[en] Methods to measure LTF do not permit the use of this data in genetic. The aim of this study was to create an equation to predict LTF (pLTF) in milk by MIR. Calibration with Partial Least Squares on 69 samples showed a cross-validation coefficient of determination equal to 75.01%. So, pLTF for 7,690 samples collected from April 2005 to April 2006 were predicted. Milk yields, %fat, %protein and somatic cells count (SCS) were added. 39,441 records on 1,910 cows were analyzed by a multi- trait mixed model including as effects: herd*test date*lactation number, lactation stage*lactation number, age*lactation number, regression on breed fractions, permanent environments within and across lactations and animal effect. Variances and standard errors were estimated by AI-REML. Number and stage of lactation influence pLTF. A difference across 7 breeds was observed. Heritability for pLTF was 19.70%. So, the genetic variability for LIF exists. The phenotypic correlation observed between SCS and pLTF was equal to 0.26. As SCS increases in presence of mastitis, this observation seems to indicate that pLTF could be an indicator of mastitis.
