Intersecting vcf files with bedtools

Question

[Maj]

Hello,

I want to intersect two vcf files :

File 1 :

CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  7.bam

Scaffold_1        223     .       G       T       10.8428 . DP=7;DPR=5,2;VDB=0.18;SGB=-0.453602;RPB=0.8;MQB=1;MQSB=1;BQB=0.9;MQ0F=0;ICB=1;HOB=0.5;AC=1;AN=2;DP4=5,0,0,2;MQ=60 GT:PL:DP:DV:SP:DP4:DPR   0/1:44,0,110:7:2:13:5,0,0,2:5,2

Scaffold_1        435     .       T       C       20.145  .       DP=3;DPR=1,2;VDB=0.1;SGB=-0.453602;RPB=1;MQB=1;MQSB=1;BQB=1;MQ0F=0;ICB=1;HOB=0.5;AC=1;AN=2;DP4=0,1,2,0;MQ=60    GT:PL:DP:DV:SP:DP4:DPR     0/1:53,0,30:3:2:0:0,1,2,0:1,2

File 2 : 

CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  S07

scaffold_11       651457  929888  G       T       .       PASS    Ty=SNP;Rk=1;UL=.;UR=.;CL=.;CR=.;Genome=C;Sd=1   GT:DP:PL:AD:HQ  0/0:4:4,16,84:4,0:70,0

scaffold_1188     2186    894758  T       G       .       PASS    Ty=SNP;Rk=1;UL=.;UR=.;CL=.;CR=.;Genome=T;Sd=1   GT:DP:PL:AD:HQ  1/1:4:84,16,4:0,4:0,70

This is the error :

bedtools intersect -a 7.variant_snp.vcf -b 7.variant_snp2.vcf | head

ERROR: file 7.variant_snp.vcf has non positional records, which are only valid for the groupBy tool.

I don't know how to deal with this issue. Does anyone have experience of this kind of error?

Thanks in advance

[steve]

might be worth simply parsing the "VCF" files into a BED format? Should be able to just include the POS field twice to stand in for the Start & Stop positions.

Answer 1

[Manvendra Singh]

First may be you get rid of headers As per I remember, I always used one vcf and one bed file to intersect by bedtools

May be you use vcftools to intersect two vcf files

It has nice command as "vcf-isec" Creates intersections and complements of two or more VCF files Nice documentation is here

[Maj]

Thank you :) I tried it but now i have a new error "The command "tabix" not found, please add it to your PATH" I'm trying to fix it. Thank you again

Answer 2

[harold.smith.tarheel]

The examples you show do not match the VCF format specs. The header line (beginning with 'CHROM') should be preceded by a hash '#'. Otherwise, 'POS' is read as a value that should be an integer but is not (i.e., non-positional).

[harold.smith.tarheel]

Also, bedtools 'intersect' does not output headers by default, and using the output for subsequent 'intersect' operations can produce errors. You can include by using the '-header' flag.