If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Teagasc Animal and Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork, IrelandDepartment of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE), University of Padova, Viale dell'Universita 16, 35020 Legnaro (PD), Italy
Department of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE), University of Padova, Viale dell'Universita 16, 35020 Legnaro (PD), Italy
Teagasc Animal and Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork, IrelandDepartment of Agronomy, Food, Natural Resources, Animals and Environment (DAFNAE), University of Padova, Viale dell'Universita 16, 35020 Legnaro (PD), Italy
The objective of the present study was to identify the factors associated with both the protein composition and free amino acid (FAA) composition of bovine milk predicted using mid-infrared spectroscopy. Milk samples were available from 7 research herds and 69 commercial herds. The spectral data from the research herds comprised 94,286 separate morning and evening milk samples; the spectral data from the commercial herds comprised 40,260 milk samples representing a composite sample of both the morning and evening milkings. Mid-infrared spectroscopy prediction models developed in a previous study were applied to all spectra. Factors associated with the predicted protein and FAA composition were quantified using linear mixed models. Factors considered in the model included the fixed effects of calendar month of the test, milking time (i.e., morning, evening, or both combined), parity (1, 2, 3, 4, 5, and ≥6), stage of lactation, the interaction between parity and stage of lactation, breed proportion of the cow (Friesian, Jersey, Norwegian Red, Montbéliarde, and other), and both the general heterosis and recombination coefficients of the cow. Contemporary group as well as both within- and across-lactation permanent environmental effects were included in all models as random effects. Total proteins (i.e., total casein, CN; total whey; and total β-lactoglobulin) and protein fractions (with the exception of α-lactalbumin) decreased postcalving until 36 to 65 days in milk and increased thereafter. After adjusting the statistical model for differences in crude protein content and milk yield separately, irrespective of stage of lactation, younger animals produced more total proteins (i.e., total CN, total whey, and total β-lactoglobulin) as well as more total FAA, Glu, and Asp than their older contemporaries. The concentration of all protein fractions (except β-CN) in milk was greatest in the evening milk, even after adjusting for differences in the crude protein content of the milk. Relative to a purebred Holstein cow, Jersey cows, on average, produced a greater concentration of all CN fractions but less total FAA, Glu, Gly, Asp, and Val in milk. Relative to their respective purebred parental average, first-cross cows produced more total CN and more β-CN. Results from the present study indicate that many cow-level factors, as well as other factors, are associated with protein composition and FAA composition of bovine milk.
Total protein content and its composition in bovine milk are among the most important milk characteristics for the dairy industry. In Europe, milk processors pay a greater premium for milk exceeding a threshold protein content than for milk exceeding a threshold fat content (
). Bovine milk generally consists of approximately 3.3% CP, of which 78% is CN, 17 to 18% is whey protein, and the remaining 4 to 5% is NPN (predominantly urea;
Associations of stage of lactation, milk protein genotype and body condition at calving on protein composition and renneting properties of bovine milk.
). Changes in the concentration of individual protein fractions in milk affect various processing attributes of the milk, including rennet coagulating time (
Effectiveness of mid-infrared spectroscopy for the prediction of detailed protein composition and contents of protein genetic variants of individual milk of Simmental cows.
Associations of stage of lactation, milk protein genotype and body condition at calving on protein composition and renneting properties of bovine milk.
) are associated with the individual protein fractions of milk. For example, mean αS-CN was reported to increase between first- and third-parity cows but plateaued thereafter, whereas β-CN decreased as parity number increased (
). For milk processing purposes, a high level of FAA is undesirable because FAA are a result of deproteinization and an indication of poorer milk quality. Therefore, bovine milk is less suitable for processing in early and late lactation, when total FAA are in greatest concentration (
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
). Nevertheless, to date, no study has investigated the variability in FAA across different parities, milking time of the day, month of the year, or dairy breeds.
