Investigation of the inheritance of birth defects in swine by complex segregation analysis

The inheritance of birth defects was investigated by analyzing a large data set with more than 30000 litters. The data were recorded as 0/1 traits in connection with the evaluation of an anomaly index for AI boats in the German state of Bavaria. The overall incidence of defects was 2.07%, and the frequencies of the individual traits varied from 0.04 % (hermaphrodism) to 0.69% (splay leg). Monogenic, digenic, polygenic, and mixed models of inheritance were compared using Maximum Likelihood methodology. Integrations necessary to calculate likelihoods were performed numerically (Gauss-Hermite), and a derivative-free algorithm (Powell method) was used for maximization. The most likely mode of inheritance was determined based on likelihood ratio tests. The inclusion of common litter environments improved the model fits remarkably. Omitting c² effects were shown to lead to false detection of major genes. In most cases, the polygenic model fitted the data best. Heritabilities estimated on the underlying scale ranged from 0.10 to 0.25, and c² effects accounted for 20-40% of the liability. Major gene effects of about one standard deviation of liability were found for scrotal hernia in the offspring of DL boars and for splayleg in the litters of DL and PI boats. Allele frequencies of the defect-causing gene were 0.02 for scrotal hernia and 0.25 for splayleg, respectively. The results were discussed in the light of previous investigations, and further analyses were suggested.