- Email: [email protected]
Genetic Variation of Residual Feed Consumption in a Selected Finnish Egg-Layer Population
BREEDING AND GENETICS Genetic Variation of Residual Feed Consumption in a Selected Finnish Egg-Layer Population N. SCHULMAN, M. TUISKULA-HAAVISTO, L. SIITONEN,1 and E. A. MANTYSAARI Agricultural Research Centre of Finland, Institute of Animal Production, Animal Breeding, FIN-31600 Jokioinen, Finland
1994 Poultry Science 73:1479-1484
INTRODUCTION Feed expenses are the main cost in egg production. Therefore, efficient layers are important for the economy of the farm. Traditionally, efficiency has been improved by selecting on egg mass production and body weight and getting a correlated response in feed efficiency (Luiting, 1990). Recently the use of feed consumption data or parameter estimates including feed consumption in the selection have been suggested to be good tools for improving feed efficiency further (Arboleda et al, 1976; Wing and Nordskog, 1982b; Nordskog et al, 1991). Residual feed consumption (RFC) is feed consump-
Received for publication December 27, 1993. Accepted for publication May 30, 1994. ^ o whom correspondence should be addressed.
tion corrected for egg mass production, body weight, and body weight gain (Byerly, 1941; Nordskog et al, 1972; Arboleda et al, 1976; Byerly et al, 1980; Bentsen, 1983a). It measures the remaining part of the variation in feed consumption not accounted for by these three traits. Studying RFC can give valuable information on the effects of using feed consumption data in the selection of layers (Bentsen, 1983b; Luiting and Urff, 1991a). The first estimated heritabilities of RFC were very small (Nordskog et al, 1972; Arboleda et al, 1976). Later, much higher values have been found, ranging from .30 to .60 (Hagger and Abplanalp, 1978; Bentsen, 1983b; Katie, 1986; Luiting and Urff, 1991b). Heritabilities are usually higher at the beginning of the laying period than later (Katie, 1986; Luiting and Urff, 1991b).
1479
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
ABSTRACT The purpose of the study was to estimate the heritability of residual feed consumption (RFC) and the genetic correlations between RFC and economically important traits. The genetic progress after four generations of selection for RFC and the changes in economically important traits were also investigated. A selection experiment for RFC was carried out from 1983 to 1987. The total data consisted of 3,750 birds and 2,661 records. The (co)variance components were calculated using derivative-free bivariate animal model restricted maximum likelihood (REML). Breeding values were estimated for calculating genetic progress in RFC and correlated responses in the other traits. The heritability of RFC calculated from the whole recorded period (16 to 42 wk) and using all 2,661 records was .46 (± .04). The genetic correlations between RFC and egg mass, number of eggs, egg weight, and body weight were not significant. The genetic correlation between RFC and feed consumption was .50 (± .04). The breeding value estimates indicated a moderate genetic progress in RFC due to selection. Feed consumption was decreased and body weight gain showed reduction in the last two generations. No change could be found in egg mass, number of eggs, egg weight, age at first egg, or body weight. (Key words: heritability, residual feed consumption, egglayer, feed conversion efficiency, restricted maximum likelihood)
1480
SCHULMAN ET AL.
MATERIALS AND METHODS Data from a selection experiment for RFC in a Finnish Leghorn population was
used. The selection experiment had been carried out in the years 1983 to 1987 and lasted five generations (Tuiskula-Haavisto, 1986; Liuttula, 1988). The total data consisted of 3,750 birds and 2,661 records. Since 1984, the birds were divided into a selection line and a control line (Table 1). The parents of the base population were known without records. Results of hens that had died or stopped laying in the recorded period were excluded from the analysis (4% of all hens). There were two hatches each in Generations 1, 2, and 3 and only one hatch in Generations 0 and 4. The chicks were reared on the floor for Generations 0, 1 (both hatches), and 3 (first hatch); otherwise, chicks were in cages. At the age of 16 wk, the hens were moved into individual cages. The hens consumed feed ad libitum. The feed used was commercial concentrate supplemented by barley, oats, and calcium. The mixture contained in average 29 g / k g Ca and 6.3 g / k g P. Calculated energy content of the feed was 2.53 kcal/kg and the protein content 16.5% of dry matter. The average temperature during the experiment was about 20 C. The lighting started with 10 h and was gradually increased to 14 h light/d except for the last generation, for which the lighting was increased to 16 h light/d. The measuring period was 26 wk over the period from 16 to 42 wk in age. The whole period was divided into six 4-wk periods and one 2-wk period. The traits measured were age at first egg, number of eggs, egg weight, adult body weight, and feed consumption.
