Multivariate analysis of mixed inheritance model of performance traits in layers using Gibbs sampling

The Gibbs sampling under a multitrait animal model was applied to detect th e single gene affecting chicken performance traits and their pleiotropic ac tions as well as co estimate the heritability and correlations for these tr aits. A total of 14 823 individuals of a Rhode Island Red line (RIR) and 18 653 individuals of a Rhode Island White line (RIW) from six generations un der long-term selection were recorded. Five performance traits were studied : initial egg production (IEP; until 38th week), egg production (until 54th week), egg weighs at 33rd week (EW), age at first egg (AFE), and body weig ht at 20th week (BW). An analysis was based on the estimated marginal densi ties of the frequencies, additive and dominance effects amd variances and c ovariances (for single gene) as well as additive generic and residual varia nces and respective covariances (for polygenes). An inference concerning th e mixed inheritance model is performed by visualising the marginal posterio r densities of major gene variance separately for all traits. The pleiotrop ic of single locus is expressed as the single gene correlation coefficient. It shows contributions genes to BW (10 and 15% of total variance for A22 a nd K44, respectively (and EW of K44 total variance). Moreover, a small posi tive pleiotropic single locus effect in line K44 registered. The polygenic heritability estimates obtained were low, except for EW and BW lines. Gener ally, the correlation estimates were in agreement with results reported in literature.