pair-wise LD was calculated by using standard formula [8] that are implemented in the EMLD program. We randomly selected a single sib pair from each family to ensure independence of the sib pairs. We then studied each family either including or excluding all parental genotype data. Multipoint and single-point linkage analyses of the affected sib-pair data were carried out using ALLEGRO [9]. For model-free multipoint linkage analyses, we used a Kong and Cox exponential model [10] and the score function of Spairs [11]. For the parametric linkage analyses, we assumed a simple dominant disease model with 100% penetrance in carriers and 0% penetrance in non-carriers, and we incorporated a heterogeneity parameter [12], thus allowing some but not all families to be linked.
