|
Abstract:
|
Proposed in 1993, the Transmission Disequilibrium Test (TDT) by Spielman et al. has been used widely to analyze genetic information from trios (affected child and parents). In 2002 (discussion: 2004), we identified a flaw in the TDT (and many of its extensions). In short, genetical "independence" of alleles transmitted to the same child does not imply statistical "independence" of observations. Stratification by parental mating types increases power and reduces other shortcomings of the TDT. More importantly, new insights open the door for new tests targeting more specific questions. Using data on risk factors for cardiovascular diseases, we present versions particularly suited for diseases with low and high dominance and extension to uni- and multivariate ordinal and semiquantitative phenotypes. Asympt/exact versions of the tests are available from mustat.rockefeller.edu.
|
