Know more

MOTS-CLES : Cattle; genetic determinism; cheese-making ability; Montbeliarde bovine breed

Mid-IR spectroscopy (MIR), which has been identified as a precise tool for identifying the fine composition of milk, was a good candidate for predicting the initial cheese-making characteristics, coagulation aptitude, acidification and yield. The From’MIR (2015-2018) project, coordinated by CEL25-90 and financed by Casdar, the Cniel and Urfac, was aimed at predicting the cheese-making aptitude of milk using MIR spectra, for genetic improvement and counseling in breeding. The storage of MIR spectra since 2011 and the development of cattle genotypes for genomic selection are two exceptionnal resources for a large scale study of these traits.

Prediction equations compiled from 420 reference cheese tests (24 parameters measured using two technologies: soft cheese and hard pressed cheese) with good to excellent precision were obtained using the PLS method for six coagulation parameters and three yield parameters; acidification aptitude was poorly predicted. These equations were then applied to more than 6 million spectra from more than 400 000 Montbeliard cows from Franch Comté, paving the way to large-scale genetic analyses.

An estimation of genetic parameters showed that these traits are heritable (between 0.37 and 0.48 on the scale for each monthly measurement) and that they have a stable determinism during lactation and between lactations. On the genetic level, all traits for yield and coagulation are favorably correlated amongst themselves, including between cheese-making technologies. Coagulation is strongly associated with the casein content and some minerals whereas yield is more associated with fat content. Cheese-making aptitude is not unfavorably associated with any currently selected traits and has experienced a rather favorable evolution over the last 10 years in this breed. An association analysis led to a device with 20 000 cows with phenotypes and genotypes imputed on the whole genome scale, using the population sequenced through the "1000 bovine genome" consortium as a reference. Fifty-nine QTL were detected as being significant on the genome scale, globally explaining between 12 and 30 % of the phenotypes for cheese trait variability. Most of the QTL known to affect milk composition also influence cheese aptitude and new QTL were detected, with causal mutation candidates. A network analysis showed a network of 736 genes that interact amongst themselves along with approximately 20 metabolic pathways, in particular for potassium and phosphates. This analysis identified several regulating genes including PPARA, ASXL3 and bta-mir-200c, strongly implicated in the regulation of these networks.

A pilot genomic evaluation indicates that the precision obtained is high, allowing for a very efficient selection on these cheese-making traits. 

This large-scale study shows the power of genetic approaches applied to a large scale of data issued from high-throughput phenotypings and genotypings. This progress will provide tools for both selection and counseling in breeding, so that cheese-making qualities may be controlled at the milk production scale. A genomic evaluation of these traits for cheese-making ability is currently being developed. Similar studies are underway with other regions and French breeds.