Mid-infrared spectroscopy (MIRS) is commonly used worldwide to predict milk fat, protein, CN, and lactose of individual animal and bulk tank milk samples. Previous studies propose MIRS as a rapid and cost-effective analytic tool for recording phenotypes at the population level (
). The ability of MIRS to predict milk technological traits, detailed protein composition (αS1-CN, β-CN, κ-CN, α-LA, β-LG A, and β-LG B), FAA, and milk color characteristics has been previously documented (
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
Effectiveness of mid-infrared spectroscopy to predict the color of bovine milk and the relationship between milk color and traditional milk quality traits.
). The objective of the present study was to quantify the associations between cow-level factors, as well as other factors, with detailed protein and FAA composition of bovine milk predicted using MIRS.
MATERIALS AND METHODS
Spectral Data
Milk samples were available from 2 sources: 7 research herds operated by the Animal and Grassland Research and Innovation Center (Teagasc, Moorepark, Fermoy, Co. Cork, Ireland) and 69 Irish commercial dairy herds. Spectra from the research herds comprised 126,845 separate morning and evening milk samples from 2,535 lactations and 1,439 cows. Spectra from the commercial herds comprised 44,976 milk samples (morning and evening milk samples combined) from 14,874 lactations and 8,733 cows. Milk samples were from 2 seasonal calving systems (spring and autumn). Milk chemical composition (milk fat, protein, CN, and lactose concentration) was predicted for all milk samples using the same Fourier transform infrared spectrometer (Foss MilkoScan FT6000; Foss Electronic A/S, Hillerød, Denmark) based at the Animal and Grassland Research and Innovation Center. The generated spectrum, containing 1,060 transmittance data in the mid-infrared region between 900 and 5,000 cm−1, was stored. Mid-infrared spectroscopy models were developed using partial least squares regressions (PROC PLS; SAS Institute Inc., Cary, NC) with untreated spectra as described in detail by
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
. Spectral regions from 926 to 1,580 cm−1, 1,717 to 2,986 cm−1, and 3,696 to 3,808 cm−1 were used to develop all prediction models based on the observed loadings for each wavelength. In brief, between the years 2013 and 2014, a calibration data set was generated using 715 individual milk samples from the same 7 research herds used in the present study. Milk samples from 345 cows appeared in both data sets. Spectral outliers were determined as milk samples with a Mahalanobis distance greater than 3 (
) relative to the mean of the calibration data set. Prediction models were developed using 557 reference values for individual proteins and up to 715 reference values for FAA determined by HPLC. The accuracy (i.e., the coefficient of correlation) of prediction from (1) split-sample cross-validation and (2) external validation on an independent 25% of the data (not included in model calibration), as reported by
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
, is provided in AppendixTable A1. The accuracy of prediction in external validation was, on average, moderate; the coefficient of determination for protein fractions ranged from 0.39 (β-LG A) to 0.74 (total CN) and for FAA ranged from 0.26 (Arg) to 0.75 (Gly;
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
to develop the prediction equations were considered as spectral outliers (1,634 milk samples in total) and discarded. Furthermore, predicted values of proteins and FAA that were >3 SD from the mean of the respective reference samples were considered to be outliers and removed as well. Milk yield over a 24-h period was available for the commercial cows; therefore, milk yield over a 24-h period was computed for the research cows as the sum of their morning and evening milk yields from the same day. Only records between 5 and 305 DIM and from parities ≤10 were retained for analysis; parities >5 were grouped together for analysis.
Contemporary group of experimental treatment by test date was defined for milk samples from cows in research herds, whereas contemporary group of herd by test date was defined for milk samples from cows in commercial herds. Only records for contemporary groups with at least 10 records were retained for analysis. The research and commercial data sets were combined for analysis. After editing, the final data set comprised 134,546 milk spectra from 9,572 cows.
