Freebayes haplotype calling

Question

[k.itokawa]

Hi all,

I am using freebayes to call variants in diploid organism. So far, my understanding of this program's behavior is as follow:

Assume that there are two SNPs at consecutive nucleotide positions, and there are multiple haplotypes such like:

• haplotype_1: XXXAGXXX
• haplotype_2: XXXTGXXX
• haplotype_3: XXXACXXX

case 1

If haplotype_1 was used as a reference, the variations at the two nucleotide positions are called as two separated SNPs in the output VCF.

 ref   1234   .   A   T ...
 ref   1235   .   G   C ...

case 2

On the other hand, if haplotype_2 was used as a reference, the two variant sites are treated as single MNP in the output VCF.

 ref   1234   .   TG   AC,AG ...

This is, basically, very understandable behavior. However, I sometime like to obtain output always like case 2 even in case 1 because this simplifies some downstream analysis such as investigating amino-acid changes and its genotype.

That is, I want VCF line like below for case 1.

 ref   1234   .   AG   AC,TG ...

Is there any option in freebayes to obtain output like this? or I appreciate if you know any other tool or methodology to get lid of this problem?

Thanks

Answer 1

[colindaven]

Not sure if I completely understand your question but I have always found multi-sample genotyping with Freebayes to be very good for comparative genotyping. I know your question pertains to different reference sequences, but this is a far more tricky question and will probably only be solved by graph genomes, for which there are a dearth of good approaches and implementations to date.

There, both positions 1234 and 1235 will be called, and indicate which of the genotypes were called for each sample at each position (towards the right hand end of the VCF line).