Modeling and genetic parameters of lactation curves in Holstein bovine breed in Colombia

Objective. Evaluate different mathematical models applied to lactation curves, determine which of them represent the best fit and estimate genetic parameters for the characteristics derived from the best model. Materials and methods. Information from 426797 monthly milk control records was used from 49108 Holstein cows coming from 470 dairy herds located in Cundinamarca, Antioquia, Valle, Boyaca, Narino and Cauca. Nonlinear mathematical models proposed by Wood, Brody, Wilmink and Papajcsik and Bodero were evaluated. After selecting the model that best adjusted the lactation curves, genetic parameters were estimated for the characteristics (beta(o)), peak milk production (y(max)), time to peak (t(pico)) and total production at 305 days (P305) where the fixed area effects, delivery and contemporary group were used. Results. Wood's model showed high values of PAIC and PBIC and high values of R-2. In relation to beta(o), y(max), t(pico) y P305 characteristics, derived from the Wood model, the average value was 16.64 +/- 6.34 lt, 27.39 lt +/- 6.85, 44.55 +/- 13.19 days and 6212 +/- 1690 lt, respectively. beta(o) y t(pico) characteristics presented a low heritability (0.02) and the y(max) and P305 showed a low to medium heritability (0.15 and 0.17, respectively). Conclusions. The model that best fit lactation curves in Holstein bovines is the Wood model. The average heritability and high genetic correlation between e ymax and P305 indicate that it is possible to include them in breeding programs.