Association of total-mixed-ration chemical composition with milk, fat, and protein yield lactation curves at the individual level

The objective of this study was to examine the effect of the chemical composition of a total mixed ration (TMR) tested quarterly from March 2006 through December 2008 for milk, fat, and protein yield curves for 27 herds in Ragusa, Sicily. Before this study, standard yield curves were generated on data from 241,153 test-day records of 9,809 animals from 42 herds in Ragusa province collected from 1995 to 2008. A random regression sire-maternal grandsire model was used to develop variance components for yields. The model included parity, age at calving, year at calving, and stage of pregnancy as fixed effects. Random effects were herd × test date, sire and maternal grandsire additive genetic effect, and permanent environmental effect modeled using third-order Legendre polynomials. Model fitting was carried out using ASReml. Subsequently, the model with estimated variance components was used to examine the influence of TMR crude protein, soluble N, acid detergent lignin, neutral detergent fiber, acid detergent fiber, starch, and ash on milk, fat, and protein yield curves. The data set contained 46,531 test-day milk yield records from 3,554 cows in the 27 herds recorded during the study period. Initially, an analysis was performed using one dietary component (one-component analysis) within each model as a fixed effect associated with the test-day record closest to the months the TMR was sampled within each herd. An interaction was included with the nutrient component and days in milk. The effect of the TMR chemical component(s) was modeled using a ninth-order Legendre polynomial. The conditional Wald F-statistic for the fixed effects revealed significant effects for acid detergent fiber, neutral detergent fiber, crude protein, starch, and their interactions with days in milk on milk, fat, and protein yield. On the basis of these results, a multicomponent analysis was performed in which crude protein, neutral detergent fiber, and starch were simultaneously included in the model with days in milk interactions. Although both analyses revealed that diet composition influenced production responses depending on lactation stage, the multiple-component analysis showed more pronounced effects of starch and neutral detergent fiber relative to crude protein for all traits throughout lactation.