vcfTools depth filter problem

[JackieBadger]

Hey guys,

I am new to vcf and having difficulty filtering my SNPs in vcfTools.

I have a vcf file (fileformat=VCFv4.1) of SNPs found in a group of individuals of a species (pooled as one individual for the time being).


sample:
Spp1_cDNA-30 211 . G A 122 . DP=15;VDB=0.0017;AF1=1;AC1=2;DP4=0,0,0,12;MQ=60;FQ=-63 PL 155,36,0

I want to apply some simple filters e.g. depth.

However the output file generated gives me depths of 0 at all SNPs despite the vcf giving sensible DP values.

Similarly, mean depth stat calculation returns 0 for the entire vcf.

What am I doing wrong here guys?

Any help will be much appreciated.

Thanks,

J

[KatsenPlatz]

Can you post the command you ran? Based on the one line of VCF you've posted, there seem to be 12 reverse reads with non-ref allele. You can filter your input using the DP4 field.

Description of DP4 from http://samtools.sourceforge.net/mpileup.shtml
"DP4: Number of 1) forward ref alleles; 2) reverse ref; 3) forward non-ref; 4) reverse non-ref alleles, used in variant calling. Sum can be smaller than DP because low-quality bases are not counted."

[JackieBadger]

OK thanks for pointing that out.
However, Im not sure the significance of what that actually means?
"12 reverse reads" is fine as far as I figure.
And overall depth is 15 (including low quality bases I presume?)

Here is my command:

vcftools --vcf file1.vcf --minDP

Testing some other functions, vcftools wont actually calulate any for me.

Eg.

vcftools --vcf SkateGenotypes.vcf --counts

I have attached a sample of my vcf file.

Thanks alot!
J

CHROM POS N_ALLELES N_CHR {ALLELE:COUNT}
Skate_cDNA-0 48 2 0 G:0 A:0
Skate_cDNA-0 49 2 0 G:0 A:0
Skate_cDNA-0 176 2 0 A:0 G:0


Addtional:

When I run vcf-stats on the compressed file I get the following error:

Use of qw(...) as parentheses is deprecated at /usr/share/perl5/VcfStats.pm line 428.
Use of qw(...) as parentheses is deprecated at /usr/share/perl5/VcfStats.pm line 433.
Could not parse the fileformat version string [VCFIDX], assuming VCFv4.1
Broken VCF header, no column names?
at /usr/share/perl5/Vcf.pm line 171, <$__ANONIO__> line 1.
Vcf::throw('VcfStats=HASH(0x1549010)', 'Broken VCF header, no column names?') called at /usr/share/perl5/Vcf.pm line 845
VcfReader::_read_column_names('VcfStats=HASH(0x1549010)') called at /usr/share/perl5/Vcf.pm line 589
VcfReader:arse_header('VcfStats=HASH(0x1549010)') called at /usr/share/perl5/VcfStats.pm line 53
VcfStats:arse_header('VcfStats=HASH(0x1549010)') called at /usr/bin/vcf-stats line 137
main::vcf_stats('HASH(0x11ac480)') called at /usr/bin/vcf-stats line 12

[JackieBadger]

No one has any insight here?

Could it be that the chromosome names (which are actually reference coltigs) are stings that do not include < > ?

Should I just remove the wording, to have just numbered chromosome IDs?

If so whats the best way to do this?

I did it in excel and with the following perl script but it then leave the output.vcf unreadable by vcftools:

perl -pe 's/^Little_Skate_cDNA-//' SkateGenotypes.vcf > SkateGenotypesError.vcf

Any help would be greatly appreciated.

Cheers,

J