Effect of the data family structure, tank replication and the statistical model, on the estimation of genetic parameters for body weight at 28 days of age in the Pacific white shrimp (Penaeus (Litopenaeus) vannamei Boone, 1931)

Real and simulated data sets were analysed to estimate the effect of data structure (i.e. the number of dams mated to one sire and the number of tanks per dam) and the model used for analysis on the bias of genetic parameter estimates of the square root of 28-day body weight (BW28) in Penaeus vannamei. Estimated parameters used were 0.13 for the heritability (h2); 0.16 for the proportion of the variance of the maternal and common environmental effects (c2); 0.15 for the proportion of the tank effect variance (t2) and 0.56 for the proportion of the environmental variance (e2). Total variance for the BW28 was 2.36 mg. Simulated data involved a total of 9000 sires and 1 444 000 offspring. Exclusion of maternal and common environmental effects inflated estimates of heritability both in real and simulated data. Exclusion of tank effects from the model inflated the estimates of heritability in real, but not in simulated data, indicating the presence of non-random effects affecting individual tanks. We conclude that replication of dams per sire is necessary for obtaining unbiased estimates of heritability. Use of replicated tanks may help in reducing overestimation of heritability if tank and genetic effects are confounded. © 2012 John Wiley & Sons Ltd.