Advertisement

The mineral profile affects the coagulation pattern and cheese-making efficiency of bovine milk

Open ArchivePublished:May 27, 2021DOI:https://doi.org/10.3168/jds.2021-20233

      ABSTRACT

      Natural variations in milk minerals, their relationships, and their associations with the coagulation process and cheese-making traits present an opportunity for the differentiation of milk destined for high-quality natural products, such as traditional specialties or Protected Designation of Origin (PDO) cheeses. The aim of this study was to quantify the effects of the native contents of Ca, P, Na, K, and Mg on 18 traits describing traditional milk coagulation properties (MCP), curd firming over time (CFt) equation parameters, cheese yield (CY) measures, and nutrient recoveries in the curd (REC) using models that either included or omitted the simultaneous effects of milk fat and casein contents. The results showed that, by including milk fat and casein and the minerals in the statistical model, we were able to determine the specific effects of each mineral on coagulation and cheese-making efficiency. In general, about two-thirds of the apparent effects of the minerals on MCP and the CFt equation parameters are actually mediated by their association with milk composition, especially casein content, whereas only one-third of the effects are direct and independent of milk composition. In the case of cheese-making traits, the effects of the minerals were mediated only negligibly by their association with milk composition. High Ca content had a positive effect on the coagulation pattern and cheese-making traits, favoring water retention in the curd in particular. Phosphorus positively affected the cheese-making traits in that it was associated with an increase in CY in terms of curd solids, and in all the nutrient recovery traits. However, a very high P content in milk was associated with lower fat recovery in the curd. The variation in the Na content in milk only mildly affected coagulation, whereas with regard to cheese-making, protein recovery was negatively associated with high concentrations of this mineral. Potassium seemed not to be actively involved in coagulation and the cheese-making process. Magnesium content tended to slow coagulation and reduce CY measures. Further studies on the relationships of minerals with casein and protein fractions could deepen our knowledge of the role of all minerals in coagulation and the cheese-making process.

      Key words

      INTRODUCTION

      One of the main factors influencing the processing characteristics of milk is its composition (
      • Troch T.
      • Lefébure É.
      • Baeten V.
      • Colinet F.
      • Gengler N.
      • Sindic M.
      Cow milk coagulation: Process description, variation factors and evaluation methodologies. A review.
      ). As fat and protein are the most important milk components for the dairy industry, they are widely included in milk quality payment systems of sheep and goat (
      • Pirisi A.
      • Lauret A.
      • Dubeuf J.P.
      Basic and incentive payments for goat and sheep milk in relation to quality.
      ) and cattle milk (
      • Sneddon N.W.
      • Lopez-Villalobos N.
      • Hickson R.E.
      • Shalloo L.
      Review of milk payment systems to identify the component value of lactose.
      ), and also in the selection indices of several cattle breeds reared for dairy purposes (
      • Ghiroldi S.
      • Nicoletti C.
      • Santus E.
      • Rossoni A.
      • Bagnato A.
      ITE: The new selection index for the Italian Brown Swiss.
      ;
      • Miglior F.
      • Muir B.L.
      • Van Doormaal B.J.
      Selection indices in Holstein cattle of various countries.
      ;
      • Pryce J.E.
      • van der Werf J.H.J.
      • Haile-Mariam M.
      • Malcolm B.
      • Goddard M.E.
      Updated index weights for the Australian Profit Ranking in dairy cattle.
      ). The importance of protein, especially casein, lies in its active influence on the coagulation pattern (i.e., it increases the speed of curd firming and curd firmness; CF) and on cheese-making ability [i.e., high cheese yield (CY)] of the processed milk (
      • Verdier-Metz I.
      • Coulon J.-B.
      • Pradel P.
      Relationship between milk fat and protein contents and cheese yield.
      ;
      • Wedholm A.
      • Larsen L.B.
      • Lindmark-Månsson H.
      • Karlsson A.H.
      • Andrén A.
      Effect of protein composition on the cheese-making properties of milk from individual dairy cows.
      ). In contrast, fat plays a passive role during the coagulation process, as fat globules are entrapped in the para-casein matrix (
      • Fox P.F.
      • Guinee T.P.
      • Cogan T.M.
      • McSweeney P.L.H.
      Cheese: Structure, rheology and texture.
      ), and thus positively affects CY and the recovery of TS and energy in the curd (
      • Pazzola M.
      • Stocco G.
      • Dettori M.L.
      • Bittante G.
      • Vacca G.M.
      Effect of goat milk composition on cheese-making traits and daily cheese production.
      ). Besides fat and protein, other milk components influence milk processing characteristics, such as lactose and bacterial and somatic cell counts (
      • Leitner G.
      • Lavon Y.
      • Matzrafi Z.
      • Benun O.
      • Bezman D.
      • Merin U.
      Somatic cell counts, chemical composition and coagulation properties of goat and sheep bulk tank milk.
      ;
      • Bobbo T.
      • Ruegg P.L.
      • Stocco G.
      • Fiore E.
      • Gianesella M.
      • Morgante M.
      • Pasotto D.
      • Bittante G.
      • Cecchinato A.
      Associations between pathogen-specific cases of subclinical mastitis and milk yield, quality, protein composition, and cheese-making traits in dairy cows.
      ). Minerals, despite representing a small proportion of milk composition (about 0.7%;
      • Kaufmann W.
      • Hagemeister H.
      Composition of milk.
      ), also have a powerful influence in defining the structural characteristics and functional properties of casein micelles, and during milk coagulation and the other phases of the cheese-making process (
      • Lucey J.A.
      • Fox P.F.
      Importance of calcium and phosphate in cheese manufacture: a review.
      ;
      • Amenu B.
      • Deeth H.C.
      The impact of milk composition on cheddar cheese manufacture.
      ). Depending on their nature (e.g., nanoclusters or crystalline) and distribution (e.g., soluble or micellar forms), they are differently involved in the processing of the milk. Several studies have dealt with artificial modification of the mineral balance of usually reconstituted whole or skim milk from bovine species, mainly by adding Ca or chelating agents, to improve its rheological properties (
      • Cooke D.R.
      • McSweeney P.L.H.
      The influence of alkaline earth metal equilibria on the rheological properties of rennet-induced skim milk gels.
      ;
      • Bauland J.
      • Famelart M.H.
      • Bouhallab S.
      • Jeantet R.
      • Roustel S.
      • Faiveley M.
      • Croguennec T.
      Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
      ). When a mineral is added, the overall salt equilibrium in the milk changes, so the specific effects of the individual minerals on the coagulation pattern cannot be quantified. In contrast, very few studies have investigated the influence of the native mineral profile of raw milk on processing characteristics (
      • Malacarne M.
      • Franceschi P.
      • Formaggioni P.
      • Sandri S.
      • Mariani P.
      • Summer A.
      Influence of micellar calcium and phosphorous on rennet coagulation properties of cows milk.
      ), and these deal mainly with the effects of Ca (
      • Tsioulpas A.
      • Lewis M.J.
      • Grandison A.S.
      Effect of minerals on casein micelle stability of cows' milk.
      ;
      • Gustavsson F.
      • Glantz M.
      • Buitenhuis A.J.
      • Lindmark-Månsson H.
      • Stålhammar H.
      • Andrén A.
      • Paulsson M.
      Factors influencing chymosin-induced gelation of milk from individual dairy cows: major effects of casein micelle size and calcium.
      ;
      • Akkerman M.
      • Larsen L.B.
      • S⊘rensen J.
      • Poulsen N.A.
      Natural variations of citrate and calcium in milk and their effects on milk processing properties.
      ). The main issue of those studies is related to the fact that the native content of a given mineral in milk is not independent from the other minerals and milk components, and that a specific coagulation or cheese-making property is often the result of a sum of actions and interactions of different milk minerals and nutrients. Natural variations in milk minerals, and their relationships with each other and with coagulation and the cheese-making process, present an opportunity for the differentiation of milk for the production of high quality natural products, such as Protected Designation of Origin (PDO) cheeses, where production specifications and restrictions prohibit milk treatments and the addition of minerals before and during cheese making. In this scenario it is important to characterize the milk supply for the native mineral profile, also considering that such studies at individual animal level are beneficial for possible genetic improvement of dairy populations for milk quality.
      Moreover, the content of some minerals in milk is highly correlated with other milk components, particularly casein (
      • Lucey J.A.
      • Horne D.S.
      Milk salts: Technological significance.
      ). This means that if these minerals are not included in the statistical models, some of their effects on coagulation and cheese-making traits as reported in the literature will be confounded with the effects of milk composition, particularly the casein content, and vice versa. However, as far as we know, none of the studies published so far on this topic has considered the simultaneous effects of milk composition and mineral contents on cheese-making efficiency. The inclusion of casein in the statistical model is particularly important for minerals such as Ca and P, as they vary in proportion to the casein content of milk. As a result, the true effects of minerals on traditional coagulation properties (MCP) are still unclear, and their effects on curd firmness over time (CFt) equation parameters (obtained from modeling individual CF values recorded with a lactodynamograph;
      • Bittante G.
      • Contiero B.
      • Cecchinato A.
      Prolonged observation and modelling of milk coagulation, curd firming, and syneresis.
      ), and on cheese-making traits aside from traditional CY, such as CY expressed as the cheese solids and water retained in the curd, and milk nutrient recoveries in the curd (REC;
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Bittante G.
      • Summer A.
      Variations in milk protein fractions affect the efficiency of the cheese-making process.
      ), are completely unknown.
      The aim of this study, therefore, was to quantify the effects of Ca, P, Na, K, and Mg on 18 traits describing traditional MCP, CFt equation parameters, different CY measures, and REC traits using models either including or omitting the simultaneous effects of the main milk components (fat and casein).

      MATERIALS AND METHODS

      Experimental Design: Selection of Herds and Cows

      This study is part of a research project (Cowplus project) aimed at quantifying the effects of different dairy breeds and farming systems, while avoiding confounding them, on milk coagulation properties and cheese-making efficiency (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ,
      • Stocco G.
      • Cipolat-Gotet C.
      • Gasparotto V.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
      ). For the present study, we selected 27 multibreed farms representing the different farming systems in the Trentino-Alto Adige region (north-eastern Italian Alps). Milk samples from 240 cows (at different parities and lactation stages) were analyzed for their mineral profiles. The cows belonged to 6 breeds: 3 specialized dairy [Holstein-Friesian (50 cows from 15 herds), Brown Swiss (50 cows from 16 herds), and Jersey (35 cows from 7 herds)], and 3 dual-purpose [Simmental (35 cows from 11 herds), Rendena (34 cows from 8 herds), and Alpine Grey (34 cows from 9 herds)]. The herds were categorized as traditional farming system using summer pastures (n = 9), traditional without summer pastures (n = 11), traditional with silages (n = 2), and modern farming system using TMR (n = 5). A detailed description of the types of farming system in the study area can be found in
      • Berton M.
      • Bittante G.
      • Zendri F.
      • Ramanzin M.
      • Schiavon S.
      • Sturaro E.
      Environmental impact and efficiency of use of resources of different mountain dairy farming systems.
      .