Pedigree data and breed composition of all animals were available from the national database managed by the Irish Cattle Breeding Federation (http://www.icbf.com). Only milk samples from Holstein (HO), Friesian (FR), Jersey (JE), Norwegian Red (NR), and Montbéliarde (MO) cows as well as their respective crosses (including progeny from crossbred parents of these breeds) were retained for analysis. The data consisted of 6,724 purebred cows (i.e., ≥75% pure), 2,848 crossbred cows, and 1,853 cows from crossbred parents. Table 1 lists the number of records, cows, and lactations and average parity of each breed and crossbreed. Coefficients of heterosis and recombination loss were calculated for each cow as
and
where sirei and dami are the proportion of genes of the breed i in the sire and the dam, respectively (
HO = Holstein; FR = Friesian; JE = Jersey; NR = Norwegian Red; MO = Montbéliarde; × indicates a cross of varying proportions of each breed. A purebred animal was deemed to be ≥75% of the breed.
n
Cows
Lactations
Parity
HO
62,653
6,562
10,568
2.7
FR
288
31
48
2.6
JE
4,872
63
107
2.22
NR
179
13
21
4.94
MO
135
55
81
4.46
HO × FR
34,690
1,826
2,954
2.68
HO × JE
24,204
488
851
2.49
HO × NR
2,280
271
501
2.07
HO × MO
728
122
188
3.65
JE × FR
2,823
67
117
2.79
JE × NR
1,518
51
102
2.25
JE × MO
51
1
1
1
NR × FR
75
12
20
2.29
MO × FR
50
10
17
4.28
1 HO = Holstein; FR = Friesian; JE = Jersey; NR = Norwegian Red; MO = Montbéliarde; × indicates a cross of varying proportions of each breed. A purebred animal was deemed to be ≥75% of the breed.
). Factors considered in the model included the fixed effects of calendar month of milk test, milking time of the day (morning, evening, or both combined), parity (1, 2, 3, 4, 5, and ≥6), stage of lactation (in 30-d intervals), an interaction between parity and stage of lactation, breed proportion of the cow fitted as separate covariates (FR, JE, NR, MO, and other), and general heterosis and recombination loss coefficients of the cow. Holstein breed proportion was not included in the model to avoid linear dependencies; therefore, breed solutions reported are relative to an HO cow. The random effects of contemporary group as well as within- and across-lactation effects were included in all models. Least squares means were estimated based on a reference cow represented as a 100% HO, third-parity cow, milked in the morning and averaged across stages of lactation and calendar months of the year of test. In a separate series of analyses, models with a protein fraction as the dependent variable were adjusted for total milk protein content (i.e., included as a covariate). Models with FAA as the dependent variable were adjusted for 24-h milk yield by including 24-h milk yield as a covariate in the models.
RESULTS
Descriptive Statistics
Mean predicted values of all milk traits in the research and commercial herds are summarized in Table 2. Mean values were similar for both the research and commercial herds. Mean values of total CN (αS1-CN, αS2-CN, β-CN, and κ-CN) and total whey (α-LA, β-LG A, and β-LG B) for both research and commercial herds were approximately in the ratio 6:1. Individual CN fractions (αS1-CN, αS2-CN, β-CN, and κ-CN) were present in the ratio 4:1:4:2 for both the research and commercial herds. The coefficient of variation differed among traits and ranged from 10% (β-CN) to 51% (β-LG-B) for the protein fractions. The FAA present in the greatest quantity in the milk was Glu, which represented 57% of total FAA (Glu, Gly, Lys, Arg, Asp, Ser, and Val) in the milk. Contemporary group accounted for between 45.34% (total FAA) and 86.11% (α-LA) of the variability in the traits investigated.
Table 2Number of records (n), mean (SD), and CV of the studied traits predicted using mid-infrared spectroscopy in research and commercial herds
Evening milk had a greater (P < 0.001) concentration of αS1-CN, αS2-CN, total whey protein, α-LA, total β-LG, and β-LG A but a reduced (P < 0.01) concentration of β-CN (13.60 vs. 13.75 g/L) compared with morning milk when adjusted for CP content (Table 3). Furthermore, although evening milk has more (P < 0.01) β-CN than morning milk, the biological difference was small. Evening milk had a greater (P < 0.001) concentration of all FAA (total FAA, Glu, Gly, Lys, Arg, Asp, and Val) except Ser compared with morning milk when adjusted for milk yield (Table 3).
