TDT analysis when both parents are missing

Suppose that the A allele of a multiallelic marker is common and that genotyping sibships containing affected subjects yields the following results:

In both cases one can deduce that the genotypes of the untyped parents must be AB and AC.

In the second sibship, if the affected sib has genotype AA this will count as two occasions when the A allele has been transmitted to an affected offspring, while if the affected sib is BC then this will count twice as the A allele not being transmitted. Hence there is no systematic bias.

However, in the first sibship when the affected sib is AA this will count in favour of allele A being transmitted to affected offspring, but if the affected sib is AB or AC this will count as one A allele being transmitted and one not being transmitted and hence will be neutral. Overall, the first sibship produces a bias in favour of the the A allele being transmitted to affected offspring.