intersect (actually: filter) a gtf file with coordinates from a bed-file

[dietmar13]

-- SOLVED -- bedtools intersect

hello,

does someone know a script which filters a gtf-file with coordinates from a bed file.

i have a bed-file with regions and want filter out all features from a gtf file which overlap (completely or largely) with this regions. is there a script or program? bedtools intersect does not work?

i know, an awk-script would not be very difficult, but why re-invent the wheel...

thank you,

dietmar

[kmcarr]

bedtools intersect does not work?

Why do you say this? Seems like bedtools intersect with the '-v' option is exactly what you are looking for.

[dietmar13]

@kmcarr

thank you - you are right. i was misled by all the examples where only bed and bam files were used for bedtools intersect examples...

dietmar

[vishal.rossi]

Hi,

I am comparing 2 different files. 1st file has 113 entries and the 2nd one has 88 entries.
I use the following command to get the differences
intersectBed -v -a 1.bed -b 2.bed or
intersectBed -v -wa -wb -a 1.bed -b 2.bed

But it shows that only 3 entries don't match in both the cases which is false.
Does anyone has the idea why?

Thanks

[syfo]

I am comparing 2 different files. [...] I use the following command to get the differences [...] only 3 entries don't match

What are you looking for exactly?
IntersectBed with the "-v" parameter will show you the intervals from "1.bed" that have nothing in common with the ones in "2.bed". Entries from "2.bed" are not supposed to be reported.
Also, one common nucleotide is enough by default to define an overlap between two intervals. For a more stringent criteria you might want to consider "-f" and "-r".

[AlexReynolds]

Another option is BEDOPS bedops, which does set operations on BED data, and BEDOPS gtf2bed, which does a lossless conversion of GTF data into BED format, which can be used with BEDOPS tools.

Let's assume that your regions-of-interest are in a file called myRegions.bed and your GTF-formatted annotations are in a file called myAnnotations.gtf.

First, we sort myRegions.bed:

$ sort-bed myRegions.bed > mySortedRegions.bed

Next, we convert the annotations to BED format:

$ gtf2bed < myAnnotations.gtf > myAnnotations.bed

Finally, we apply a --not-element-of set operation to show elements of the annotations file which do not overlap mySortedRegions.bed, if there is one or more bases of overlap (i.e., any overlap at all):

$ bedops --not-element-of -1 myAnnotations.bed mySortedRegions.bed > myAnswer.bed

As the gtf2bed conversion step was lossless, it is easy to convert myAnswer.bed back to GTF:

$ awk '{print $1"\t"$7"\t"$8"\t"($2+1)"\t"$3"\t"$5"\t"$6"\t"$9"\t"(substr($0, index($0,$10)))}' myAnswer.bed > myAnswer.gtf

[eieneg]

To intersect coordinates

Just use GFF-Intersector

https://github.com/PriceJon/GFF_Intersector

it can intersect GFF files with multiple other coordinates!! have you got R? if so just 2 commands and you don't have to worry about the visualisation issue