Filter intergenic regions in a .bed file

Question

[paulklein05]

Hello everyone,

I am looking for a way to filter out intra-genic regions (thus exons + introns), to only keep inter-genic regions from a .bed file. My first guess was to use the intersect command from bedtools, by comparing my bed file with the human reference genome

bedtools intersect -a my_file.bed -b hg38_genome.bed  -v > out.bed

I extracted the hg38_genome.bed file from UCSC webpage (with default parameters) but I am unsure of its content: does this file give exons associated chromosome locations and coordinates ? Or does it give entire genes (exons + introns) boundaries, which is what I am looking for ?

I am very new to bioinformatics so do not hesitate to suggest a different approach.

Best,
Paul

Answer 1

[kevin.stachelek]

you can select a variety of bedfiles from UCSC according to your regions of interest. If you select the default "Whole Gene" you will get exons+introns.

[paulklein05]

Hello @kevin, thanks for you reply !

I think that is what I am looking for, but when I check the first few lines of the generated file, I do not understand why I get overlapping regions regarding the coordinates:

Am I missing something ?

[kevin.stachelek]

Overlapping features are usually different transcripts of the same gene. You can get a visual confirmation of this if you load the downloaded bed file as a custom track in the genome browser. You can also see the transcript id annotated in the fourth column of your screenshot.

Answer 2

[ATpoint]

Use bedtools complement. Intergenic is the complement of the genome with everything that is annotated as "gene" (and its subsets like exon, intron, UTRs...) so just complement the genome with a GTF file, e.g. from GENCODE filtered for type gene and you have your intergenic regions

[paulklein05]

I did not know this bedtools command, thanks for sharing :)