How to specify how the consensus sequence should be built?

[larahurban]

Hi,

I am sorry for asking a quite simple question, but I can unfortunately not find any documentation about this, and I am also not sure if I am reading the raw code of the function correctly:

I am using FastaVariant to create a consensus sequence per sample; all exemplary code uses the default parameter of "het/hom = True", but I was wondering what they mean? Also, what happens in the case of a missing genotype call, and is there a way to specify that the alternative allele should be preferred in the case of a heterozygous variant (I am investigating a dominant phenotype)?

Thank you very much!

Best wishes, Lara

[mdshw5]

Hi @larahurban. I've just come back from vacation so please excuse the delay in answering. I really never did document the FastaVariant class very well, but here's the way it works:

• Only SNVs are supported, so MNVs and indels are ignored.
• Only called sites are used, so any position that does not have a called genotype in the sample of interest will be using the reference allele.
• The het and hom parameters determine whether to use the alternate allele (true) for each type genotype call.
• There is also a call_filter parameter that accepts expressions that evaluate against the FORMAT fields. The example I include in the README excludes homozygous sites using call_filter, but you could use it to exclude calls based on GQ or DP.

[larahurban]

Thank you very much, Matt!