The objective of this simulation study was to evaluate the suitability of risk ratio (lambda R) as indicator of the number of genotyped animals necessary for mapping genes for binary traits, such as mastitis in cattle. Data were simulated for different prevalences, genetic models and parameters of these, in families with five affected halfsibs in average. The data were analyzed with methodology of "affected sib pairs". The number of genotyped halfsib families needed increase with the prevalence. As risk ratio is lower with higher prevalences, lambda R is an adequate indicator for comparing power for traits with different prevalences. The number of animals needed was substantially higher for dominant as for recessive traits. Incomplete penetrance had little influence on the power, but phenocopies increase the number of needed animals. The presence of a second locus decreased the power especially in genetic models with locus heterogeneity. It can be concluded that lambda R is a useful indicator to determine the number of animals necessary in mapping experiments in halfsib families.
binary traits; cattle; experimental design; genetic mapping; risk ratio
