May 11, 1998 - favorable genetic trend in the milk coagulation properties ... MATERIALS AND METHODS. Data. A total of 789 FAy and 86 Finnish Friesian ( FFr) cows from 51 ... A variance-covariance structure between studied traits 1 and 2 ...
Genetic Parameters for the Milk Coagulation Properties and Prevalence of Noncoagulating Milk in Finnish Dairy Cows T. IKONEN, K. AHLFORS, R. KEMPE, M. OJALA, and O. RUOTTINEN Department of Animal Science, University of Helsinki, PO Box 28, 00014 Helsinki University, Finland
ABSTRACT The genetic parameters were estimated for milk coagulation properties and milk production traits, and the prevalence of noncoagulating milk in the Finnish dairy cattle population was investigated. Data were included for 789 Finnish Ayrshire cows and 86 Finnish Friesian cows from 51 herds. The animal model used for estimation included fixed effects for parity, stage of lactation, breed, and herd. Further, effects of milk protein genotypes on phenotypic and genetic variation in the studied traits were examined. Heritability estimates for the milk coagulation properties were moderately high. The k-casein B allele was associated with the best phenotypic and genetic values for curd firmness, and the A and E alleles were associated with the poorest. About 24% of the additive genetic variation in the curd firmness was due to milk protein polymorphism. About 8% of the Finnish Ayrshire cows in the present study produced noncoagulating milk. Because of the occurrence of the noncoagulating milk and a possibly unfavorable genetic trend in the milk coagulation properties, it would be important to improve these traits in the Finnish Ayrshire breed. Milk coagulation properties could be improved directly by selecting for these traits or indirectly by favoring the k-casein B allele or by selecting against genetic markers associated with poorly coagulating or noncoagulating milk. ( Key words: milk coagulation, noncoagulation, genetic parameters, Finnish dairy cows) Abbreviation key: FAy = Finnish Ayrshire, FFr = Finnish Friesian, NC = noncoagulating. INTRODUCTION The importance of cheese production has increased during the past 15 yr in Finland, and currently about
Received May 11, 1998. Accepted September 9, 1998. 1999 J Dairy Sci 82:205–214
40% of the milk produced there is used for cheese production. The coagulation ability of milk is essential in cheese making. Milk with favorable coagulation properties (i.e., short coagulation and curdfirming times and a firm curd) is expected to give more cheese with desirable composition than milk with unfavorable properties (3, 6). There are no published data on variation or trend in the milk coagulation properties over the past decades in Finland. However, according to the observations made in Finnish dairies, the average coagulation ability of milk has been deteriorating during the past 20 to 30 yr. Consequently, to be able to produce the same amount of cheese, more milk is needed today than in recent decades. In addition, two Finnish data sets collected in 1980s ( 2 9 ) and 1990s ( 1 2 ) showed that 3.6% of 168 Finnish Ayrshire ( FAy) cows and 10.2% of 59 FAy cows, respectively, produced noncoagulating ( NC) milk. Because of the importance of cheese production in Finland, there is an interest in halting the undesirable trend in milk coagulation properties and in improving these traits. Consequently, reasons for the unfavorable changes and variation in the coagulation properties have to be identified. Should a reasonable part of the variation in the milk coagulation properties be genetic, these characteristics could be improved by selection. In addition, it is important to find out whether the relatively high frequency of NC milk that has been reported in (12, 29) was by chance or is a common phenomenon in the Finnish dairy cattle population. Heritability estimates for the milk coagulation properties have been estimated in only a few studies (15, 25, 29), none of which used an animal model to account for all known genetic relationships among animals. Because of the small data sets and statistical procedures assumed, some heritability estimates were not reliable. The results of the previous studies implied, however, that a moderate part of the variation in the milk coagulation properties could be due to additive genetic effects.
205
206
IKONEN ET AL.