Table 3Least squares means (SE in parentheses) of individual protein fractions adjusted for CP content and of individual free AA (μg/mL of milk) adjusted for milk yield in both morning and evening milkings
The observed interaction between stage of lactation and parity on the concentration of protein fractions persisted regardless of whether adjustments were made in the statistical model for either differences in CP content or 24-h milk yield (results not shown). When adjusted for CP content, total CN and protein fractions (except for α-LA) decreased postcalving to between 36 and 65 DIM across all parities but gradually increased thereafter (Figures 1, 2, and 3). In younger cows, α-LA in milk decreased between 5 and 155 DIM (P < 0.05) and then plateaued when adjusted for CP content; however, in older cows, α-LA in milk remained constant until mid to late lactation, after which it decreased in concentration (Figure 3). Across all stages of lactation, younger animals produced milk with a greater concentration of total CN (P < 0.05), total whey (P < 0.001), and total β-LG (P < 0.001) than their older contemporaries when adjusted for CP content. Younger cows produced more αS1-CN (P < 0.01) and β-CN (P < 0.001) in milk than older cows, but first-parity cows produced less αS2-CN (0.001) and κ-CN (0.01) than multiparous cows when adjusted for CP content.
Figure 1Trends in concentration of proteins in milk (a) total protein, (b) total CN, (c) total whey, and (d) total β-LG adjusted for CP content across stage of lactation for animals in parity 1 (•), parity 2 (▪), parity 3 (▴), parity 4 (○), parity 5 (□), and parity ≥6 (Δ). Error bars represent the mean SE across parities.
Figure 2Trends in concentrations of CN fractions in milk (a) αS1-CN, (b) αS2-CN, (c) β-CN, and (d) κ-CN adjusted for CP content across stage of lactation for animals in parity 1 (•), parity 2 (▪), parity 3 (▴), parity 4 (○), parity 5 (□), and parity ≥6 (Δ). Error bars represent the mean SE across parities.
Figure 3Trends in concentrations of whey proteins in milk (a) α-LA, (b) β-LG A, and (c) β-LG B adjusted for CP content across stage of lactation for animals in parity 1 (•), parity 2 (▪), parity 3 (▴), parity 4 (○), parity 5 (□), and parity ≥6 (Δ). Error bars represent the mean SE across parities.
Figure 4 illustrates the interaction between stage of lactation and parity (P < 0.001) on total FAA, Glu, Gly, and Lys concentration in milk adjusted for milk yield. Irrespective of cow parity, Lys and Val concentration decreased in milk until 36 to 65 DIM; subsequently, Lys concentration plateaued and Val concentration continued to increase across stage of lactation. Total FAA and Gly concentration decreased from 5 to 125 DIM, after which total FAA continued to decrease across stage of lactation in earlier parities but plateaued in later parities and Gly concentration plateaued irrespective of parity. Across stage of lactation, younger cows had a greater (P < 0.001) concentration of total FAA, Glu, and Asp in milk compared with older contemporaries. The concentration of Gly was the same across parities (P < 0.05), and the concentration of Lys and Arg was lower in earlier parities compared with later parities (P < 0.001).
Figure 4Trends in total and individual free AA (FAA) in milk (a) total FAA, (b) Glu, (c) Gly, (d) Lys, (e) Arg, (f) Asp, (g) Ser, and (h) Val adjusted for milk yield across stage of lactation for animals in parity 1 (•), parity 2 (▪), parity 3 (▴), parity 4 (○), parity 5 (□), and parity ≥6 (Δ). Error bars represent the mean SE across parities.
After adjusting for CP content, a peak in the concentration of all CN fractions was evident in the months of August, September, and October (Figure 5; P < 0.001), whereas the concentration of α-LA remained relatively constant across the year (P < 0.05). The concentration of Glu was greater (P < 0.001) during the months of February, March, April, and June, whereas the concentration of Gly was greater (P < 0.001) during the months of February, March, and June when adjusted for milk yield. The change in the concentration of Asp, Ser, and Val across calendar month of the year adjusted for milk yield was small but significant (P < 0.05).