      Milk Sampling and Analysis of Milk Composition and Mineral Profiles

      Samples were taken from the cows once during the evening milking (2 L of milk/cow) to carry out analyses of the milk chemical components, mineral profiles, and processing characteristics (coagulation properties and cheese-making traits). Immediately after collection, the samples were stored at 4°C and were analyzed within 24 h of the time of sampling. The fat, protein, casein, lactose, and TS contents of each milk sample were measured with a MilkoScan FT2 infrared analyzer (Foss Electric A/S), calibrated according to reference methods [ISO 1211/IDF for fat (
      • ISO-IDF (International Organization for Standardization and International Dairy Federation)
      Milk - Determination of fat content. International Standard ISO 1211 and IDF 1:2010.
      ); ISO 8968–2/IDF 20–2 for protein (
      • ISO-IDF (International Organization for Standardization and International Dairy Federation)
      Milk and milk products - Determination of nitrogen content - Part 1: Kjeldahl principle and crude protein calculation. International Standard ISO 8968–1 and IDF 1:2014.
      ); ISO 17997–1/IDF 29–1 for casein (
      • ISO-IDF (International Organization for Standardization and International Dairy Federation)
      Milk - Determination of casein-nitrogen content - Part 1: Indirect method. International Standard ISO 17997–1 and IDF 29–1:2004.
      ); ISO 26462/IDF 214 for lactose (
      • ISO-IDF (International Organization for Standardization and International Dairy Federation)
      Milk - Determination of lactose content - Enzymatic method using difference in pH. International Standard ISO 26462:2010 and IDF 214:2010.
      ); and ISO 6731/IDF 21 for TS (
      • ISO-IDF (International Organization for Standardization and International Dairy Federation)
      Milk, cream and evaporated milk - Determination of total solids content. International standard ISO 6731 and IDF 21:2010.
      )].
      Mineral contents (Ca, P, Na, K, and Mg) were determined using a Spectro Arcos EOP ICP-OES (Spectro A.I. GmbH). All instrument operating parameters were optimized for a 10% nitric acid solution as follows: axial plasma observation, Crossflow nebulizer, Scott Double Pass spray chamber, 3.0 mm diameter quartz injector torch, plasma power 1,400 W, coolant gas 12.0 L/min, auxiliary gas 0.6 L/min, nebulizer gas 0.85 L/min, additional gas 0.20 L/min, sample uptake rate 2.0 mL/min, replicate read time 28 s, 3 replicates, pre-flush time 60 s. The milk samples were analyzed after microwave closed vessel digestion (Ethos 1600; Milestone S.r.l.). Subsamples of between 1.950 and 2.050 g of each milk sample were placed in a vessel with 2 mL of 30% hydrogen peroxide and 7 mL of concentrated (65%) nitric acid, both Suprapur quality (Merck Chemicals GmbH). These subsamples were subjected to microwave digestion as follows: step 1, 25 to 200°C in 18 min at 1,500 W with maximum pressure 45 bar; step 2, 200°C for 15 min at 1,500 W with maximum pressure 45 bar; step 3, 200 to 110°C for 15 min. After cooling to room temperature, the dissolved sample was diluted with ultrapure water (resistivity 18.2 M Ω cm at 25°C) to a final volume of 20 mL. Calibration standards were prepared using multi-element and single-element standard solutions (Inorganic Ventures Inc.) in 10% Suprapur nitric acid (Merck Chemicals GmbH) to obtain matrices similar to the samples. Calibration solutions of the analytes were prepared at common concentrations of 0, 0.002, 0.005, 0.02, 0.05, 0.2, 0.5, and 2 mg/L, as well as further concentrations of 5, 20, 50, and 200 mg/L, respectively, of calcium, potassium, magnesium, sodium, and phosphorus. The accuracy and precision of these calibration solutions were tested by analyzing a blank solution, a low-level control solution (recovery limits ± 30%), a medium-level control solution (recovery limits ± 10%), and the international standard reference material BCR-063R “Skim milk powder” (Institute for Reference Materials and Measurements), prepared as described above. The measured values and the certified values were in excellent agreement for all 5 minerals. Detailed macro- and micro-mineral profiles of these milk samples and the effects of dairy system, breed, parity, and lactation stage of the cows were reported in a previous study (
      • Stocco G.
      • Summer A.
      • Malacarne M.
      • Cecchinato A.
      • Bittante G.
      Detailed macro- and micromineral profile of milk: Effects of herd productivity, parity, and stage of lactation of cows of 6 dairy and dual-purpose breeds.
      ).

      Traditional Milk Coagulation Properties

      Milk coagulation properties were measured using a mechanical lactodynamograph (Formagraph, Foss Electric A/S), with pendula calibration carried out before each session of the trial. Each sample (10 mL of milk) was heated to 35°C, then mixed with 200 μL of rennet solution (Hansen Standard 215 with 80 ± 5% chymosin and 20 ± 5% pepsin; 215 international milk clotting units/mL; Pacovis Amrein AG) freshly diluted to 1.2% (wt/vol) in distilled water. Coagulation temperature was maintained at 35°C and the duration of the analysis was 60 min. Traditional single-point measurements of each milk sample [rennet coagulation time (RCT; min), time interval between gelation and attainment of curd firmness of 20 mm (k20; min), and curd firmness at 30, 45, and 60 min after rennet addition (a30, a45, and a60, respectively, mm)] were obtained directly from the instrument.

      Modeling the Coagulation Pattern

      The Formagraph recorded the width (in mm) of the oscillatory graph of the pendulum submerged in the milk-filled wells every 15 s. Thus, 240 CF values were recorded for each milk sample. A 4-parameter model was used to fit the CF over time values of each sample. This model, which uses all the information available for estimating the 4 coagulation parameters (
      • Bittante G.
      • Contiero B.
      • Cecchinato A.
      Prolonged observation and modelling of milk coagulation, curd firming, and syneresis.
      ), was as follows:
      CFt=CFP×[1ekCF×(tRCTeq)]×ekSR×(tRCTeq),
      [1]


      where CFt is curd firmness at time t (mm); CFP is the asymptotic potential value of CF at an infinite time (mm); kCF is the curd firming instant rate constant (%/min); kSR is the syneresis instant rate constant (%/min); and RCTeq is RCT estimated by the CFt equation on the basis of all data points (min). These parameters provide additional information to the traditional MCP because (1) CFP is conceptually independent of test duration and does not depend on RCT (as a30 does); (2) kCF describes the increase in CF after RCT toward CFP; (3) kSR represents the expulsion of whey from the coagulum and describes the apparent decrease in CF after RCT; soon after RCT the effect of kCF prevails over the effect of kSR, and the CFt curve increases until its maximum firmness value (CFmax) is reached at a point in time (tmax) when the 2 effects are equal; after tmax, the effect of kSR prevails over the effect of kCF, and the CFt curve declines asymptotically toward zero; (4) the RCTeq has the same meaning as the traditional RCT, but is now estimated using all available data. To avoid convergence and estimation problems, the procedure described by
      • Bittante G.
      • Contiero B.
      • Cecchinato A.
      Prolonged observation and modelling of milk coagulation, curd firming, and syneresis.
      was modified to include CF measurements up to 45 min from the addition of rennet, whereas CFP was calculated by multiplying CFmax by 1.34, which is the coefficient resulting from the linear regression between CFP and CFmax values obtained in a preliminary analysis. The other 3 CFt model parameters (RCTeq, kCF, and kSR) were estimated by curvilinear regression using the nonlinear procedure (PROC NLIN) in the SAS software (SAS Institute Inc.). The parameters of each individual equation were estimated using the Marquardt iterative method (350 iterations and a 10−5 level of convergence), according to
      • Bittante G.
      • Contiero B.
      • Cecchinato A.
      Prolonged observation and modelling of milk coagulation, curd firming, and syneresis.
      . The coefficient of determination of individual equations was very good, being higher that 0.99 for 99.9% of the total observations (data not shown).

      Model Cheese-Making and Related Traits

      We used the individual cheese-making procedure described by
      • Stocco G.
      • Cipolat-Gotet C.
      • Gasparotto V.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
      to measure CY and REC traits. Briefly, milk samples (1.5 L of milk/cow) were heated to 35°C (30 min), then mixed with 8 mL of rennet solution (Hansen Standard 215 with 80 ± 5% chymosin and 20 ± 5% pepsin; 215 international milk clotting units/mL; Pacovis Amrein AG). Gelation time was determined by visual observation of gelation of the milk with the aid of a spoon. Curd firming occurred at 50°C (cooking phase, 20 min). The curd was cross-cut 10 min after gelation had occurred, then 10 min after cross-cutting the curd was separated from the whey (draining phase, 30 min). During the draining phase, the curd was gently pressed and turned over to facilitate whey expulsion. In the last 10 min of this phase, the curd was shaped into wheels in small cylindrical molds and was left in the whey. The model cheeses thus formed were pressed for 30 min, turning over every 10 min, and were then immersed in liquid brine for 30 min. The whey was analyzed for chemical composition (fat, protein, lactose, and TS) with a MilkoScan FT2 infrared analyzer.
      Cheese-making traits were calculated from the weights of the milk and whey (g) and their chemical compositions, as described by
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Bittante G.
      • Summer A.
      Variations in milk protein fractions affect the efficiency of the cheese-making process.
      . Briefly, the traits measured were CYCURD, CYSOLIDS, and CYWATER, calculated as the ratio of the weight (g) of fresh curd, curd DM, and curd water, respectively, to the weight of the milk processed (g); RECPROTEIN, RECFAT, and RECSOLIDS, calculated as the ratio of the weight (g) of the component (protein, fat, and DM, respectively) in the curd to the weight of the corresponding component in the milk (g). Recovery of energy in the curd (RECENERGY) was determined by estimating the energy in the milk and in the curd using the equation proposed by
      • NRC
      Nutrient Requirements of Dairy Cattle.
      and converted into MJ/kg.

