Pedigree structure with half-sibs

[Andrey Yurchenko]

Dear Pasi,
I have several families with 1 mother and 2-3 fathers each (reptiles). Is it possible to use Lep-MAP3 with this type of pedigree? How can I specify the parentage relationship in my case? I tried to assign the fathers to the families with mothers but received the error : too many parents.
Thank you!

[Pasi Rastas]

Dear Andrey,

Thank you for your question.

And yes, you can analyze such half-sib data. You have to split the data to full-sib families (unique mother-father pairs). Then you can provide the "halfSibs=1" parameter to ParentCall2 and analyze the output data as of consisting of full-sib families.

Do you know the father for each offspring?

Cheers,
Pasi

[Andrey Yurchenko]

I do not have genoypes of the fathers, but know the likely relationship inside the families (from COLONY testing). I am thinking about using the ParentalCall to impute the fathers genotypes when it is possible. What do you think?

[Pasi Rastas]

Yes, I think COLONY will find find full-sib families. And ParentCall2 will impute the missing parents if the families are sufficiently large (maybe over 8 individuals). Grandparents would also help here if you have data on them.

Cheers,
Pasi

[Andrey Yurchenko]

Dear Pasi,
Another couple of questions for my case:
1) If I make unique mother - father pairs for half-sibs, should I or should not introduce a new family ID for these mother-fathers and halfsibs? Or should I leave the previous family ID for these familes?
For example, a family 1 consists of mother1, father1, fahter2, offspring1, offspring2.
Should I separate them into 2 familes with family1:
mother1, father 1, offspring1
and family 2:
mother 1, father2, ofspring2
Os it right logic?

2) Is it possible to leave the fathers genotypes uncalled for familes with small number of ofspring and use only mothers genotypes?

Thank you very much!
Andrey

[Pasi Rastas]

Dear Andrey,

Sorry for taking some time to answer.

1) The family ID should be unique for each mother, father combination, like "FATHER_ID|MOTHER_ID".

2) It is possible to have missing genotypes, but both parents have to be resolved in order to use any SNPs.

For example, if you know that the mother is heterozygous A/B, then the other parent is quite uncertain with only a few offspring. And the other parent is crucial for getting linkage information.

Grandparents and half-sibs can help to resolve the parents.

Cheers,
Pasi