1000 Genomes VCF Files

[ashkot]

Hi all,
I am conducting whole genome analysis on 1K Genomes data and I need some help understanding the VCF file

First, my VCF file only contains 1/0 and 1/1 genotypes. Why i am not seeing 0/0. Also, does it mean that if a genomic location (Chr:location) is not in the VCF then there is no variation detection at that location? (which would be the same as 0/0)

Second, if i have to find out the genotype at a particular location how does one do that. We have some variants which are not in dbSNP, how can I view them in a VCF file?

I am using SAMTools and VCFUtils all under standard alrorithm settings.

Thanks,
Ashwin

[laura]

First there being no variation at a particular site is not the same as an individual being 0/0. The site might not be variable at all or the data you have may not contain evidence of variation so be very careful about the assumptions you make about the absence of data

While it would seem unlikely that one individual was only homozygous non ref or heterozygous it isn't totally impossible especially if you are only considering a small range

Can you actually give us your command lines and which individual you are considering as this will make it a lot easier for us to try and explain what is going on

[ashkot]

Hi there,
I am following the following workflow:

Sample ID: HG00096

First, indexing the bam file using samtool index
Second, sorting the bam file using samtools sort eg2.bam eg2.sorted
Third, generating a bcf file using samtools mpileup -uf reference_genome_file eg2.sorted.bam | bcftools view -bvcg - > eg2.raw.bcf

The reference human genome file is the one obtained from 1kGenomes, Human_37_*.fasta

Lastly, converting my bcf file to vcf using bcftools view *.bcf | perl /usr/share/samtools/vcfutils.pl varFilter -D100 > *.vcf

I am interested in knowing genotypes for rs numbered dbSNPs and other variants for which I have HG37 genomic coordinates.

When i follow the above workflow, i am not getting 0/0 genotypes and at the same time some of the genomic locations also do not show up. This is why my first question was if absence of variation should be considered a 0/0 genotype?

Thanks,
Ashwin

[laura]

Is there a particular reason you are generating the snps yourself rather than using our release data?

ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/

If there are no homozygous ref snps across the whole genome I would run your process again to make sure you didn't make an error then I would report the problem to the samtools-help mailing list to see if they know what is going on as if it is a problem it is either with the way you ran the caller or with the results the caller produced not the vcf file

[laura]

You might want to consider retitling this thread or asking a new question about the best use of samtools mpileup as this is not my area of expertise

[ashkot]

the reason why i am calling SNPs is because i am trying to setup an analysis pipeline using the 1KGenomes data. I want to make sure we have got all the parameters correct but the 0/0 issue still remains unresolved.

[laura]

Ask a question about samtools and snp calling then rather than 1000 genomes and I suspect you will get a lot more help

I would recommend looking at the genotypes we do have for the individual you are trying with as that will be a good indiciator of how accurate what you are doing is

[ashkot]

Already posted, titled Samtools mpileup, I will also check the genotypes as you suggested.

[vyellapa]

Im looking for a link of genotypes vcf of the latest 1000 genome release and the corresponding reference. I can see one at the link below but it seems to the older release (629 individuals and vcf 4.0). Can I know where I can find the same file in the newer release.

ftp://ftp.1000genomes.ebi.ac.uk/vol1...notypes.vcf.gz

Thank you,
Teja