      Statistical Analysis

      The values of the 25 traits examined here (composition, mineral profile, coagulation, and cheese-making traits) outside the interval of the mean ± 3 standard deviations were designated as outliers and excluded. All traits were analyzed using 2 mixed linear models (MIXED procedure; SAS Institute Inc.). The first comprehensive linear mixed model (M1) was
      yfghijklmnopqr = μ + Herdf + Breedg + Parityh + DIMi + Caj + Pk + Nal + Km + Mgo + fatp + caseinq + efghijklmnopqr,
      [2]


      where yfghijklmnopqr is the observed trait (fat, casein, Ca, P, Na, K, Mg, RCT, k20, a30, a45, a60, RCTeq, kCF, kSR, CFmax, tmax, CFP, CYCURD, CYSOLIDS, CYWATER, RECFAT, RECPROTEIN, RECSOLIDS, RECENERGY); μ is the overall intercept of the model; Herdf is the random effect of the fth herd (f = 1 to 27); Breedg is the random effect of the gth breed (g = Holstein-Friesian, Brown Swiss, Jersey, Simmental, Rendena, and Alpine Grey); Parityh is the fixed effect of the hth parity (h = 1 to ≥ 3; first parity = 80 cows; second parity = 59 cows; ≥third parity = 99 cows); DIMi is the fixed effect of the ith class of DIM (i = 1 to 7; class 1 = 8–49 d, 25 cows; class 2 = 50–91 d, 27 cows; class 3 = 92–133 d, 39 cows; class 4 = 134–175 d, 42 cows; class 5 = 176–217 d, 43 cows; class 6 = 218–259 d, 32 cows; class 7 = >259 d, 30 cows); each mineral was included in classes according to quintiles based on its contents in the milk. Ranges of minerals per each quintile and the number of cows per each quintile is reported in Supplemental Table S1 (https://figshare.com/articles/dataset/Supplemental_Table_S1/14602095;
      • Stocco G.
      Supplemental Table S1. Figshare. Dataset.
      ). Caj is the fixed effect of the jth quintile of Ca (j = 1 to 5); Pk is the fixed effect of the kth quintile of P (k = 1 to 5); Nal is the fixed effect of the lth quintile of Na (l = 1 to 5); Km is the fixed effect of the mth quintile of K (m = 1 to 5) Mgo is the fixed effect of the oth quintile of Mg (o = 1 to 5); fatp is the fixed effect included in the model as a linear covariate; caseinq is the fixed effect included in the model as a linear covariate; and efghijklmnopqr is the random residual ∼N(0,σe2), where σe2 is the residual variance. When fat, casein, or one of the minerals in milk was considered a dependent variable, it was, of course, excluded from the model's independent variables.
      A reduced version of model M1, named model M2, was obtained by excluding the fat and casein covariates. This model was used to carry out an auxiliary analysis to quantify the effects of the 5 minerals not corrected for fat and casein contents (i.e., the confounding effect of milk composition and mineral profile). The results obtained from the M2 model are not described and discussed analytically in this paper, but they are reported as Supplemental Table S2 (https://figshare.com/articles/dataset/Supplemental_Table_S2/14602110;
      • Stocco G.
      Supplemental Table S2. Figshare. Dataset.
      ). Pearson's product-moment correlations were estimated among fat, casein, and the minerals, and are presented as supplemental material (Supplemental Figure S1, https://figshare.com/articles/figure/Supplemental_Figure_S1/14602113;
      • Stocco G.
      Supplemental Figure S1. Figshare. Figure.
      ).

      RESULTS AND DISCUSSION

      The rationale of this study relies on many aspects, among which the most important are (1) data are based on the comparison of individual milk samples of different characteristics in terms of composition, origin, farming system, breed, and animals; (2) the results obtained are representative of many conditions, given that the experimental design and the statistical models adopted are able to avoid overlapping effects and multicollinearity; (3) the coagulation and cheese-making ability of milk in relation to the mineral content has never been studied before in terms of CFt parameters (i.e., RCTeq, kCF, CFmax, tmax, kSR, and CFP), different measures of CY (i.e., CYCURD, CYSOLIDS, and CYWATER), and REC traits (i.e., RECFAT, RECPROTEIN, RECSOLIDS, and RECENERGY). Beyond scientific relevance, we believe that the present study is important for the dairy industry because the highest priced cheeses are often those protected by designations (such as PDO by the European Union, or organic products) that forbid any addition of chemicals during cheese-making.

      Major Sources of Variation in Milk Fat, Casein, and Mineral Contents

      Table 1 reports the descriptive statistics and results of the ANOVA of fat, casein, and minerals using the comprehensive model (M1). The effects of the 6 breeds, herds, parity, and DIM on the minerals were previously investigated and reported by
      • Stocco G.
      • Summer A.
      • Malacarne M.
      • Cecchinato A.
      • Bittante G.
      Detailed macro- and micromineral profile of milk: Effects of herd productivity, parity, and stage of lactation of cows of 6 dairy and dual-purpose breeds.
      using the same data, so they will not be discussed here. These factors, together with fat and casein, were of course included in the models to correctly quantify the effect of the minerals on the dependent variables. It is just worth noting that the importance of the effects of herd and breed of cow varied greatly according to the different traits: together they represented about half the total variance in the P content of milk, and only 13% in Ca and K (Table 1). Stage of lactation was very important for casein and Na contents (P < 0.001), less important for Ca and P (P < 0.05), and not significant for fat, K, and Mg contents, whereas the cow's parity affected only P (P < 0.01), Na, and Mg (P < 0.001).
      Table 1Descriptive statistics (mean ± SD) and ANOVA of milk components (fat and casein) and of milk minerals (calcium, phosphorus, sodium, potassium, and magnesium)
      ItemMilk component, %Mineral, mg/L
      FatCaseinCaPNaKMg
      Mean4.232.651,2231,0143571,694103
      ±SD2.000.421321144610914
      Random factor,
      The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      %
       Herd11.211.49.826.713.711.815.1
       Breed16.145.73.223.97.31.215.0
      Fixed factor (F-value)
       DIM0.86.5
      P < 0.001.
      3.0
      P < 0.05
      2.5
      P < 0.05
      4.0
      P < 0.001.
      2.02.0
       Parity0.50.51.86.7
      P < 0.01
      19.9
      P < 0.001.
      0.79.0
      P < 0.001.
       Fat0.34.5
      P < 0.05
      0.37.6
      P < 0.01
      3.35.5
      P < 0.05
       Casein1.519.9
      P < 0.001.
      56.5
      P < 0.001.
      0.225.4
      P < 0.001.
      9.1
      P < 0.001.
       Ca0.73.9
      P < 0.01
      5.4
      P < 0.001.
      1.21.62.3
       P1.113.4
      P < 0.001.
      6.3
      P < 0.001.
      2.45.2
      P < 0.001.
      1.3
       Na0.71.81.11.12.6
      P < 0.05
      6.5
      P < 0.001.
       K0.86.4
      P < 0.001.
      0.58.2
      P < 0.001.
      2.11.4
       Mg1.72.42.4
      P < 0.05
      2.310.1
      P < 0.001.
      1.2
      RMSE
      RMSE = root mean square error.
      0.90.280.458.532.684.48.4
      1 The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      2 RMSE = root mean square error.
      * P < 0.05
      ** P < 0.01
      *** P < 0.001.
      As expected, fat and casein were associated with the mineral profile of milk. Milk fat affected the contents of Ca, Na, and Mg, but it was not in turn affected by any of the minerals (Table 1). Casein was much more interrelated with macro-minerals: it influenced all the minerals, except Na, and was in turn affected by all the minerals, except Na and Mg (Table 1).
      Relationships among minerals were also observed: 8 out of 20 possible mineral-on-mineral effects were significant (Table 1). When we compared these results with those obtained from the model that did not include fat and casein (M2), summarized in Supplemental Table S2, we found differences in the relationships among the minerals, as also evidenced by the different number of significant mineral-on-mineral effects (13 out of 20), as summarized in Figure 1. This means that some of the mineral-on-mineral effects are most likely due to an indirect effect of milk gross composition, especially the casein content. In particular, the effects of Ca on K, Ca on Mg, P on Na, P on Mg, and Mg on P were significant in the model that did not correct for milk composition, but were no longer significant when milk fat and casein were taken into account. The other 8 mineral-on-mineral effects reported in Table 1 were still significant, although their effect tended to lessen after fat and casein correction.
      Figure thumbnail gr1
      Figure 1Total number of effects of minerals on minerals and milk technological traits tested, number of effects significant in the base model (P < 0.05), and number of effects significant in the model including the covariate of milk fat and casein contents (P < 0.05 F+C). Minerals = milk content of Ca, P, Na, K, and Mg (5 traits); MCP = milk traditional coagulation properties (5 traits); CFt = parameters of the curd firming equation (6 traits); CY = cheese yields (3 traits); REC = recovery of milk nutrients in the curd (4 traits).

