Informativeness of twin-nuclear family and nuclear family designs for segregation analysis

This brief note examines the precision of segregation parameter estimates from both twin-nuclear family and conventional nuclear family designs. The program MENDEL was used to derive expected frequencies of different family patterns of affectation (with fixed sibship size) under various single gene models, and then to estimate the standard errors (and information) associated with gene frequency and penetrance values at given sample sizes. As might be expected, a 2- to 5-fold increase in relative efficiency was found for the same family size if families included an MZ twin pair among their offspring. The methods used allow convenient calculation of expected informativeness of a given study design.