations.gtf) > hits_not_overlapping_any_annotations.bed

See the documentation for bedops, bam2bed, and gtf2bed for more information.