      Major Sources of Variation in Coagulation and Cheese-Making Traits

      The statistical analyses of the traditional MCP and CFt equation parameters are summarized in Table 2. After including in the model the breed, parity, and lactation stage of the cows, and the composition and mineral profile of the milk, the effect of herd on coagulation and curd firming traits was moderate (6.3 to 18.0% of total variance) for all traits, except for the curd firming (kCF) and curd syneresis (kSR) instant rate constants, which were almost unaffected by herd.
      Table 2Descriptive statistics (mean ± SD) and ANOVA of traditional milk coagulation properties (MCP) and of curd firming over time (CFt) equation parameters
      ItemMCP
      RCT = measured rennet gelation time; k20 = time interval between gelation and attainment of curd firmness of 20 mm; a30, a45, and a60 = curd firmness 30, 45, and 60 min after rennet addition.
      Curd firming (CFt) equation parameter
      RCTeq = rennet coagulation time estimated by CFt modeling; kCF = curd firming instant rate constant; kSR = syneresis instant rate constant; CFP = asymptotic potential curd firmness; tmax = time at achievement of maximum curd firmness (CFmax).
      RCT, mink20, mina30, mma45, mma60, mmRCTeq, minkCF, %/minkSR, %/minCFP, mmtmax, min
      Mean15.83.5845.251.551.616.09.150.7473.146.8
      ±SD5.41.6316.312.512.05.422.680.2615.69.2
      Random factor,
      The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      %
       Herd12.68.418.012.611.413.53.10.513.36.3
       Breed9.10.26.91.00.69.54.03.21.57.7
      Fixed factor (F-value)
       DIM2.8
      P < 0.05
      0.93.0
      P < 0.01
      1.31.33.0
      P < 0.01
      1.21.11.61.4
       Parity2.41.92.64.1
      P < 0.05
      2.52.52.02.54.0
      P < 0.05
      4.3
      P < 0.05
       Fat0.11.11.01.82.20.10.82.30.80.4
       Casein0.214.5
      P < 0.001.
      13.2
      P < 0.001.
      24.9
      P < 0.001.
      25.6
      P < 0.001.
      0.31.40.738.5
      P < 0.001.
      0.0
       Ca2.6
      P < 0.05
      2.42.5
      P < 0.05
      1.31.52.7
      P < 0.05
      3.1
      P < 0.05
      1.71.72.7
      P < 0.05
       P1.61.51.30.91.01.51.10.70.80.7
       Na2.01.92.00.50.52.21.31.40.33.1
      P < 0.05
       K1.52.21.40.40.91.71.60.90.71.7
       Mg0.62.5
      P < 0.05
      1.52.11.40.70.10.11.90.8
      RMSE
      RMSE = root mean square error.
      4.51.312.29.69.34.52.30.310.88.1
      1 RCT = measured rennet gelation time; k20 = time interval between gelation and attainment of curd firmness of 20 mm; a30, a45, and a60 = curd firmness 30, 45, and 60 min after rennet addition.
      2 RCTeq = rennet coagulation time estimated by CFt modeling; kCF = curd firming instant rate constant; kSR = syneresis instant rate constant; CFP = asymptotic potential curd firmness; tmax = time at achievement of maximum curd firmness (CFmax).
      3 The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      4 RMSE = root mean square error.
      * P < 0.05
      ** P < 0.01
      *** P < 0.001.
      The effect of breed was even lower than that of herd (0.2 to 9.1%, Table 2), due to the inclusion of milk composition and mineral profile in the model, which explained a large part of the differences among breeds observed for these traits in a previous study (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ). The effects of parity and lactation stage after including milk composition and mineral profile were also smaller here compared with those reported by
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      .
      Milk fat content did not have a direct effect on milk coagulation, curd firming, and syneresis, whereas casein, as expected, exerted quite a large influence on these traits. Casein favorably affected all traditional MCP, except RCT, as well as the CFP (and CFmax) of the CFt equation parameters (Table 2). As the casein content is interrelated with the milk mineral profile, as can be seen in Table 1, including it in the statistical model together with the minerals made it possible to interpret the results more accurately. It is worth noting that the minerals were significantly involved in 9 of the 25 possible effects on traditional MCP, and 10 of the 30 possible effects on the CFt equation parameters when the experimental data set was analyzed using the model that did not include the milk fat and casein covariates (Figure 1). However, the effect of some minerals remained significant after including milk composition: Ca content on RCT, a30, RCTeq, kCF, and tmax; Na content on tmax; and Mg content on k20 (Table 2). This confirms that about two-thirds of the apparent effects of minerals on MCP and the CFt equation parameters are in fact mediated by their association with milk composition, especially casein content; only one-third of the effects on these traits can be directly attributed to the minerals (particularly Ca) independent of milk composition.
      Moving to cheese-making traits (Table 3), the herd effect was moderate (15.6 to 19.5% of total variance) for the 3 CY and for RECSOLIDS, and much smaller (3.1 to 6.6%) for the other recovery traits. Again, the effect of breed was much smaller (<7.0%), with the only exception of RECFAT (10.8%). The effects of parity and lactation stage on cheese-making traits were never significant (Table 3), unlike in other studies where milk composition and mineral profile were not included in the statistical model (
      • Cipolat-Gotet C.
      • Cecchinato A.
      • De Marchi M.
      • Bittante G.
      Factors affecting variation of different measures of cheese yield and milk nutrients recovery from an individual model cheese manufacturing process.
      ;
      • Stocco G.
      • Cipolat-Gotet C.
      • Gasparotto V.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
      ).
      Table 3Descriptive statistics (mean ± SD) and ANOVA of cheese yield (CY) measures and nutrient recovery traits (REC)
      ItemCY,
      The traits measured were CYCURD, CYSOLIDS, and CYWATER, calculated as the ratio of the weight (g) of fresh curd, curd DM, and curd water, respectively, to the weight of the milk processed (g).
      %
      REC,
      The traits measured were RECPROTEIN, RECFAT, and RECSOLIDS, calculated as the ratio of the weight (g) of the component (protein, fat, and DM, respectively) in the curd to the weight of the corresponding component in the milk (g). Recovery of energy in the curd (RECENERGY) was determined by estimating the energy in the milk and in the curd using the equation proposed by NRC (2001) and converted into MJ/kg.
      %
      CYCURDCYSOLIDSCYWATERRECFATRECPROTEINRECSOLIDSRECENERGY
      Mean15.78.427.2485.179.453.469.0
      ±SD3.01.691.384.321.95.04.2
      Random factor,
      The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      %
       Herd15.815.619.56.63.115.76.6
       Breed6.16.21.710.83.92.37.0
      Fixed factor (F-value)
       DIM0.60.60.41.10.70.50.8
       Parity0.30.20.61.80.70.20.6
       Fat286.4
      P < 0.001.
      1,891.5
      P < 0.001.
      26.6
      P < 0.001.
      2.08.6
      P < 0.01
      344.6
      P < 0.001.
      153.8
      P < 0.001.
       Casein169.0
      P < 0.001.
      248.0
      P < 0.001.
      102.4
      P < 0.001.
      1.48.8
      P < 0.01
      55.8
      P < 0.001.
      9.6
      P < 0.01
       Ca3.8
      P < 0.01
      2.43.3
      P < 0.05
      4.0
      P < 0.01
      1.52.5
      P < 0.05
      3.9
      P < 0.01
       P0.73.6
      P < 0.01
      0.24.4
      P < 0.01
      4.0
      P < 0.01
      3.6
      P < 0.01
      3.9
      P < 0.01
       Na1.71.11.52.18.9
      P < 0.001.
      2.6
      P < 0.05
      1.9
       K1.50.91.61.31.30.80.8
       Mg3.1
      P < 0.05
      2.6
      P < 0.05
      2.5
      P < 0.05
      1.21.02.42.1
      RMSE
      RMSE = root mean square error.
      0.80.20.73.61.41.82.1
      1 The variance of each random factor is expressed as percentage of the sum of variances of all random factors (including residual variance).
      2 The traits measured were CYCURD, CYSOLIDS, and CYWATER, calculated as the ratio of the weight (g) of fresh curd, curd DM, and curd water, respectively, to the weight of the milk processed (g).
      3 The traits measured were RECPROTEIN, RECFAT, and RECSOLIDS, calculated as the ratio of the weight (g) of the component (protein, fat, and DM, respectively) in the curd to the weight of the corresponding component in the milk (g). Recovery of energy in the curd (RECENERGY) was determined by estimating the energy in the milk and in the curd using the equation proposed by
      • NRC
      Nutrient Requirements of Dairy Cattle.
      and converted into MJ/kg.
      4 RMSE = root mean square error.
      * P < 0.05
      ** P < 0.01
      *** P < 0.001.
      As expected, milk fat and casein contents played an essential role in explaining the variability in the 3 CY measures and in the recovery traits, with the only exception being RECFAT (Table 3). Nevertheless, the numbers of significant effects of milk mineral content on these traits changed little whether or not the milk fat and casein covariates were included in the statistical model (Figure 1). Although insoluble minerals (Ca and phosphate) associated with the para-casein matrix are known to influence %CY (
      • Fox P.F.
      • Guinee T.P.
      • Cogan T.M.
      • McSweeney P.L.H.
      Cheese yield.
      ), our results suggest that the effects of minerals on cheese-making traits are barely mediated by their association with the milk composition.

