Abstract #T51

# T51
Statistical power of the Bayesian analysis for simulated transmission ratio distortion in cattle.
S. Id-Lahoucine1, A. Cánovas2, C. Jaton*2,3, F. Miglior4,2, F. S. Schenkel2, J. P. Chesnais3, S. Miller5, M. Sargolzaei2,3, J. F. Medrano6, J. Casellas1, 1Departament de Ciència Animal i dels Aliments, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain, 2Centre for Genetic Improvement of Livestock, Department of Animal Biosciences, University of Guelph, Guelph, ON, Canada, 3Semex Alliance, Guelph, ON, Canada, 4Canadian Dairy Network, Guelph, ON, Canada, 5Angus Genetics Inc, St. Joseph, MO, 6Department of Animal Sciences, University of California-Davis, Davis, CA.

Studies on transmission ratio distortion (TRD), defined as the deviation from the expected Mendelian inheritance of alleles from heterozygous parents, remain rare, especially in livestock species. This phenomenon can be caused by various biological mechanisms affecting gametes, embryos, fetuses, or even postnatal offspring. Therefore, the study of TRD can lead to the identification of genetic factors involved in fertility and reproduction traits. Specific Bayesian models have been recently developed for the analysis of TRD, accommodating a wide range of population structures. This parameterization can differentiate between sire- and dam-specific TRD, or merge both effects into an overall TRD estimate. The statistical relevance of TRD can be tested by a Bayes factor (BF), a ratio of probabilities between the models with and without TRD effects. The results obtained on simulated cattle population data sets showed that statistical power increased with the population size, the mean between intra-sex proportion of relevant heterozygous parents, and the magnitude of the TRD itself (Table 1). The sex ratio did not directly influence the statistical power, but it influenced the number of sires and, consequently, the available genetic variability and the proportion of heterozygous sires. The rate of false positives was 0.29% when BF ≥ 10 (strong evidence), and 0.02% when BF ≥ 100 (decisive evidence). Pearson correlation coefficients between simulated and estimated TRD were greater than 0.97 when BF ≥ 10. The Bayesian analyses showed great statistical power to detect TRD with high accuracy in the simulated cattle population. Table 1. Statistical power to detect overall TRD or sire-specific TRD with Bayes factor ≥100 (250 replicates with one offspring per dam and 1:25 sex ratio)
Population sizeTRD1Sire-specific TRD2
0.10.30.50.10.30.5
250.010.180.540.000.231.00
1250.030.860.990.050.991.00
2500.140.981.000.201.001.00
5000.461.001.000.651.001.00
1Random minor allele frequency ≥0.05. 2All sires were heterozygous; the same results were observed for dam-specific TRD (heterozygous dams).

Key Words: transmission ratio distortion, heterozygous parents, Bayes factor