The lack of suitable equipment for routine determination of the milk coagulation properties in the national cow population would restrict possibilities of direct selection for these traits. An indirect way of improving the milk coagulation properties might possibly be to favor the k-CN B allele or alleles at other loci that are associated with favorable coagulation properties. The k-CN B allele is known to be associated with more desirable coagulation properties (1, 4, 12, 19, 26, 28, 29, 30) and protein composition of milk (12, 19, 22, 28) than is the A allele, but no reliable estimates exist of the effect of the k-CN E allele on these traits. The k-CN E allele is rather common (30%) in the FAy (13), the main dairy breed in Finland. Another indirect way of improving the milk coagulation properties might be to breed for routinely recorded dairy traits that correlate favorably with the milk coagulation properties. Genetic correlations between the milk coagulation properties and milk production traits have thus far been estimated from only a few, small data sets (15, 25). Reliable estimates for the genetic correlations between the milk coagulation properties and milk production traits are therefore needed. The objectives of this study were to estimate genetic parameters (heritabilities and genetic correlations) for milk coagulation properties and milk production traits and to estimate the prevalence of NC milk in the Finnish dairy cattle population. MATERIALS AND METHODS Data A total of 789 FAy and 86 Finnish Friesian ( FFr) cows from 51 herds in southern Finland were sampled once during the period from February to April 1995. The milk samples were a mixture of evening and morning milkings; samples were analyzed for milk coagulation properties, pH, milk yield, fat percentage, protein percentage, SCC, and major milk protein genotypes. Background information about the cows and herds was obtained from national milk recording data sets from Agricultural Data Processing Centre (Vantaa, Finland). It was not possible to get detailed information about the feeding of the cows, although feeding procedures can affect milk coagulation properties (1, 8, 19). Laboratory Analyses The milk coagulation properties were determined at 32°C using a Formagraph (Foss Electric A/S, Journal of Dairy Science Vol. 82, No. 1, 1999
Hillerød, Denmark) in the laboratory of the Food Research Institute of Agricultural Research Centre (Jokioinen, Finland). Rennet (Renco Calf Rennet Liquid; New Zealand Rennet Company Ltd., Eltham, New Zealand) was diluted in 0.07 M sodium acetate buffer (1:100, vol/vol; pH 5.5). The pH of milk was determined (PHM82 Standard pH Meter; Radiometer, Copenhagen, Denmark). The milk samples were allowed to coagulate for 30 min because, during the cheese-making process, curd is usually cut 30 min after the addition of rennet to the milk. The three milk coagulation properties determined were milk coagulation time in minutes, curdfirming time in minutes, and firmness of the curd in millimeters. Milk coagulation time was the time from the addition of rennet to milk to the beginning of coagulation. Curd-firming time was the time from the beginning of coagulation to the moment the width of the curve was 20 mm. Firmness of the curd was the width of the curve 30 min after the addition of rennet. The milk samples that did not form curd in 30 min (i.e., a straight line on the output paper) were defined as NC samples. Fat and protein percentages were determined using a Milko Scan 605 (Foss Electric A/S), and SCC was determined by means of a Fossomatic 360 (Foss Electric A/S) in local dairy laboratories. The frequency distribution for SCC was not normal, so SCC were logarithmically transformed. Genotypes for the as1-CN, b-CN, k-CN, and b-LG were determined in Finnish Animal Breeding Association laboratory (Vantaa, Finland) using isoelectric focusing as described by Erhardt ( 5 ) . Statistical Analyses Univariate and bivariate models. Heritabilities for the studied traits and genetic correlations between the milk coagulation properties and milk production traits were first estimated using an univariate and a bivariate model: yijklmn = m + parityi + stagej + breedk + herdl + animm + eijklmn, [1] where yijklmn m parityi stagej breedk
= = = =
milk coagulation or milk production trait, mean, fixed effect of parity i ( i = 1 to 4), fixed effect for stage of lactation j ( j = 1 to 6), = fixed effect of breed k ( k = 1 to 2),
GENETIC PARAMETERS FOR MILK COAGULATION
herdl = fixed effect of herd l ( l = 1 to 51), animm = random additive genetic effect of animal m, N ( 0, As2a) , and eijklmn = random residual effect, N ( 0, Is2) . e
A variance-covariance structure between studied traits 1 and 2 for a bivariate model was assumed, where sa1a2 and se1e2 are additive genetic and residual covariances between traits 1 and 2:
Var
a1 a2 = e1 e2
As2a1 Asa1a2 0 0
Asa1a2 0 0 As2a2 Is2e1 0 Ise1e2 0
0 0 Ise1e2
Is2e2
Parity was grouped in four classes: first, second, third, and fourth to ninth parities; stage of lactation was grouped into six classes: 5 to 30 d, 31 to 60 d, 61 to 120 d, 121 to 180 d, 181 to 240 d, and >240 d after calving; and breed was grouped into two classes: FAy and FFr. Herd was treated as a fixed effect because the herds were a group selected from a certain area, and differences in the milk coagulation properties between them were of interest. Of the 51 herds, 33 were pure FAy herds, 1 was a pure FFr herd, and 17 were mixed herds. The number of animals per herd ranged from 6 to 47. The 789 FAy and 86 FFr cows with records were daughters of 246 FAy and 41 FFr sires, respectively. The average number of daughters per sire was only 3 but ranged from 1 to 79. The pedigree information consisted of parents and grandparents of the cows with records, and the total number of animals in the statistical analyses was 2757. Variance and covariance components for the random effects were estimated from the data using the REML VCE package (10). Solutions for the fixed and random effects in the models were obtained using the PEST package ( 9 ) , and statistical significance of the fixed effects was tested using the F test provided by the PEST package ( 9 ) . Multivariate models. In addition to univariate and bivariate analyses, heritabilities for the studied traits and genetic correlations between the milk coagulation properties and milk production traits were estimated using a corresponding multivariate model, in which a milk coagulation characteristic was analyzed simultaneously with pH, milk yield, fat percentage, protein percentage, and SCC. The multivariate model was used to determine whether more reliable and more accurate estimates for the genetic parameters would be obtained when information on phenotypic and genetic association between all studied traits was available.