      Calcium

      Calcium is one of the most important minerals in milk. In the aqueous phase, Ca is present in ionic form, and is associated with citrate and inorganic phosphate to form calcium citrate and calcium phosphate, respectively. In the micellar phase, Ca is bound to phosphoseryl residues of casein molecules and inorganic phosphate (i.e., colloidal calcium phosphate, CCP). The presence of calcium phosphate clusters in the micelles is essential to the structure of the protein particles and to their technological functionality (
      • Dalgleish D.G.
      • Corredig M.
      The structure of the casein micelle of milk and its changes during processing.
      ). In this study, Ca appeared to be the mineral with the greatest effect on milk quality and technological properties: 19 of the 23 traits studied were significantly affected by milk Ca content. This number decreased after milk fat and casein were also included in the model (Figure 2), but nonetheless remained substantial (11 out of 23 traits). Milk Ca was associated with fat and casein, and was also related to the contents of P and Mg (Supplemental Figure S1).
      Figure thumbnail gr2
      Figure 2Total number of effects of each individual mineral on the 23 milk technological traits tested, number of effects significant in the base model (P < 0.05), and number of effects significant in the model including the covariate of milk fat and casein contents (P < 0.05 F+C).
      Figure 3 depicts the pattern of the CFt equation parameters across different concentrations of Ca in milk. Clearly, the overall coagulation process improved at increasing levels of Ca in milk. In particular, coagulation time traits (RCT and RCTeq) were shortened by about 4 min moving from the lowest (1,059 mg/L) to the highest (1,445 mg/L) average Ca concentration quintile. Curd firming was faster (about + 2%/min of kCF), so that at 30 min the curd was also firmer (about +10 mm of a30), and CFmax was reached faster (about −4 min of tmax) in milk samples with high compared with low Ca concentrations. Even after including fat and casein in the model, the effect of Ca remained substantial on coagulation traits, as depicted in Supplemental Figure S2a (https://figshare.com/articles/figure/Supplemental_Figure_S2/14625513,
      • Stocco G.
      Supplemental Figure S2. Figshare. Figure.
      ).
      Figure thumbnail gr3
      Figure 3Effect of Ca on gelation, curd firming, and syneresis of individual milk samples. Mineral concentrations are reported as mean values (mg/L) of each quintile of the distribution.
      Our results are in agreement with those reported by
      • Tsioulpas A.
      • Lewis M.J.
      • Grandison A.S.
      Effect of minerals on casein micelle stability of cows' milk.
      for the effects of the natural mineral contents of 235 milk samples on casein micelle stability and some technological traits (i.e., RCT and coagulum firmness, measured by rheometer), and by
      • Akkerman M.
      • Larsen L.B.
      • S⊘rensen J.
      • Poulsen N.A.
      Natural variations of citrate and calcium in milk and their effects on milk processing properties.
      , who investigated the natural variation in Ca and citrate contents in skim milk in relation to RCT and the curd firming rate (measured by rheometer).
      • Ketto I.A.
      • Knutsen T.M.
      • Øyaas J.
      • Heringstad B.
      • Ådn⊘y T.
      • Devold T.G.
      • Skeie S.B.
      Effects of milk protein polymorphism and composition, casein micelle size and salt distribution on the milk coagulation properties in Norwegian Red cattle.
      analyzed the correlations between Ca, P, and Mg contents and coagulation properties (measured by a mechanical instrument) in 99 milk samples and found that Ca was associated negatively with RCT (r = −0.21, P < 0.01) and k20 (r = −0.23, P < 0.01), and positively with a30 (r = 0.27, P < 0.001), although the coefficients were low. Those authors used Pearson correlations to assess only the linear relationships between minerals and coagulation properties, without correcting for any other affecting factor (i.e., herd, animal, and milk components).
      Other studies have found several differences in the mineral contents of milk between samples exhibiting good and poor coagulation. In an investigation of the causes of noncoagulating milk from Danish-Holstein cows (n = 20),
      • Frederiksen P.D.
      • Andersen K.K.
      • Hammersh⊘j M.
      • Poulsen H.D.
      • S⊘rensen J.
      • Bakman M.
      • Qvist K.B.
      • Larsen L.B.
      Composition and effect of blending of noncoagulating, poorly coagulating, and well-coagulating bovine milk from individual Danish Holstein cows.
      found no differences in total Ca, P, and Mg contents between well and poorly coagulating milk samples. In contrast,
      • Jensen H.B.
      • Poulsen N.A.
      • Andersen K.K.
      • Hammersh⊘j M.
      • Poulsen H.D.
      • Larsen L.B.
      Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
      , also looking at the underlying causes of poorly coagulating milk from Holstein-Friesian and Jersey cows (n = 102), found some differences in the total, soluble, and micellar fractions between milk samples exhibiting good and poor coagulation. They found that total and micellar Ca, and soluble and micellar P were higher in well than in poorly coagulating Jersey milk samples, whereas both total and micellar Ca and P, and micellar Mg were higher in well than in poorly coagulating Holstein-Friesian milk samples. However, those authors did not study the direct effects of each mineral on the coagulation properties of their samples, and therefore did not quantify them.
      Milk Ca content also strongly affected CY measures and REC traits (Table 3). Fresh CY was higher in milk samples with elevated Ca concentrations than in milk samples with low Ca (about +1% on an average of 15.7%, i.e., a favorable effect of +6%; Figure 4a). This effect seems due mainly to the increased retention of water in the curd (CYWATER), although the trends were rather cubic. Regarding the REC traits, a higher Ca content in milk also resulted in higher RECFAT (about +5%) and RECSOLIDS (+2%; Figure 4b), leading to a 3% higher RECENERGY (data not shown). Although not related to the native mineral content of milk, previous studies evidenced that the positive effects of the addition of CaCl2 on the recovery of fat and protein and cheese yield were probably due to the increased aggregation of caseins (
      • Fox P.F.
      • Guinee T.P.
      • Cogan T.M.
      • McSweeney P.L.H.
      Cheese yield.
      ).
      Figure thumbnail gr4
      Figure 4Effect of Ca on CYCURD and CYWATER [calculated as the ratio of the weight (g) of fresh curd and curd water, respectively, to the weight of the milk processed (g)] [a], RECFAT and RECSOLIDS [calculated as the ratio of the weight (g) of the component (fat and DM, respectively) in the curd to the weight of the corresponding component in the milk (g)] [b] of individual milk samples, and the coefficient of determination (R2) of the regression. CY = cheese yield; REC = recovery of milk nutrients in the curd.
      It is important to remember that our results for Ca are adjusted for the effects of fat and casein, the main factors influencing MCP, the CFt equation parameters, and cheese-making traits (
      • Bland J.H.
      • Grandison A.S.
      • Fagan C.C.
      Evaluation of milk compositional variables on coagulation properties using partial least squares.
      ;
      • Pazzola M.
      • Stocco G.
      • Dettori M.L.
      • Bittante G.
      • Vacca G.M.
      Effect of goat milk composition on cheese-making traits and daily cheese production.
      ;
      • Cipolat-Gotet C.
      • Malacarne M.
      • Summer A.
      • Cecchinato A.
      • Bittante G.
      Modeling weight loss of cheese during ripening and the influence of dairy system, parity, stage of lactation, and composition of processed milk.
      ). Possibly, further understanding could be achieved by analyzing the mineral profile of standardized milk samples (e.g., fat to protein ratio) and by quantifying the effect of each mineral in milk samples with the same composition (
      • Auldist M.J.
      • Johnston K.A.
      • White N.J.
      • Fitzsimons W.P.
      • Boland M.J.
      A comparison of the composition, coagulation characteristics and cheesemaking capacity of milk from Friesian and Jersey dairy cows.
      ). Moreover, since
      • Bauland J.
      • Famelart M.H.
      • Bouhallab S.
      • Jeantet R.
      • Roustel S.
      • Faiveley M.
      • Croguennec T.
      Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
      confirmed that the soluble and colloidal forms of Ca are important in explaining the changes in the coagulation properties of milk, it would be interesting to assess the effect of each mineral form on coagulation and cheese-making properties of milk.

      Phosphorus

      Phosphorus is present in milk as organic (i.e., bound to casein) and inorganic phosphates (i.e., ions). Inorganic phosphates are equally distributed between the aqueous and micellar phases (i.e., CCP) at a milk pH of 6.7. In this study, P appeared to have the second largest effect on milk quality and technological properties after Ca. Thirteen of the 23 traits studied here were significantly affected by milk P concentrations when analyzed with the M2 statistical model, and the number of traits decreased to 7 when milk fat and casein were also included in the model (Figure 2). Milk P content was not associated with fat, but it was the mineral with the strongest association with casein content (Table 1). It should also be pointed out that P, Ca, and K contents are mutually influential (Table 1).
      The quantity of CCP and the number of phosphate groups in the casein micelle seem to influence the rennet coagulation of milk (
      • Malacarne M.
      • Franceschi P.
      • Formaggioni P.
      • Sandri S.
      • Mariani P.
      • Summer A.
      Influence of micellar calcium and phosphorous on rennet coagulation properties of cows milk.
      ), and the interaction of caseins with CCP enhances the aggregation of the para-casein micelles (
      • Bauland J.
      • Famelart M.H.
      • Bouhallab S.
      • Jeantet R.
      • Roustel S.
      • Faiveley M.
      • Croguennec T.
      Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
      ). However, in the present study, the P content of milk had no effect on either MCP or the CFt equation parameters when fat and casein were included in the model, but when they were not included, P showed significant associations with coagulation, curd firming, and syneresis traits (Supplemental Figure S2b, https://figshare.com/articles/figure/Supplemental_Figure_S2/14625513,
      • Stocco G.
      Supplemental Figure S2. Figshare. Figure.
      ). This means that the effects of P sometimes reported in the literature were probably mediated by its strong association with casein. The effects of P on coagulation traits were not linear, since milk samples with P concentrations between 983 and 1,047 mg/L showed shorter RCT, faster kCF, the highest kSR, and higher CF compared with both low and high milk P concentrations.
      • Ketto I.A.
      • Knutsen T.M.
      • Øyaas J.
      • Heringstad B.
      • Ådn⊘y T.
      • Devold T.G.
      • Skeie S.B.
      Effects of milk protein polymorphism and composition, casein micelle size and salt distribution on the milk coagulation properties in Norwegian Red cattle.
      reported low linear correlation coefficients between P and some coagulation traits: −0.22 for k20, and 0.22 for and a30 and gel firming rate.
      • Gustavsson F.
      • Glantz M.
      • Buitenhuis A.J.
      • Lindmark-Månsson H.
      • Stålhammar H.
      • Andrén A.
      • Paulsson M.
      Factors influencing chymosin-induced gelation of milk from individual dairy cows: major effects of casein micelle size and calcium.
      observed a significant effect of P on the gelation time of 98 individual milk samples from Swedish Red cows. The nonlinearity of the relationship between P and coagulation found in our study could be due to the several interactions between P and the other milk components and minerals, especially casein (i.e., organic phosphate linked to phosphoserine residues) and Ca (i.e., CCP).
      • Jensen H.B.
      • Poulsen N.A.
      • Andersen K.K.
      • Hammersh⊘j M.
      • Poulsen H.D.
      • Larsen L.B.
      Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
      reported differences in the proportions of soluble and micellar fractions of P between well and poorly coagulating milk samples from Jersey and Holstein-Friesian cows. In particular, micellar P was higher than soluble P in well compared with poorly coagulating Jersey milk samples, but in Holstein-Friesian cows they found only higher micellar P, but not lower soluble P, in well coagulating compared with poorly coagulating milk samples. The differences can probably be attributed to the different total casein contents and casein profiles of the 2 breeds. In fact, the cation binding ability of the casein fractions for the organic form of P decreases moving from αS2, αS1, β, to κ-casein, corresponding with their decreasing phosphoserine residues (
      • Lucey J.A.
      • Otter D.
      • Horne D.S.
      A 100-Year Review: Progress on the chemistry of milk and its components.
      ). In fact, caseins and whey proteins are the main mineral-binding components in milk. For example, αS2 and αS1 caseins bind Ca and Fe; β-casein, α-LA, and β-LG bind Ca, Zn, Mg, Mn, and Cu; and lactoferrin binds Fe and Zn (
      • Vegarud G.E.
      • Langsrud T.
      • Svenning C.
      Mineral-binding proteins and peptides; occurrence, biochemical and technological characteristics.
      ). A recent study on the detailed protein fractions of milk from 1,504 cows of the same breeds as in this study reported large differences among breeds in their protein profiles (caseins, whey proteins, and minor NPN compounds;
      • Amalfitano N.
      • Stocco G.
      • Maurmayr A.
      • Pegolo S.
      • Cecchinato A.
      • Bittante G.
      Quantitative and qualitative detailed milk protein profiles of 6 cattle breeds: Sources of variation and contribution of protein genetic variants.
      ). Because of the mineral-binding ability of casein and whey proteins, it would be very interesting to combine these data with data on minerals to further elucidate the effects of each single component and their interactions on coagulation and cheese-making traits.
      Although P did not seem to be strictly associated with milk coagulation, it exerted a large effect on the cheese-making traits (Table 3 and Figure 5). Unlike Ca, which increased CYCURD mainly through increased water retention, milk samples with elevated P concentrations showed higher CYSOLIDS and CYCURD followed the same trend, but was not significant (Figure 5a). This effect was not linear, but quadratic, with the highest values being for milk samples in the fourth quintile (1,048–1,100 mg/L). This pattern is a clear consequence of 2 different trends observed in milk fat and protein recovery in the curd. Milk samples with intermediate concentrations of P (second, third, and fourth quintiles) showed higher RECFAT (Figure 5b); RECPROTEIN, on the other hand, followed a linear pattern, with the highest values corresponding to the highest concentrations of P in the milk (Figure 5b). As expected, RECSOLIDS and RECENERGY followed a similar trend to CYSOLIDS, with the highest values corresponding to the fourth quintile, and a slight decrease in the last quintile of P (Figure 5c). A possible explanation for these nonlinear associations could lie in the interaction of P with the other milk components and minerals, especially Ca (formation of CCP). Calcium and inorganic phosphates are in dynamic equilibrium between the aqueous and micellar phases, and this equilibrium is influenced by the physico-chemical conditions of milk (e.g., pH, temperature), while modifications occurring between the aqueous and micellar phases affect the structure and stability of casein micelles (

      Gaucheron, F. 2013. Importance of the mineral fraction in dairy science and technology. IV SIMLEITE, October 2013, Vicosa, Brazil.