Figure 5(a) Least squares means of concentrations (g/L of milk) of CN fractions [primary vertical axis; αS1-CN (•), αS2-CN (▪), β-CN (▴), and κ-CN (♦)] and whey fractions [secondary vertical axis; α-LA (○), β-LG A (□), and β-LG B (Δ)] adjusted for total milk protein content across calendar month of the year. (b) Least squares means of concentrations (µg/mL of milk) of Glu (•; primary vertical axis) and Gly (▪), Lys (▴), Arg (♦), Asp (○), Ser (□), and Val (Δ; all on the secondary vertical axis) in milk adjusted for milk yield across calendar month of the year. Error bars represent SE.
Table 4 provides breed regression coefficient estimates for concentration of proteins adjusted for CP content expressed relative to a purebred HO for FR, JE, NR, and MO breeds and associated heterosis and recombination estimates. Milk produced by JE cows had the greatest concentration of all CN fractions (P < 0.001), and JE cows produced milk with 3.91, 3.14, 2.86, and 4.64 g/L more total CN than HO, FR, NR, and MO cows, respectively. Also, JE cows produced milk that had a greater concentration of the CN fractions (αS1-, αS2-, β-, and κ-CN) in addition to a greater concentration of total whey, α-LA, total β-LG, and β-LG A relative to HO cows. The concentration of total whey protein in milk of JE cows was 0.42, 0.23, 0.42, and 0.53 g/L greater compared with the milk of HO, FR, NR, and MO cows, respectively. Milk produced by JE cows had less (P < 0.001) total FAA, Glu, Gly, and Asp than any other breed of cow, including HO (Table 5).
Table 4Breed regression coefficient estimates (SE in parentheses) for concentration of proteins (g/L of milk) adjusted for CP content expressed relative to a purebred Holstein (HO) for Friesian (FR), Jersey (JE), Norwegian Red (NR), and Montbéliarde (MO) breeds and associated heterosis and recombination estimates
Table 5Breed regression coefficient estimates (SE in parentheses) for concentration of free AA (μg/mL of milk) adjusted for milk yield expressed relative to a purebred Holstein (HO) for Friesian (FR), Jersey (JE), Norwegian Red (NR), and Montbéliarde (MO) breeds and associated heterosis and recombination estimates
Both heterosis and recombination estimates for all traits were small in magnitude. Relative to the purebred parent average, first-cross cows produced milk that had 0.27 g/L more total CN and 0.13 g/L more β-CN. Positive recombination estimates were observed for all protein fractions in milk. Heterosis estimates for all FAA (except Lys) in milk were not different from zero, and recombination estimates for all FAA were not different from zero.
DISCUSSION
The objective of the present study was to determine the factors associated with protein composition and FAA composition of bovine milk. In the present study, mean values of αS1-, αS2-, β-, and κ-CN were in the ratio 4:1:4:2, which was not in agreement with results by
for 1,336 Simmental cows only; the present study, however, contained records from cows of multiple breeds and crossbreds. Total mean proteins, determined using HPLC (42.53 g/L for the research herds; 41.73 g/L for the commercial herds), were higher than those recorded using MIRS (37.45 g/L for the research herds; 36.67 g/L for the commercial herds). This is most likely attributable to cumulative variation during summation of the individual protein values when integrating peak areas from the HPLC data (
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
Effectiveness of mid-infrared spectroscopy for the prediction of detailed protein composition and contents of protein genetic variants of individual milk of Simmental cows.
also reported high mean protein values using HPLC (i.e., up to 40.12 and 40.68 g/L, respectively) in milk from Simmental cows.