207
Modification of the univariate model. In order to estimate the effects of b-CN, k-CN, and b-LG genotypes on the milk coagulation properties and to study how their inclusion in the model affects additive genetic variation in the coagulation properties, the following model was assumed: yijklmnop = m + parityi + stagej + breedk + herdl + b-k-CNm + b-LGn + animo + eijklmnop [2] where b-k-CNm = fixed effect of b-k-CN genotype class m ( m = 1 to 9), and b-LGn = fixed effect of b-LG genotype ( n = 1 to 3). The b-CN and k-CN were included in Model [2] as composite b-k-CN grouped into nine genotype classes (Figure 1). Because the k-CN B allele was rare (0.08), the AB, BB, and BE genotypes of k-CN were combined within each b-CN genotype. Being almost monomorphic, as1-CN was not considered in the formation of the composite casein genotypes. The b-LG polymorphism was grouped into three genotype classes: AA ( n = 88), AB ( n = 366), and BB ( n = 421). Other effects, assumptions, and statistical procedures used with Model [2] were identical to those used with Model [1]. RESULTS AND DISCUSSION Means and Variation The milk coagulation properties varied considerably (Table 1). About 8% of the milk samples did not coagulate in 30 min (curd firmness = 0.0 mm), and, thus, the distribution for the curd firmness was skewed toward the lowest (i.e., the most undesirable) values. In addition, every third milk sample did not reach a curd firmness of 20 mm in 30 min. The number of samples for the coagulation time and, especially, for the curd-firming time was, therefore, lower than for curd firmness. The curd-firming time was thus excluded from further statistical analyses. Means and variation for the milk coagulation and milk production traits (Table 1 ) were of the same magnitude as those reported by Ikonen et al. (12). Factors Affecting the Milk Coagulation Properties Parity. Milk yield, SCC, and pH increased with parity (Table 2). Both SCC and pH usually have an Journal of Dairy Science Vol. 82, No. 1, 1999
208
IKONEN ET AL.
Figure 1. Estimates of effect of the b-k-CN genotypes on the milk coagulation properties. 1AB ( n = 5 ) + BB ( n = 1 ) + BE ( n = 26); 2AB ( n = 80) + BB ( n = 2 ) + BE ( n = 6); 3AB ( n = 13) + BB ( n = 1).
unfavorable effect on the milk coagulation properties [e.g., (26, 27)]. Parity had, however, no statistically significant ( P = 0.423) effect on the milk coagulation properties. Similar results were observed also in some other studies (4, 15, 26), but, in the study of Schaar et al. (28), the milk coagulation properties improved with parity. Stage of lactation. The milk coagulation properties were at their best during the 1st mo of lactation ( 5 to 30 d after calving) and again from the 9th mo (>240 d after calving) onward (Table 3). The changes in the curd firmness over the course of lactation were parallel to those in protein and fat percentages and in SCC and were almost opposite to the
changes in milk yield over lactation (Table 3). This result agrees with those reported in (12, 14), but, in others (4, 24), the milk coagulation properties deteriorated as lactation proceeded. In (15, 26, 28), stage of lactation had no effect on the milk coagulation properties. Confounded effects for stage of lactation and season in some of the previous studies may partially explain the contradictory results for the effect of stage of lactation on the milk coagulation properties. Breed. The milk coagulation properties were on average better for the FFr than for the FAy cows (Table 4), which was in part because NC milk was found only in the FAy. Also a difference between the breeds in pH of milk explained some of the differences
TABLE 1. Means and variation in the milk coagulation and milk production traits.
Coagulation time,3 min Curd-firming time, min Curd firmness, mm Curd firmness,3 mm pH Daily milk yield, kg Fat content, % Protein content, % Log-transformed SCC 1Number
n1
X
Minimum
Maximum
CV2
809 587 875 809 875 875 875 875 875
12.3 9.6 23.2 25.1 6.78 24.8 4.41 3.35 4.28
3.5 1.5 0.0 1.0 6.58 6.0 2.29 2.45 1.39
30.0 23.0 56.0 56.0 7.12 53.2 7.80 4.81 8.58
41 47 57 50 1 31 16 11 29
of cows and observations. of variation, expressed as a percentage. 3Milk samples that coagulated in 30 min. 2Coefficient
Journal of Dairy Science Vol. 82, No. 1, 1999
209
GENETIC PARAMETERS FOR MILK COAGULATION TABLE 2. Estimates (Est.) of effect of parity relative to the first parity class on the milk coagulation and milk production traits. Parity 1 ( n = 313) Coagulation time,1 min Curd firmness, mm pH Daily milk yield, kg Fat content, % Protein content, % Log-transformed SCC 1Milk
2 ( n = 240) Est. 0.2 –0.1 0.03 2.4 0.00 0.09 0.44
0 0 0 0 0 0 0
SE 0.5 1.1 0.01 0.5 0.05 0.02 0.10
3 ( n = 148) Est. –0.4 2.1 0.05 3.3 –0.04 –0.02 0.46
SE 0.5 1.2 0.01 0.5 0.06 0.03 0.12
4 to 9 ( n = 174) Est. –0.6 2.5 0.04 5.1 –0.10 –0.03 0.58
SE 0.5 1.2 0.01 0.5 0.06 0.03 0.11
P 0.423 0.066
