Browsing by Author "Kaksonen, Sanna"

Energy requirement of dairy cows can be higher than energy intake during early lactation. When energy balance is negative cows mobilize energy from body tissue. This increases the concentration of non-esterified fatty acids (NEFA) in the plasma. NEFA uptake to mammary gland is directly related to the plasma NEFA concentration. Therefore, negative energy balance and the plasma NEFA concentration may be predicted from fatty acid concentration of the milk determined from mid-infrared (MIR) spectral data. The objective of this study was to analyze the relationship between milk fatty acid profile, single fatty acid concentrations and plasma NEFA concentration. In addition, the aim was to test, if the MIR spectrum results can be used in predicting plasma NEFA concentration and subsequently negative energy balance in dairy cows. This study was a part of the Nordic Feed Efficiency -project. Data were collected in three research farms from primiparous Nordic Red dairy cows between September 2013 and August 2016. Second lactation data were collected from the same cows when possible. There were 610 records from 143 primiparous cows, and 199 records from 49 cows in second lactation. Data were analyzed with Mixed- procedure of the SAS software. The association between the plasma NEFA concentration and predictor traits were studied with correlation analysis. Plasma NEFA concentration was predicted by using regression analysis. Regression models were examined separately for both parities. Among milk fatty acids C18:1c9 had the strongest correlation to plasma NEFA concentration in both lactations. Plasma NEFA concentration predicted based on the model including C18:1c9 concentration and days in milk (DIM) had a strong correlation with observed plasma NEFA concentration being 0.84 in first lactation and 0.89 in second lactation. Models including concentrations of two milk fatty acids and DIM had higher coefficients of determination than models including one fatty acid concentration and DIM. Best two fatty acid models included concentrations of C14:0 and C18:1c9 and DIM as predictor variables. The coefficients of determination for these models were 0.54 in first lactation and 0.69 in second lactation. Prediction error was smaller in second lactation models models (RMSE 0.08 mmol/L) than in first lactation models (RMSE 0.16 mmol/L). Adding more prediction variables did not improve the models. Prediction models in this research underestimated the highest plasma NEFA concentrations, because in some cases plasma NEFA concentration can reflect stress in addition to negative energy balance. These models are useful and reliable in predicting existence and severity of negative energy balance of dairy cows with similar feeding as in this study.