      ). This certainly affects the cheese-making process. For example, the lower RECFAT observed at the highest P concentrations could be explained by excessive mineralization of the casein micelle (high CCP content), which determines a reduction in the phosphate groups of caseins, and as a consequence, a reduction in the interaction between these groups and soluble ionic Ca during the second phase of the coagulation process (
      • Malacarne M.
      • Franceschi P.
      • Formaggioni P.
      • Sandri S.
      • Mariani P.
      • Summer A.
      Influence of micellar calcium and phosphorous on rennet coagulation properties of cows milk.
      ). Similarly, an excess of phosphates in soluble form could sequester soluble ionic Ca, leading to a weak coagulum that is no longer able to retain fat globules in the casein network.
      Figure thumbnail gr5
      Figure 5Effect of P on CYCURD and CYSOLIDS [calculated as the ratio of the weight (g) of fresh curd and curd DM, respectively, to the weight of the milk processed (g)] [a], RECFAT and RECPROTEIN [calculated as the ratio of the weight (g) of the component (fat and protein, respectively) in the curd to the weight of the corresponding component in the milk (g)] [b], RECSOLIDS [calculated as the ratio of the weight (g) of DM in the curd to the weight of the corresponding component in the milk (g)] and RECENERGY [determined by estimating the energy in the milk and in the curd using the equation proposed by
      • NRC
      Nutrient Requirements of Dairy Cattle.
      and converted into MJ/kg] [c] of individual milk samples, and the coefficient of determination (R2) of the regression. CY = cheese yield; REC = recovery of milk nutrients in the curd.

      Sodium

      Sodium is present in milk mainly in the aqueous phase, where it is free or weakly associated with ions of the opposite charge. Together with K and Cl, Na contributes to the ionic strength of milk (

      Gaucheron, F. 2013. Importance of the mineral fraction in dairy science and technology. IV SIMLEITE, October 2013, Vicosa, Brazil.

      ). Since it is in osmolar equilibrium between milk and blood, a higher milk Na concentration than normal is often indicative of an inflammatory process affecting the mammary gland, and is associated with increased solubilization of casein and proteolytic activity in milk (
      • El Zubeir I.E.M.
      • El Owni O.A.O.
      • Mohamed G.E.
      Effect of mastitis on macro-minerals of bovine milk and blood serum in Sudan.
      ;
      • Batavani R.A.
      • Asri S.
      • Naebzadeh H.
      The effect of subclinical mastitis on milk composition in dairy cows.
      ). Five of the 23 traits we studied were significantly affected by Na concentration, whether or not fat and casein were included in the statistical model. Milk Na was negatively associated with fat, but was not associated with casein content (Table 1). Moreover, Na content was influenced by Mg, and affected K and Mg contents (Table 1).
      The CFt curves across different concentrations of Na in milk are illustrated in Figure 6. Although Na was significant only on tmax, it is interesting that both the lowest and the highest levels of Na were associated with delayed RCT and RCTeq, and with the lowest a30 values. Milk samples with intermediate Na contents showed the most favorable coagulation and curd-firming patterns (Figure 6). The effect of the natural content of Na in milk on coagulation and curd firming was therefore not linear, but instead curvilinear.
      Figure thumbnail gr6
      Figure 6Effect of Na on gelation, curd firming, and syneresis of individual milk samples. Mineral concentrations are reported as mean values (mg/L) of each quintile of the distribution.
      Most of the previous studies have focused on the effect of adding NaCl to milk on the dissociation between Ca and P in the casein micelles (
      • Lucey J.A.
      • Fox P.F.
      Importance of calcium and phosphate in cheese manufacture: a review.
      ), and the coagulation properties of reconstituted (
      • Sbodio O.A.
      • Tercero E.J.
      • Coutaz R.
      • Revelli G.R.
      Effect of rennet and sodium chloride concentration on milk coagulation properties.
      ) or fresh pasteurized milk (
      • Awad S.
      Effect of sodium chloride and pH on the rennet coagulation and gel firmness.
      ).
      • Awad S.
      Effect of sodium chloride and pH on the rennet coagulation and gel firmness.
      showed that RCT slowed and CF decreased with increasing NaCl concentrations in milk. In this study, however, the high natural content of milk Na had only a marginal effect on coagulation, which could be due to the fact that we examined the native content instead of its addition, and investigated only the mineral Na and not the compound NaCl. The contribution of this mineral to coagulation and the cheese-making traits did not change when fat and casein were included in the statistical model (Figure 2). This is probably due to the fact that we did not sample any clinically mastitic cows, and sampled only a few cows with high SCC. The range of variation in Na here (281–488 mg/L) was much narrower than when mastitic milk was also included (
      • El Zubeir I.E.M.
      • El Owni O.A.O.
      • Mohamed G.E.
      Effect of mastitis on macro-minerals of bovine milk and blood serum in Sudan.
      ;
      • Batavani R.A.
      • Asri S.
      • Naebzadeh H.
      The effect of subclinical mastitis on milk composition in dairy cows.
      ). In goat milk, a high native content of NaCl (i.e., >319 mg/dL) impaired coagulation (i.e., slowed k20, decreased CF traits, and inhibited syneresis;
      • Stocco G.
      • Pazzola M.
      • Dettori M.L.
      • Cipolat-Gotet C.
      • Summer A.
      • Vacca G.M.
      Variation in caprine milk composition and coagulation as affected by udder health indicators.
      ) and the overall cheese-making process (
      • Stocco G.
      • Pazzola M.
      • Dettori M.L.
      • Paschino P.
      • Summer A.
      • Cipolat-Gotet C.
      • Vacca G.M.
      Effects of indirect indicators of udder health on nutrient recovery and cheese yield traits in goat milk.
      ).
      Regarding bovine cheese-making traits, native milk Na affected RECPROTEIN and RECSOLIDS. The former was about 2% lower in milk samples with a high Na content compared with samples with a low Na content, although the trend was not linear, but quadratic, whereas the trend for RECSOLIDS was rather erratic (Figure 7).
      Figure thumbnail gr7
      Figure 7Effect of Na on RECPROTEIN and RECSOLIDS [calculated as the ratio of the weight (g) of the component (protein and DM, respectively) in the curd to the weight of the corresponding component in the milk (g)] of individual milk samples, and the coefficient of determination (R2) of the regression. REC = recovery of milk nutrients in the curd.

      Potassium

      Potassium is a monovalent ion contributing a quarter of the osmolality of bovine milk together with Na and Cl (
      • Atkinson S.
      • Alston-Mills B.
      • Lönnerdal B.
      • Neville M.C.
      Major Minerals and Ionic Constituents of Human and Bovine Milks.
      ). Potassium balance closely interacts with glucose and electrolyte metabolism (
      • Berg M.
      • Plöntzke J.
      • Leonhard-Marek S.
      • Müller K.E.
      • Röblitz S.
      A dynamic model to simulate potassium balance in dairy cows.
      ), and its concentration in milk is regulated mostly by secretion mechanisms in the mammary cell. The dairy industry's use of K salts (e.g., KCl) is aimed at reducing the Na content of cheese (
      • Grummer J.
      • Bobowski N.
      • Karalus M.
      • Vickers Z.
      • Schoenfuss T.
      Use of potassium chloride and flavor enhancers in low sodium Cheddar cheese.
      ), but this practice is generally not favored because the salts tend to impart a bitter flavor to the cheese.
      • Bauland J.
      • Famelart M.H.
      • Bouhallab S.
      • Jeantet R.
      • Roustel S.
      • Faiveley M.
      • Croguennec T.
      Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
      reported that the addition of KCl to milk did not affect mineral partitioning between the colloidal and soluble phase, or the aggregation of casein micelles and curd firming. However, no studies are available on the effect of native milk K on coagulation and cheese making. Potassium interacts with casein and with the minerals P and Na, as can be seen in Table 1 and in Supplemental Figure S1. However, the correlation coefficient between K and casein was low, and indeed K has a weak affinity with caseins, as does Na (
      • Le Graet Y.
      • Brulé G.
      Les équilibres minéraux du lait: influence du pH et de la force ionique.
      ). According to our results, this mineral seemed not to have a specific role of its own during coagulation and the cheese-making process when fat, casein, and the other minerals were included in the model (Table 2, Table 3, Figure 2). In fact, when fat and casein were not included in the model, it was found to affect k20, a30, kCF, CYCURD, and CYWATER: all these traits worsened at increasing levels of K in the milk (Supplemental Table S2 and Supplemental Figure S2c). Given the general unfavorable association of K with casein and the concentrations of the other minerals, and that it was found to have an effect only after removing fat and casein from the model, we can speculate that the apparent contribution of K to coagulation and the cheese-making traits is instead attributable to casein and to the changes in the equilibrium of the other milk constituents and the overall mineral profile.