TABLE 1. Number of females and males in the selection and control lines and numbers selected Generation Selection line 0 1 2 3 4 Control line 0 1 2 3 4 iFull brothers were not selected.
Number of females
Selected females
Number of males
Selected males1
485 466 255 561 239
78 55 79 80
292 141 70 156
26 19 19 22
485 168 259 119 109
184 143 132 109
292 51 84 88
86 43 43 41
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
There are not many reports on genetic correlations between RFC and other economically important traits. Luiting and Urff (1991b) found a positive genetic correlation between RFC and feed consumption. The genetic correlations between RFC and egg mass, number of eggs, egg weight, age at first egg, body weight, and body weight gain are all reported to be close to zero (Bentsen, 1983b; Katie and Kolstad, 1991; Luiting and Urff, 1991b). The response to selection for RFC and the effects of selection on other economically important traits have been studied by selection experiments (Bordas and M6rat, 1984; Katie and Kolstad, 1991; Bordas et al., 1992). Residual feed consumption was changed due to selection in these experiments. The aim of this study was to estimate the heritability of RFC in a Finnish Leghorn population and to see whether there are significant genetic correlations between RFC and other economically important traits (e.g., egg mass, number of eggs, egg weight, age at first egg, body weight, body weight gain, and feed consumption). The genetic progress in RFC and the genetic changes in the other traits over four generations of selection for RFC were also investigated. The (co)variance component estimation was carried out using bivariate animal model REML (restricted maximum likelihood).
GENETIC VARIATION OF RESIDUAL FEED CONSUMPTION
E(FC) = a + ba AW + b 2 EM + b 3 BW75
where E(FC) = expected feed consumption (grams per day); AW = body weight gain (grams per day); EM = egg mass production (grams per day); BW75 = metabolic body weight (grams); bi, b2, b 3 = regression coefficients; and a = constant. The hens were selected using their own record for RFC. In the base generation, hens with less than the population mean 30.4 g egg mass/d were excluded from the selection. At the base generation, the males were selected on their full sibs's RFC. Four males were culled because of the extremely low egg mass production of their full sibs. In the subsequent genera-
tions, a selection index including full- and half-sib records was used (Becker, 1967). The heritability used in the index was .24, calculated from the half-sib correlation in the base population. The numbers selected are shown in Table 1. An unselected control line was kept during the experiment. In the control line the aim of the mating system was to increase the effective number of parents and thus to avoid inbreeding by choosing as sires one male offspring from each sire and one female offspring from each dam in the previous generation (Gowe et al, 1959) (Table 1). When producing the second generation, 26 female offspring from as many different full-sib groups were brought from the selection line to the control line because of a failure in the incubation of the control. The (co)variance components were estimated using a REML procedure (Patterson and Thompson, 1971), and the derivativefree method was applied as the computing procedure (Graser et al, 1987; Meyer, 1991). The following animal model was used: y = Xb + Zu + e where y = the vector of n observations of one or two different traits, of which one is always residual feed consumption; b = the vector of fixed generation-hatch classes; u = the vector of random animal effects; e = the vector of residuals; X = the incidence matrix for fixed effects; Z = the incidence matrix for animal effects;
TABLE 2. Regression coefficients used for calculation of expected feed consumption [E(FQ]. The equation where the coefficients were used is E(FC) = a + bj AW + b 2 EM + b 3 BW-75 1 Period
Age
1 2 3 4 5 6 7
(wk) 16 to 20 to 24 to 28 to 32 to 36 to 40 to
20 24 28 32 36 40 42
Constant a 6.48 5.57 7.00 2.04 10.10 7.35 1.91
^
EM b2
BW-75 b3
R2
.75 1.00 .69 .91 1.10 .74 .53
1.00 .46 .47 .62 .64 .83 .90
50.7 52.7 36.4 56.5 43.8 36.1 40.5
58.8 59.5 51.3 56.8 58.6 65.0 68.9
AW
1AW = body weight gain; EM = egg mass; BW-75
=
metabolic body weight.
