0
$\begingroup$

This might be a very basic question for many here. With the basic understanding of inheritance, eventhough there is a possibility of multiple genotypes due to multi alleles, the resulting genotype can only have two alleles(paternal and maternal) that way after variant calling, an allele at one position can be homozygous or heterozygous. So there can be max of two alleles, but why do we see multiple alleles at a given position in VCF. I am trying to understand the science behind this. Please help out. Thanks!

chr5    127640782   .   AG  A,AA    .   .   .   GT:AD:DP    1/2:0,28,409:437
$\endgroup$
  • $\begingroup$ Can you please clarify what you mean by multiple alleles at a given position in VCF.? Maybe, you could show an extract of a VCF file to show us what you mean. Are you surprised that in the population there can be more than 2 alleles at a given site? $\endgroup$ – Remi.b Sep 28 '18 at 16:34
2
$\begingroup$

In your example the 4th column contains the reference allel on the given position. It is not said, that your sample has this. The 5th column contains all alternative alleles found on the given position.

Which alleles are present in your sample is given in the last column by 1/2. This means you have one allel with the first value in the 5th column and one allele with the second value given in the 5th column. You see, that your sample has exactly alleles, but none of them is the reference. This would be indicated by a 0.

$\endgroup$
0
$\begingroup$

A vcf can contain multiple columns representing allele data for multiple samples. I also see no reason why you could not represent a tetraploid sample's data in a vcf. One might also have a mix of organisms i one sampe, like a mixed population of bacteria. If you could get your SNP caller to call a tri-allelic SNP, the vcf format can handle it.

$\endgroup$
0
$\begingroup$

The other possibility I can imagine is that the format is designed to accommodate heterogeneity between individual cells in a sample, and inevitable sequencing errors.

If you sequence a sample from an organism which (mostly) has two alleles at a locus (e.g. A and G), there may still be individual cells which have a C or T at that locus, and there may also be sequencing errors introducing a C or T in reads which were actually an A or G. The VCF format would then still be able to represent the low-probability calls, in addition to the higher-probability ones.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.