Bedtools and VCF files

[morangal]

i'm trying to run intersectBed with a VCF file (produced by complete genomics, version vcf4.1) and a Bed file of the RefSeq genes(downloaded from UCSC table browser), but i don't get any results,
i do know about certains variants included in my list that should appear in the intersection.
i changed the first colomn of the VCF file from "1,2,3...X,Y" to "chr1. chr2... chrX, chrY" (even though i did not see any need for that in the documentation).
my command is :
intersectBed -wa -a RefSeqGenes.bed -b myfile.vcf

strangely, when i run this with a minifile that contains only a few lines (with variants that i expect to be intersected) it works. but when i run it, with a larger vcf file (that contains the whole genome, or even a single chromosome's variants) i get nothing.

any idea why??

your help will be very appreciated :-)))
Moran

[aggp11]

Hi,

I think intersectBed expects the format of chromosome name to be the same in both the input files, i.e. if you have "chr1, chr2 etc.." in the vcf file, make sure you have the chromosome names in the same format in the bed file as well.

[morangal]

thanks! i've noticed that :-))
but the big mystery, is why it does not work with the complete vcf file (after th echromosomes were prefixed with "chr"), but do work with a sample of it....

[aggp11]

Could you post the first few lines of the big vcf file (that doesn't work),the smaller file(that works) and the bed file?

[morangal]

the mini-file that works contains the VCF header + this 1 lines:
chr16 2121925 . G T . . NS=2;SS=Germline;CGA_FI=7249|NM_000548.3|TSC2|CDS|MISSENSE&7249|NM_001077183.1|TSC2|CDS|MISSENSE&7249|NM_001114382.1|TSC2|CDS|MISSENSE;CGA_PFAM=PFAM|PF03542|Tuberin GT:PS:FT:GQ:HQ:CGA_EHQP:AD:CGA_RDP:CGA_ODP:CGA_OAD:CGA_ORDP 1/0:.:PASS:472:472,499:499,499:61:29,32:32:60:30,30:30 1/0:.:PASS:571:571,571:571,571:60:30,30:30:61:29,32:32
(btw, it works with and without the header lines)

the full original file, that does not work, contains the same header and these first 3 lines:
chr1 1 . N <CGA_NOCALL> . . END=10000 GT ./. ./.
chr1 10001 . T <CGA_CNVWIN> . . NS=2;CGA_WINEND=12000 GT:CGA_GP:CGA_CP:CGA_PS:CGA_CT:CGA_TS:CGA_CL:CGA_LS:CGA_SCL:CGA_SLS:CGA_LAF:CGA_LLAF:CGA_ULAF .:0.341:.:0:.:0:0.997:326:.:.:.:.:. .:0.553:2:20:=:20:1.003:86:1.003:86:0.29:0.17:0.5
chr1 10001 . T <CGA_NOCALL> . . END=10472
GT ./. ./.


it should be mentioned that in order to add "chr" to the first colomn of the original VCF, i had to remove the header lines, add "chr" to the rest (using awk '{print "chr" $0}' originalfile.vcf ) and then re-concatenate the header lines. i don't see how these text manipulations should affect the program, but maybe i missing something here ?


the Bed file to be intersected contains these first lines:
chr1 67000041 67000051 NM_032291_cds_0_0_chr1_67000042_f 0 +
chr1 67091529 67091593 NM_032291_cds_1_0_chr1_67091530_f 0 +
chr1 67098752 67098777 NM_032291_cds_2_0_chr1_67098753_f


Thanks :-)))