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
The selection was based on total RFC expressed as a percentage of expected feed consumption. The RFC was calculated at each period as the difference between observed and expected feed consumption (Bentsen, 1983a). Finally, the results for each period were combined to give the total RFC for each hen. The expected feed consumption was derived from multiple regression equations. Every period had its own regression coefficients. The coefficients were calculated from the base population and used unchanged in all the other generations (Table 2), except in the base population, for which only one regression equation covering the whole laying period was used. The regression equation was:
1481
1482
SCHULMAN ET Ah.
and the (co)variances V(u) = G = A <8> G 0 , V(e) = R = I ® RQ, cov (u,e') = 0, and V(y) = V
RESULTS AND DISCUSSION Heritability of Residual Feed Consumption The heritability of total RFC (from 16 to 42 wk of age) estimated for the base population from the whole data (base generation, selection, and control line) was .46 with a standard error of .04. When only animals from the base population and the selection line were used the estimated heritability was .41 (± .06), and when the base population and control line were used it was .48 (± .06). These values are in good agreement with those found in the literature (Hagger and Abplanalp, 1978; Bentsen, 1983b; Luiting and Urff, 1991b). The heritability estimated by Bentsen (1983b) was .50
Genetic Correlations Between Residual Feed Consumption and Economically Important Traits The genetic correlations between RFC and egg mass, number of eggs, egg weight, and body weight were not significant. This is in agreement with the results obtained by Bentsen (1983b) and Luiting and Urff (1991b). However, negative genetic correlations have been also reported between RFC and egg mass, egg production, and egg weight (Fairfull and Chambers, 1984; Katie and Kolstad, 1991). A very small positive genetic correlation between RFC and body weight gain was found (.09 ± .06). Also, Bentsen (1983b) found a very small genetic correlation between body weight gain and RFC. However, nonsignificant correlations have been also reported (Katie and Kolstad, 1991; Luiting and Urff, 1991b). There seemed to be a small negative genetic correlation between RFC and age at first egg (-.11 ± .06). Negative correlations have also been found by Bentsen (1983b). However, a positive genetic correlation between RFC and age at first egg has been also reported (Fairfull and Chambers, 1984). The genetic correlation between RFC and feed consumption was .50 ± .04. This is well in agreement with the literature values reviewed by Luiting (1991). However, Luiting and Urff (1991b) reported a smaller positive value.
Genetic Progress in Residual Feed Consumption The average breeding value estimates showed a moderate genetic progress in RFC from 1983 to 1987 due to selection for RFC as a percentage of expected feed consumption (RFC%) (Figure 1). The genetic progress was in total -6.13 g / h e n per d. This is approximately -1.53 g / h e n per d per generation. In the control line, the RFC stayed close to zero. When the genetic progress was estimated by comparing the
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
where <8> is the Kronecker product; G 0 is the genetic (co)variance matrix between the traits; A is the relationship matrix between animals; and RQ is the residual (co)variance matrix between traits. The variance components were first estimated using univariate procedures and then with multivariate procedures. Only two traits were considered simultaneously in the multivariate procedure. The standard errors of the genetic correlations were calculated according to Falconer (1989). Breeding values for all animals were estimated to compute the genetic progress in RFC and a correlated change in the other traits. The breeding values were estimated using animal model best linear unbiased prediction (AM-BLUP). The same model as for (co)variance component estimation was used. The heritability used was .46. Inbreeding coefficients were obtained as diagonal elements of the numerator relationship matrix computed using the procedure presented by Quaas (1976) and averaged within lines and generations.
(from 16 to 22 wk of age), and the one estimated by Luiting and Urff (1991b) was .48 (from 20 to 44 wk of age). A lower estimate has been reported by Wing and Nordskog (1982a) and a slightly higher estimate by Katie (1986).
GENETIC VARIATION OF RESIDUAL FEED CONSUMPTION
1483
Generation
results of the selection and the control line the progress was estimated as -1.43 g/hen per d per generation. The final inbreeding coefficient was 1.9% in the selection line and .8% in the control. Because the coefficients were low in both lines, there was no reason to assume that inbreeding would have invalidated the use of the control line to estimate genetic trend or reduced the rate of response in the selection line. Moderate response in RFC has been found also in other selection experiments (Bordas and M6rat, 1984; Katie and Kolstad, 1991; Bordas et al, 1992). Genetic Changes in Economically Important Traits The average breeding value estimates showed that feed consumption was decreased due to the positive genetic correlation between RFC and feed consumption (Figure 2). The same kind of response was reported by Bordas et al (1992). In the present experiment, body weight gain tended to decrease slightly in the selection line in the last two generations, which was in agreement with the estimated genetic correlation. In the selection experiment of Bordas et al. (1992) no changes in body weight gain were observed. No genetic changes were found in egg mass, number of eggs, egg weight, age at first egg, or body weight. This is in agreement with the estimated genetic correlations except that the small negative genetic correlation found between RFC and age at first egg did not appear as a change in the age at first egg after selection for RFC as percentage of expected feed consumption.
