A critical aspect of the simulation concerns the ascertainment/diagnostic schemes used to construct the datasets. The Aipotu families were selected when at least two offspring were present who had either Phenotype 1 (P1), Phenotype 2 (P2), or Phenotype 3 (P3) (representing three different phenotypes). Other family members were also counted as affected if they had either P1, P2, or P3. Similarly, the Karangar families were chosen for at least two affected offspring with P2 or P3, and other family members were also counted as affected if they had P2 or P3. The Danacaa families were also collected based on the presence of P1 in two offspring and only P1 was counted as affected in the other family members. These ascertainment/diagnostic schemes were chosen to simulate the variability of ascertainment/diagnosis in actual psychiatric disease. The NYC study chose families based on the presence of at least 4 members affected with P1 or P2 or P3 and at least one of each phenotype was required to be present in each family. This ascertainment fact was deliberately not noted in the data description.