      Magnesium

      The technological importance of Mg in milk has been largely eclipsed by Ca, which plays an essential role in the structure and stability of casein micelles via CCP (
      • Oh H.E.
      • Deeth H.C.
      Magnesium in milk.
      ). However, these 2 minerals act cooperatively during coagulation, as they have different coupling sites on casein, and in particular, Ca aids the binding of Mg by making more casein sites available (
      • Cuomo F.
      • Ceglie A.
      • Lopez F.
      Temperature dependence of calcium and magnesium induced caseinate precipitation in H2O and D2O.
      ).
      • Bauland J.
      • Famelart M.H.
      • Bouhallab S.
      • Jeantet R.
      • Roustel S.
      • Faiveley M.
      • Croguennec T.
      Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
      evidenced that after addition of MgCl2, Mg was mainly exchanged with casein micelles through the bound form, whereas 70% of added Ca precipitated as CCP. In our study, we were able to disentangle the contribution of each mineral to coagulation and the cheese-making traits from the other minerals included in the model and milk composition. Unlike the other minerals, the effect of Mg was more evident when fat and casein were included in the statistical model (Figure 2). It seems that fat and casein, which are associated with Mg content (Table 1), masked the effect of this mineral. Magnesium was also influenced by Na content, and it affected Ca and Na (Table 1). Table 2, Table 3 show clearly that Mg had an effect on k20 and the 3 CY measures. In particular, moving from low to high levels of Mg in milk, a slight linear increase in k20 values was observed (more than 1 min difference between low and high Mg content). The results on the effect of Mg on coagulation traits reported in the literature are limited to the association between this mineral and the overall good or poor coagulation ability of milk.
      • Ketto I.A.
      • Knutsen T.M.
      • Øyaas J.
      • Heringstad B.
      • Ådn⊘y T.
      • Devold T.G.
      • Skeie S.B.
      Effects of milk protein polymorphism and composition, casein micelle size and salt distribution on the milk coagulation properties in Norwegian Red cattle.
      reported weak associations between Mg and the gel firming rate (r = 0.18, P < 0.01) and gel firmness at 30 min (r = 0.22, P < 0.01);
      • Frederiksen P.D.
      • Andersen K.K.
      • Hammersh⊘j M.
      • Poulsen H.D.
      • S⊘rensen J.
      • Bakman M.
      • Qvist K.B.
      • Larsen L.B.
      Composition and effect of blending of noncoagulating, poorly coagulating, and well-coagulating bovine milk from individual Danish Holstein cows.
      did not find any differences in Mg content between well and poorly coagulating milk samples, but
      • Jensen H.B.
      • Poulsen N.A.
      • Andersen K.K.
      • Hammersh⊘j M.
      • Poulsen H.D.
      • Larsen L.B.
      Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
      did find some differences in the milk of Holstein-Friesian cows.
      Regarding cheese-making traits, high levels of Mg were associated with reduced CYSOLIDS (about −0.2%) and CYWATER (the trend here was erratic), that consequently tended to reduce the total CYCURD (about −0.5%; data not shown). The correlation coefficient between Mg and casein found in our study (r = 0.62, P < 0.001; Supplemental Figure S1) was similar to that between Mg and protein reported by
      • Bijl E.
      • van Valenberg H.J.F.
      • Huppertz T.
      • van Hooijdonk A.C.M.
      Protein, casein, and micellar salts in milk: Current content and historical perspectives.
      ; r = 0.64, P < 0.01). It is interesting that this linear relationship was not accompanied by the same trend when Mg was associated with cheese-making traits. Because no published studies provide this type of information, we can only speculate that these results are related to different interactions with the other minerals (i.e., inorganic phosphates, Ca) and milk components (i.e., citrate, nanoclusters of casein micelles), and some of the enzymatic reactions in which Mg is involved (i.e., β-galactosidase, alkaline phosphatase activities;
      • Rankin S.A.
      • Christiansen A.
      • Lee W.
      • Banavara D.S.
      • Lopez-Hernandez A.
      Invited review: The application of alkaline phosphatase assays for the validation of milk product pasteurization.
      ;
      • Banerjee G.
      • Ray A.
      • Hasan K.N.
      Is divalent magnesium cation the best cofactor for bacterial β-galactosidase?.
      ).

      CONCLUSIONS

      The results presented here provide new knowledge about the relationships between the mineral contents, coagulation ability, and cheese-making traits of bovine milk. We found that a high Ca content had a positive effect on both the coagulation pattern and cheese-making traits, favoring water retention in the curd. Phosphorus positively affected the cheese-making traits, increasing CY in terms of curd solids, and all the nutrient recovery traits, although a very high P content in milk was associated with less fat recovered in the curd. The variation in the Na content of milk only mildly affected coagulation, whereas protein recovery was negatively associated with high concentrations of this mineral, probably reflecting the association with subclinical mastitis. The role of K during coagulation and the cheese-making process seemed to be more passive and linked to milk composition and the overall milk salt equilibrium, whereas high Mg content tended to slow coagulation and reduce CY traits.

      ACKNOWLEDGMENTS

      The authors thank the Autonomous Province of Trento (Italy) for funding. The authors have stated no conflicts of interest.

      REFERENCES

        • Akkerman M.
        • Larsen L.B.
        • S⊘rensen J.
        • Poulsen N.A.
        Natural variations of citrate and calcium in milk and their effects on milk processing properties.
        J. Dairy Sci. 2019; 102 (31155267): 6830-6841
        • Amalfitano N.
        • Stocco G.
        • Maurmayr A.
        • Pegolo S.
        • Cecchinato A.
        • Bittante G.
        Quantitative and qualitative detailed milk protein profiles of 6 cattle breeds: Sources of variation and contribution of protein genetic variants.
        J. Dairy Sci. 2020; 103 (33069399): 11190-11208
        • Amenu B.
        • Deeth H.C.
        The impact of milk composition on cheddar cheese manufacture.
        Aust. J. Dairy Technol. 2007; 62: 171-184
        • Atkinson S.
        • Alston-Mills B.
        • Lönnerdal B.
        • Neville M.C.
        Major Minerals and Ionic Constituents of Human and Bovine Milks.
        in: Jensen R.G. Food Science and Technology, Handbook of Milk Composition. Academic Press, 1995: 593-622
        • Auldist M.J.
        • Johnston K.A.
        • White N.J.
        • Fitzsimons W.P.
        • Boland M.J.
        A comparison of the composition, coagulation characteristics and cheesemaking capacity of milk from Friesian and Jersey dairy cows.
        J. Dairy Res. 2004; 71 (15068067): 51-57
        • Awad S.
        Effect of sodium chloride and pH on the rennet coagulation and gel firmness.
        Lebensm. Wiss. Technol. 2007; 40: 220-224
        • Banerjee G.
        • Ray A.
        • Hasan K.N.
        Is divalent magnesium cation the best cofactor for bacterial β-galactosidase?.
        J. Biosci. 2018; 43 (30541954): 941-945
        • Batavani R.A.
        • Asri S.
        • Naebzadeh H.
        The effect of subclinical mastitis on milk composition in dairy cows.
        Majallah-i Tahqiqat-i Dampizishki-i Iran. 2007; 8: 205-211
        • Bauland J.
        • Famelart M.H.
        • Bouhallab S.
        • Jeantet R.
        • Roustel S.
        • Faiveley M.
        • Croguennec T.
        Addition of calcium and magnesium chlorides as simple means of varying bound and precipitated minerals in casein micelle: Effect on enzymatic coagulation.
        J. Dairy Sci. 2020; 103 (32921475): 9923-9935
        • Berg M.
        • Plöntzke J.
        • Leonhard-Marek S.
        • Müller K.E.
        • Röblitz S.
        A dynamic model to simulate potassium balance in dairy cows.
        J. Dairy Sci. 2017; 100 (28987591): 9799-9814
        • Berton M.
        • Bittante G.
        • Zendri F.
        • Ramanzin M.
        • Schiavon S.
        • Sturaro E.
        Environmental impact and efficiency of use of resources of different mountain dairy farming systems.
        Agric. Syst. 2020; 181102806
        • Bijl E.
        • van Valenberg H.J.F.
        • Huppertz T.
        • van Hooijdonk A.C.M.
        Protein, casein, and micellar salts in milk: Current content and historical perspectives.
        J. Dairy Sci. 2013; 96 (23849643): 5455-5464
        • Bittante G.
        • Contiero B.
        • Cecchinato A.
        Prolonged observation and modelling of milk coagulation, curd firming, and syneresis.
        Int. Dairy J. 2013; 29: 115-123
        • Bland J.H.
        • Grandison A.S.
        • Fagan C.C.
        Evaluation of milk compositional variables on coagulation properties using partial least squares.
        J. Dairy Res. 2015; 82 (25287607): 8-14
        • Bobbo T.
        • Ruegg P.L.
        • Stocco G.
        • Fiore E.
        • Gianesella M.
        • Morgante M.
        • Pasotto D.
        • Bittante G.
        • Cecchinato A.
        Associations between pathogen-specific cases of subclinical mastitis and milk yield, quality, protein composition, and cheese-making traits in dairy cows.
        J. Dairy Sci. 2017; 100 (28365113): 4868-4883
        • Cipolat-Gotet C.
        • Cecchinato A.
        • De Marchi M.
        • Bittante G.
        Factors affecting variation of different measures of cheese yield and milk nutrients recovery from an individual model cheese manufacturing process.
        J. Dairy Sci. 2013; 96 (24094531): 7952-7965
        • Cipolat-Gotet C.
        • Cecchinato A.
        • Malacarne M.
        • Bittante G.
        • Summer A.
        Variations in milk protein fractions affect the efficiency of the cheese-making process.
        J. Dairy Sci. 2018; 101 (30122415): 8788-8804
        • Cipolat-Gotet C.
        • Malacarne M.
        • Summer A.
        • Cecchinato A.
        • Bittante G.
        Modeling weight loss of cheese during ripening and the influence of dairy system, parity, stage of lactation, and composition of processed milk.
        J. Dairy Sci. 2020; 103 (32475671): 6843-6857
        • Cooke D.R.
        • McSweeney P.L.H.
        The influence of alkaline earth metal equilibria on the rheological properties of rennet-induced skim milk gels.
        Dairy Sci. Technol. 2014; 94: 341-357
        • Cuomo F.
        • Ceglie A.
        • Lopez F.
        Temperature dependence of calcium and magnesium induced caseinate precipitation in H2O and D2O.
        Food Chem. 2011; 126: 8-14
        • Dalgleish D.G.
        • Corredig M.
        The structure of the casein micelle of milk and its changes during processing.
        Annu. Rev. Food Sci. Technol. 2012; 3 (22385169): 449-467
        • El Zubeir I.E.M.
        • El Owni O.A.O.
        • Mohamed G.E.
        Effect of mastitis on macro-minerals of bovine milk and blood serum in Sudan.
        J. S. Afr. Vet. Assoc. 2005; 76 (15900896): 22-25
        • Fox P.F.
        • Guinee T.P.
        • Cogan T.M.
        • McSweeney P.L.H.
        Cheese: Structure, rheology and texture.
        in: Fundamentals of Cheese Science. 2nd ed. Springer, 2017: 475-532
        • Fox P.F.
        • Guinee T.P.
        • Cogan T.M.
        • McSweeney P.L.H.
        Cheese yield.
        in: Fundamentals of Cheese Science. 2nd ed. Springer, 2017: 279-331
        • Frederiksen P.D.
        • Andersen K.K.
        • Hammersh⊘j M.
        • Poulsen H.D.
        • S⊘rensen J.
        • Bakman M.
        • Qvist K.B.
        • Larsen L.B.
        Composition and effect of blending of noncoagulating, poorly coagulating, and well-coagulating bovine milk from individual Danish Holstein cows.
        J. Dairy Sci. 2011; 94 (21943730): 4787-4799
      1. Gaucheron, F. 2013. Importance of the mineral fraction in dairy science and technology. IV SIMLEITE, October 2013, Vicosa, Brazil.