Similar trends across stage of lactation for αS1-CN, β-CN, β-LG, and α-LA in milk adjusted for CP content (i.e., decrease in early lactation followed by a gradual increase) were observed both in the present study and elsewhere (
observed a similar stage of lactation trend when β-CN was expressed relative to total CN. The observed decline in total proteins and protein fractions in early lactation coincides with the period of negative energy balance typically seen in dairy cows in early lactation (
); negative protein balance (when the quantity of protein broken down by the cow exceeds the quantity ingested) may also occur. It can take a cow up to 20 wk to regain a positive energy and protein balance and for actual milk protein content to increase again (
The reduction in concentration of total FAA and the individual FAA (Glu, Gly, Lys, Arg, and Asp) in milk up to 65 d postcalving corroborates documented reports by
, who studied the concentration of these FAA in bovine milk from 7 to 60 DIM. Both total FAA and Glu in milk decreased as parity increased and, despite the part–whole relationship between them (Glu makes up >55% of total FAA), the lactation profile of Glu and total FAA in milk differed in younger animals (Figure 4). To our knowledge, no study to date has investigated the association between parity and FAA composition of milk.
Variability in Milk Quality
Considerable variability in protein fractions and FAA existed in both populations studied. The coefficient of variation of β-LG B and Gly was 51 and 69%, respectively, in the research herds, which is considerably greater than the coefficient of variation of 31% observed for milk yield in the same population (results not shown). The potential of milk MIRS to predict protein fractions and FAA (
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
) provides an opportunity to generate large quantities of data for use in genetic evaluations and thus breeding programs. The attributes and tools therefore appear to exist to facilitate breeding programs for superior milk quality characteristics if such characteristics are deemed to be heritable.
Contemporary group accounted for between 45.34% (total FAA) and 86.11% (α-LA) of the variation in the data set used in this study, indicating that the combination of herd and test date has large effects on both protein and FAA composition of the milk. These differences offer the potential for herds to be selected on the basis of their protein and FAA profile; further milk price premiums could be paid to herds that are producing milk with a protein or FAA profile that better fits the processor's needs for production. Herd-level estimates of milk quality can be readily obtained as a by-product from national genetic evaluations, and thus the data can be readily available. These herd solutions would be independent of genetic merit of the producing animals and therefore more closely reflect the influence of management on milk quality. Moreover, the ability to monitor the trend in milk quality over time within a herd would help provide producers and processors with information to help support their decisions about the factors affecting the quality of their milk.
Decision Support Tool
The observed trends in milk protein fractions and FAA across month of the year (Figure 5) suggest differences in the suitability of milk for producing milk products across the year and provide evidence of the difficulty in acquiring a stable product of constant composition across time. The effect on consistency of product is further compounded in seasonal calving herds that exist in Ireland (
) and elsewhere because milk protein fractions and FAA also vary across stage of lactation, which is synchronized with calendar month. The structure of the data, coupled with the statistical model, implies that the observed effects reported in the present study are independent of each other and are therefore additive. Commercially available infant formulas have a ratio of total CN to total whey protein that is close to that of human milk, but β-LG, which is not present in human milk, is present in the greatest amount in cow milk. It is advantageous for infant formula producers to select cow milk with a higher concentration of α-LA and a lower concentration of β-LG. Figure 3 demonstrates that earlier parity animals in early lactation produce the highest concentration of α-LA. Results from the present study also show that milk produced by young JE cows in the months of August, September, and October could achieve a greater concentration of CN fractions in milk.
Bovine and human milk also differ in their amino acid profiles (
). Free amino acids are often added to infant formula by processors, especially for the production of formula for infants with allergies to CN and whey protein fractions in milk (
). Results from the present study may also help infant formula processors select milk that is naturally higher in the sought-after protein or FAA profile, thereby minimizing the requirement for protein and FAA additives.
Previous studies have revealed that greater concentrations of all CN fractions in milk significantly increase cheese yield (
Effectiveness of mid-infrared spectroscopy for the prediction of detailed protein composition and contents of protein genetic variants of individual milk of Simmental cows.
