A study was conducted to find early predictors of the Booroola gene in several generations of a crossbred sheep population. Merino carriers of the Booroola gene were mated with Texel sheep to improve prolificacy of the latter. Ovulation rate at 8 months of age, litter size at 1 and 2 years of age and FSH and inhibin levels at 3, 4, 5 and 6 weeks of age were determined in about 700 females.
Gibbs sampling was applied for inference in a mixed inheritance model. Estimates for the gene effect in heterozygous females were +1·5 corpora lutea and +1·3 lambs at 2 years of age. The gene effect on litter size at 1 year was small. The only significant major gene effect for hormone levels found was for lnINH4 (–0·66).
A number of hormone levels and combinations of hormone levels appeared to be useful predictors of carrier status of individual animals. In comparison with a situation where only parents' genotype is known, posterior probabilities for non-carriers were on average increased from 50 to over 95% when FSH levels were used. However, the combined posterior probabilities of carriers and non-carriers increased only up to 67%. So in general, classification with Gibbs sampling resulted in too few animals being identified as carrier. The sum of FSH levels at 3, 4, 5 and 6 weeks of age is proposed as a predictor of presence of the Booroola gene in an animal. Multivariate analysis of mixed inheritance models could help to find more effective combinations.