Association tests using discordant sibs
For some diseases it may be difficult to obtain parents for a TDT sample. It may sometimes be possible to deduce parental alleles from the siblings of cases and then perform the TDT, but subtle biases can arise with this procedure.
Instead, one can use an unaffected sibling of each case as a control. The power is enhanced by choosing a sibling who is as genetically distinct from the case as possible, and this does not introduce any bias.
In the first sibship, the case is AC and the control is BD. In the second, the case is AA and the control is BC
