Searching causal structure among calving traits in first-parity Holstein cattles of Iran

In this research the causal structure among calving traits of 29950 first-parity Holstein cattles of Iran including calving difficulty (CD), birth weight of calves (BW) and gestation length (GL) was revealed applying data collected by Iranian Animal Breeding in 131 herds from 1995 to 2004 by Inductive Causation (IC) searching algorithm. Significant structural coefficients were found for causal effects of BW on CD (0.060±0.002) and of GL on CD (0.007±0.002). Furthermore, the causal effect of GL on BW was significant (0.219±0.005). Considering the revealed causal structure, standard and recursive multivariate models were compared applying deviance Information criterion (DIC) and predictive ability of models in terms of two measures including mean square of error and correlation between observed and predicted values. The obtained results revealed the causal effect of BW and GL on CD and the plausibility of recursive multivariate model over standard multivariate one. Therefore, considering the causal structure among calving traits is of crucial importance.