The first year of the project (2015) has focused on method development. At AU-FOOD, we have tested LC-ESI/MS as a possible method for quantification of one of the important high-value proteins, osteopontin (OPN), using OPN standards purified at AFI and at AU-MBG, Laboratory for Protein Chemistry. These tests proved to be very challenging, as OPN is very heterogeneous with regard to post translational modifications, and furthermore consists of different molecular chain lengths. We have further worked with another one of the proteins we are going to quantify, β-casein (β-CN). We have purified β-CN from milk to use as a standard, and have tested the dynamic range of the protein in the analysis. Further, a LC-MS tripleQ equipment has been invested in, and we have worked on setting up a so-called MRM method for quantification of β-CN  and α-lactalbumin (α-LA), which is now close to implementation. The parameters whey/casein ratio and casein/total protein have also been calculated, based on relative concentrations of data from LS-ESI/MS from approximately 700 animals and the data forwarded to AU-MBG for initial genetic analyses of these parameters. The aim is to develop absolute quantitative methods for both α-LA, β-CN and the whey/casein ratio. In previous studies we have found relatively high variability in the relative concentration of α-LA within and between breeds, but increasing economic value is related to the amount, which the absolute quantification will provide. At AU-FOOD, a master student has been carrying out his master project linked to the project (Jonas Lund Rebsdorf).

At AU-MBG (protein part), we have in the first year of the project (2015) been focusing on setup and testing of in-house sandwich and competitive ELISAs for quantification of native full length osteopontin and the N-terminal fragment, which constitutes the majority of OPN in bovine milk. No commercial antibodies against bovine milk OPN exists, and generation of such in both chicken (eggs) and rabbits have shown to be difficult/unsuccesfull (probably due to the heavy and heterogenous modification of OPN which makes the epitopes highly variable). Bovine OPN has contains sequence regions that are very homologous to sequences in human and murine OPN. Hence we have selected and tested a number of monoclonal antibodies against human and murine OPN for cross-reactivity with the bovine milk OPN. One of these antibodies showed the desired cross-reactivity and this antibody is now used together with a rabbit-anti-bovine OPN polyclonal antibody to setup an ELISA. Though we have had good results with this setup, we may have to aquire or generate hybridomas (and make another monoclonal antibody against bovine OPN) as the background in the assay is rather high – and the optimal ELISA for measuring many samples should be based on two monoclonal antibodies.

In the genetic part at AU-MBG the focus in 2015 has been screening for genetic variability of selected milk proteins. We are currently re-sequencing four pools containing a large number of individuals each and these will soon be evaluated for genetic variation. Further, we already found a number of interesting SNP markers to test from previous sequencing projects and the 1000 Bulls Genome project, but these needs to be confirmed in the pools before TaqMan assays are created. A number of SNPs within the osteopontin gene have been tested using TaqMan assays in relation to mastitis and milk protein level and the results are currently being processed. Moreover, the four pools were designed in order to assess if additional SNPs within the sequenced genes would affect the somatic cell count (mastitis) or milk protein levels. By selecting cows with extreme phenotypes and then pool the extremes (High vs. Low) we will easily be able to evaluate from the sequencing if additional SNPs are likely to affect these two traits.  Genotyping of the 35 animals is completed and the dataset now encompasses 750 animals in total.

In addition in 2015 we have estimated the heritability of the major milk proteins using a multi-trait approach. Results show that the multi-trait approach gave a better estimate of the heritability (i.e. lower standard error) compared to the single trait approach. This is an important result for the project as the phenotypes we are working with are difficult to be measured in large scale (10,000+) and we still would like to get an estimate of the genetic parameters. A peer reviewed paper on this has been published in J. Dairy Sci..

Further we have performed a genome-wide scan of the Danish Holstein and Danish Jersey breed for the major milk proteins and their post translational modified forms. A number of genomic regions have been detected which could be interesting to explore further. The results have been submitted to BMC Genetics.