). Figure 2, Figure 3 indicate that the concentration of certain protein fractions (αS1-CN, β-CN, κ-CN, and β-LG B) in early lactation could possibly be too low for cheese production; this could be useful information for cheese manufacturers to optimize quality and yield of cheese. Mid- and late-lactation milk has a greater concentration of κ-CN than early-lactation milk, and first-parity animals have a lower concentration of κ-CN than their older contemporaries. A higher concentration of κ-CN in milk results in a smaller CN micelle (
) and therefore in a shorter rennet coagulation time, contributing to a stronger curd and more cheese yield. However, it is the genetic polymorphism of κ-CN B that is of importance (
), suggesting that a way to improve milk process ability for cheese production would be to select animals with genes coding for κ-CN B. Jersey cows produced more CN fractions but less total FAA in milk compared with HO cows, indicating that JE cows may produce milk that is more suitable for cheese production and of a better processing quality than HO cows.
Crossbreeding
Heterosis is defined as the difference between the performance of a crossbred animal and the average of the parents (
). Relative to the purebred parent average, first-cross cows produced 0.13 g/L more β-CN. Studies have indicated that the consumption of β-CN A1 is associated with higher mortality rates from coronary heart disease in humans (
). However, further research is required to determine the genetic composition of β-CN in the milk analyzed in the present study based on its genetic composition. Recombination loss is defined as the disintegration of epistatic associations to form nonparent interloci combinations of alleles in crossbred animals (
), even though favorable recombination estimates were calculated for concentrations of all the individual protein fractions (total CN, αS1-CN, αS2-CN, β-CN, κ-CN, total β-LG, β-LG A, and β-LG B) in milk. An unfavorable effect is expected because recombination normally affects traits such as milk production that have been under long-term selection intensity (
Milk production and fertility performance of Holstein, Friesian, and Jersey purebred cows and their respective crosses in seasonal-calving commercial farms.
showed recombination to have positive effects on milk compositional traits, including protein percentage, and suggested that different population breeding goals may be a causative factor in the inconsistencies among studies on the effect of recombination on milk compositional traits. Traditionally, Irish dairy cows may have been naturally selected for fertility and survivability as a result of the seasonal calving system operated in Ireland (
Milk production and fertility performance of Holstein, Friesian, and Jersey purebred cows and their respective crosses in seasonal-calving commercial farms.
). Results from the present study indicate that crossbred cows had a greater concentration of β-CN in milk than purebred HO, which is advantageous for cheese production, thus demonstrating another advantage of crossbreeding.
CONCLUSIONS
Results from the present study indicate that factors including stage of lactation, parity, calendar month of the year, milking time, and breed are all associated with protein and FAA composition of bovine milk. Of particular interest was that younger animals produced more total CN, total whey, and total β-LG across early and mid lactation and more Glu and Asp in milk across lactation than their older contemporaries. Milk produced by JE cows had a greater concentration of all CN fractions but a lower concentration of total FAA than that produced by HO cows. This study provides information on how individual milk proteins and FAA change across calendar months of the year and across stages of lactation, which could be useful input parameters for decision support tools in the management of product portfolios by processors over time.
ACKNOWLEDGMENTS
Funding for this work was provided by the Irish Department of Agriculture, Food and the Marine (Dublin, Ireland) Research Stimulus Fund project 11/SF/311, Breed Quality.
APPENDIX
Table A1Number of records (n), root mean squared error (RMSE), correlation coefficient between true and predicted values in cross-validation (rc), external validation (rv), and ratio performance deviation (RPD) tested using the split-sample cross-validation and external validation from
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
Effectiveness of mid-infrared spectroscopy for the prediction of detailed protein composition and contents of protein genetic variants of individual milk of Simmental cows.
Milk production and fertility performance of Holstein, Friesian, and Jersey purebred cows and their respective crosses in seasonal-calving commercial farms.
Prediction of individual milk proteins including free amino acids in bovine milk using mid-infrared spectroscopy and their correlations with milk processing characteristics.
Effectiveness of mid-infrared spectroscopy to predict the color of bovine milk and the relationship between milk color and traditional milk quality traits.
Associations of stage of lactation, milk protein genotype and body condition at calving on protein composition and renneting properties of bovine milk.
30592-1&id=gr1.jpg){kind=link}
30592-1&id=gr2.jpg){kind=link}
30592-1&id=gr3.jpg){kind=link}
30592-1&id=gr4.jpg){kind=link}
30592-1&id=gr5.jpg){kind=link}