Abstract #T222

# T222
Regression analysis of the relationship between milk de novo synthesized fatty acids, trans-10 C18:1, and milk fat percent using treatment means from the literature.
Cesar I. Matamoros*1, R. Klopp1, AR Clarke1, Kevin J. Harvatine1, 1The Pennsylvania State University, University Park, PA.

Biohydrogenation-induced milk fat depression (BH-MFD) is caused by bioactive conjugated linoleic acid (CLA) isomers produced during altered rumen biohydrogenation. These CLA isomers appear in low concentration in milk during BH-MFD, but trans-10 C18:1 also increases during BH-MFD and provides a biomarker of alternate ruminal fermentation and the production of bioactive fatty acids (FA) that inhibit milk fat synthesis. The concentration of de novo synthesized FA (<16 carbons) are also decreased during BH-MFD. The objective of this study was to characterize the relationship between milk fat de novo FA and trans-10 C18:1 and milk fat concentration allowing prediction of BH-MFD based on milk FA profile. A database was constructed from peer-reviewed papers (n = 928 treatments) and treatment means of non-published papers from Penn State (n = 149 treatments). The inclusion criteria were that the papers reported milk yield, milk fat concentration, and milk FA profile in post-peak Holstein cows. Regression analysis was conducted using JMP Pro 13.0. The linear, quadratic, and a single and 2 pool exponential fit were tested. The 2-pool exponential model allowed modeling of a responsive and non-responsive portion of milk fat as BH-MFD results in a maximal 50% reduction in milk fat. Additionally, a comparison between methylation methods (2-step acid, 1-step acid, and sodium methoxide methylation) was also tested because they differ in the literature. The concentration of even chain saturated FA between 6 and 14 carbons and C14:1 best fit to milk fat percent using a linear model (MSE = 0.3; RMSE = 0.548; R2 = 0.168). Milk trans-10 18:1 best fit to milk fat percent using a 2-pool exponential model (MSE = 0.149; RMSE = 0.385]. The exponential fit did not differ between data from papers using acid and base methylation procedures or between the literature and only data from our laboratory. In conclusion, BH-MFD was best predicted by milk trans-10 C18:1 using a 2-pool exponential model allowing prediction of expected milk fat concentration from a milk fatty acid analysis.

Key Words: trans-10 C18:1, milk fat, milk fat depression