A crucial resource for bio-economic optimization of pig production is an accurate growth model. The question is which model suits best to model the commercial growth traject of the pre-maturely slaughtered finishing pig. The objective of this study was twofold: (1) to compare the accuracy of several published growth models; and (2) to provide up-to-date parameters for accurate and applicable growth curves for high-lean pigs of four different sexes (boar, barrow, GnRH-vaccinated boar and gilt).Data from four trials of which two with individually housed and two with commercially group housed pigs were used. Animals were kept ad libitum on a multiphase diet. Body weight and feed intake were recorded from 9 weeks of age until slaughter. The serial
body weight data were fitted to two dynamic models (Gompertz and Bridges), describing growth as a function of age and one static model (Monomolecular), describing growth as function of cumulative feed intake, using PROC NLMIXED in SAS. Residual sum of squares (RSS) and root mean squares (RMS) were calculated to evaluate model goodness-of-fit. RSS were used to test significant differences in goodness-of-fit between models. Residual plots were evaluated to investigate under or over prediction. The static monomolecular model showed significant (P<0.05) better goodness-of-fit for the four sexes compared to the dynamic models, except for the barrows in one trial. The Gompertz and Bridges models did not significantly differ in goodness-of-fit. Residual plots showed no systematic over or under prediction under 125 kg. However, because of selection bias, due to slaughtering the animals at a fixed weight instead of age, over prediction occurred beyond 125 kg. Application of the current parameters for weight predictions beyond 125 kg is therefore not recommended.
|Titel||Book of Abstracts Annual Meeting of the European Federation of Animal Science|
|Uitgeverij||Wageningen Academic Publishers|
|Artikel nummer||Theatre 9|
|ISBN van geprinte versie||978-90-8686-284-9|
|ISBN van elektronische versie||978-90-8686-830-8|
|Status||Gepubliceerd - 29-aug-2016|