        • Ghiroldi S.
        • Nicoletti C.
        • Santus E.
        • Rossoni A.
        • Bagnato A.
        ITE: The new selection index for the Italian Brown Swiss.
        Interbull Bull. 2005; 33: 222
        • Grummer J.
        • Bobowski N.
        • Karalus M.
        • Vickers Z.
        • Schoenfuss T.
        Use of potassium chloride and flavor enhancers in low sodium Cheddar cheese.
        J. Dairy Sci. 2013; 96 (23332837): 1401-1418
        • Gustavsson F.
        • Glantz M.
        • Buitenhuis A.J.
        • Lindmark-Månsson H.
        • Stålhammar H.
        • Andrén A.
        • Paulsson M.
        Factors influencing chymosin-induced gelation of milk from individual dairy cows: major effects of casein micelle size and calcium.
        Int. Dairy J. 2014; 39: 201-208
        • ISO-IDF (International Organization for Standardization and International Dairy Federation)
        Milk - Determination of casein-nitrogen content - Part 1: Indirect method. International Standard ISO 17997–1 and IDF 29–1:2004.
        ISO and IDF, 2004
        • ISO-IDF (International Organization for Standardization and International Dairy Federation)
        Milk - Determination of fat content. International Standard ISO 1211 and IDF 1:2010.
        ISO and IDF, 2010
        • ISO-IDF (International Organization for Standardization and International Dairy Federation)
        Milk, cream and evaporated milk - Determination of total solids content. International standard ISO 6731 and IDF 21:2010.
        ISO and IDF, 2010
        • ISO-IDF (International Organization for Standardization and International Dairy Federation)
        Milk - Determination of lactose content - Enzymatic method using difference in pH. International Standard ISO 26462:2010 and IDF 214:2010.
        ISO and IDF, 2010
        • ISO-IDF (International Organization for Standardization and International Dairy Federation)
        Milk and milk products - Determination of nitrogen content - Part 1: Kjeldahl principle and crude protein calculation. International Standard ISO 8968–1 and IDF 1:2014.
        ISO and IDF, 2014
        • Jensen H.B.
        • Poulsen N.A.
        • Andersen K.K.
        • Hammersh⊘j M.
        • Poulsen H.D.
        • Larsen L.B.
        Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
        J. Dairy Sci. 2012; 95 (23040012): 6905-6917
        • Kaufmann W.
        • Hagemeister H.
        Composition of milk.
        in: Gravert H.O. Dairy-Cattle Production. Elsevier, 1987: 107-171
        • Ketto I.A.
        • Knutsen T.M.
        • Øyaas J.
        • Heringstad B.
        • Ådn⊘y T.
        • Devold T.G.
        • Skeie S.B.
        Effects of milk protein polymorphism and composition, casein micelle size and salt distribution on the milk coagulation properties in Norwegian Red cattle.
        Int. Dairy J. 2017; 70: 55-64
        • Le Graet Y.
        • Brulé G.
        Les équilibres minéraux du lait: influence du pH et de la force ionique.
        Lait. 1993; 73: 51-60
        • Leitner G.
        • Lavon Y.
        • Matzrafi Z.
        • Benun O.
        • Bezman D.
        • Merin U.
        Somatic cell counts, chemical composition and coagulation properties of goat and sheep bulk tank milk.
        Int. Dairy J. 2016; 58: 9-13
        • Lucey J.A.
        • Fox P.F.
        Importance of calcium and phosphate in cheese manufacture: a review.
        J. Dairy Sci. 1993; 76: 1714-1724
        • Lucey J.A.
        • Horne D.S.
        Milk salts: Technological significance.
        in: Advanced Dairy Chemistry, Volume 3: Lactose, Water, Salts and Minor Constituents. 3rd ed. Springer Science+Buisness Media, 2009: 351-390
        • Lucey J.A.
        • Otter D.
        • Horne D.S.
        A 100-Year Review: Progress on the chemistry of milk and its components.
        J. Dairy Sci. 2017; 100 (29153180): 9916-9932
        • Malacarne M.
        • Franceschi P.
        • Formaggioni P.
        • Sandri S.
        • Mariani P.
        • Summer A.
        Influence of micellar calcium and phosphorous on rennet coagulation properties of cows milk.
        J. Dairy Res. 2014; 81 (24345431): 129-136
        • Miglior F.
        • Muir B.L.
        • Van Doormaal B.J.
        Selection indices in Holstein cattle of various countries.
        J. Dairy Sci. 2005; 88 (15738259): 1255-1263
        • NRC
        Nutrient Requirements of Dairy Cattle.
        7th rev. ed. Natl. Acad. Press, 2001
        • Oh H.E.
        • Deeth H.C.
        Magnesium in milk.
        Int. Dairy J. 2017; 71: 89-97
        • Pazzola M.
        • Stocco G.
        • Dettori M.L.
        • Bittante G.
        • Vacca G.M.
        Effect of goat milk composition on cheese-making traits and daily cheese production.
        J. Dairy Sci. 2019; 102 (30827544): 3947-3955
        • Pirisi A.
        • Lauret A.
        • Dubeuf J.P.
        Basic and incentive payments for goat and sheep milk in relation to quality.
        Small Rumin. Res. 2007; 68: 167-178
        • Pryce J.E.
        • van der Werf J.H.J.
        • Haile-Mariam M.
        • Malcolm B.
        • Goddard M.E.
        Updated index weights for the Australian Profit Ranking in dairy cattle.
        Proc. Assoc. Adv. Anim. Breed. Genet. 2009; 18: 143-146
        • Rankin S.A.
        • Christiansen A.
        • Lee W.
        • Banavara D.S.
        • Lopez-Hernandez A.
        Invited review: The application of alkaline phosphatase assays for the validation of milk product pasteurization.
        J. Dairy Sci. 2010; 93 (21094726): 5538-5551
        • Sbodio O.A.
        • Tercero E.J.
        • Coutaz R.
        • Revelli G.R.
        Effect of rennet and sodium chloride concentration on milk coagulation properties.
        CYTA J. Food. 2006; 5: 182-188
        • Sneddon N.W.
        • Lopez-Villalobos N.
        • Hickson R.E.
        • Shalloo L.
        Review of milk payment systems to identify the component value of lactose.
        Proc. N.Z. Soc. Anim. Prod. 2013; 73: 33-36
        • Stocco G.
        Supplemental Table S1. Figshare. Dataset.
        • Stocco G.
        Supplemental Table S2. Figshare. Dataset.
        • Stocco G.
        Supplemental Figure S1. Figshare. Figure.
        • Stocco G.
        Supplemental Figure S2. Figshare. Figure.
        • Stocco G.
        • Cipolat-Gotet C.
        • Bobbo T.
        • Cecchinato A.
        • Bittante G.
        Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
        J. Dairy Sci. 2017; 100 (27837976): 129-145
        • Stocco G.
        • Cipolat-Gotet C.
        • Gasparotto V.
        • Cecchinato A.
        • Bittante G.
        Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
        Animal. 2018; 12 (28712377): 434-444
        • Stocco G.
        • Pazzola M.
        • Dettori M.L.
        • Cipolat-Gotet C.
        • Summer A.
        • Vacca G.M.
        Variation in caprine milk composition and coagulation as affected by udder health indicators.
        Int. Dairy J. 2019; 98: 9-16
        • Stocco G.
        • Pazzola M.
        • Dettori M.L.
        • Paschino P.
        • Summer A.
        • Cipolat-Gotet C.
        • Vacca G.M.
        Effects of indirect indicators of udder health on nutrient recovery and cheese yield traits in goat milk.
        J. Dairy Sci. 2019; 102 (31351732): 8648-8657
        • Stocco G.
        • Summer A.
        • Malacarne M.
        • Cecchinato A.
        • Bittante G.
        Detailed macro- and micromineral profile of milk: Effects of herd productivity, parity, and stage of lactation of cows of 6 dairy and dual-purpose breeds.
        J. Dairy Sci. 2019; 102 (31477292): 9727-9739
        • Troch T.
        • Lefébure É.
        • Baeten V.
        • Colinet F.
        • Gengler N.
        • Sindic M.
        Cow milk coagulation: Process description, variation factors and evaluation methodologies. A review.
        Biotechnol. Agron. Soc. Environ. 2017; 21: 276-287
        • Tsioulpas A.
        • Lewis M.J.
        • Grandison A.S.
        Effect of minerals on casein micelle stability of cows' milk.
        J. Dairy Res. 2007; 74 (17291390): 167-173
        • Vegarud G.E.
        • Langsrud T.
        • Svenning C.
        Mineral-binding proteins and peptides; occurrence, biochemical and technological characteristics.
        Br. J. Nutr. 2000; 84 (11242452): S91-S98
        • Verdier-Metz I.
        • Coulon J.-B.
        • Pradel P.
        Relationship between milk fat and protein contents and cheese yield.
        Anim. Res. 2001; 50: 365-371
        • Wedholm A.
        • Larsen L.B.
        • Lindmark-Månsson H.
        • Karlsson A.H.
        • Andrén A.
        Effect of protein composition on the cheese-making properties of milk from individual dairy cows.
        J. Dairy Sci. 2006; 89 (16899662): 3296-3305