FIGURE 2. The genetic changes in feed consumption over four generations of selection on residual feed consumption as a percentage of expected feed consumption.
In some selection experiments for RFC, no genetic changes in egg mass production, number of eggs, or body weight gain were found (Bordas and M6rat, 1984; Bordas et al, 1992). Katie and Kolstad (1991) did not observe changes in low lines, but in the high line egg mass was decreased. Inconsistent changes were found in age at first egg and egg weight (Bordas et al, 1992). In the experiment of Katie and Kolstad (1991), body weight was found to increase due to selection for RFC as a percentage of expected feed consumption in low lines, whereas Bordas et al (1992) did not find any differences in body weight between their high and low lines. When the genetic changes were investigated by comparing the phenotypic values of the different traits in the selection and control lines, the results were similar to those above except for body weight gain. This method did not show any change in body weight gain due to selection on RFC as a percentage of expected feed consumption. ACKNOWLEDGMENTS We are very grateful to Anne Nylander for help in doing the REML analysis and to Anni Jarvinen and Laura Lauttamaki for excellent technical assistance.
REFERENCES Arboleda, C. R., D. L. Harris, and A. W. Nordskog, 1976. Efficiency of selection in layer-type chick-
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
FIGURE 1. The genetic progress in residual feed consumption over four generations of selection.
1484
SCHULMAN ET AL. Katie, J., and N. Kolstad, 1991. Selection for efficiency of food utilisation in laying hens: Direct response in residual food consumption and correlated responses in weight gain, egg production and body weight. Br. Poult. Sci. 32:939-953. Liuttula, M., 1988. Selection experiment for feed efficiency in egg laying hens. 144th Nordisk Jordbrugsforskning seminar, Wik, Sweden. Luiting, P., 1990. Genetic variation of energy partitioning in layer hens: causes of variation in residual feed consumption. World's Poult. Sci. 46:133-152. Luiting, P., 1991. The Value of Feed Consumption Data for Breeding in Laying Hens. Ph.D. Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Luiting, P., and E. M. Urff, 1991a. Optimization of a model to estimate residual feed consumption in the laying hen. Livest. Prod. Sci. 27:321-338. Luiting, P., and E. M. Urff, 1991b. Residual feed consumption in laying hens. 2. Genetic variation and correlations. Poultry Sci. 70:1663-1672. Meyer, K., 1991. Estimating variances and covariances for multivariate animal models by restricted maximum likelihood. Genet. Sel. Evol. 23:67-83. Nordskog, A. W., H. L. French, C. R. Arboleda, and D. W. Casey, 1972. Breeding for efficiency of egg production. World's Poult. Sci. J. 28:175-188. Nordskog, A. W., Y. H. Hou, H. Singh, and S. Cheng, 1991. Selecting for efficiency of egg production using food consumption records in layer-type chickens. Br. Poult. Sci. 32:87-101. Patterson, H. D., and R. Thompson, 1971. Recovery of interblock information when block sizes are unequal. Biometrika 58:545-554. Quaas, R. L., 1976. Computing the diagonal elements of a large numerator relationship matrix. Biometrics 32:949-953. Tuiskula-Haavisto, M., 1986. Resultater og information fra finske forsag med selektion for foderudnyttelse hos verpelwms. 103rd Nordisk Jordbrugsforskning seminar, Foulum, Denmark. Wing, T. L., and A. W. Nordskog, 1982a. Use of individual feed records in a selection program for feed efficiency. I. Heritability of the residual component of feed efficiency. Poultry Sci. 61: 226-230. Wing, T. L., and A. W. Nordskog, 1982b. Use of individual feed records in a selection program for feed efficiency. II. Effectiveness of different selection indexes. Poultry Sci. 61:231-235.
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
ens by using supplementary information on feed consumption. II Application to net income. Theor. Appl. Genet. 48:75-83. Becker, W. A., 1967. Manual of Quantitative Genetic Procedures. 2nd ed. Washington State University, Puyallup, WA. Bentsen, H. B., 1983a. Genetic variation in feed efficiency of laying hens at constant body weight and egg production. I. Efficiency measured as a deviation between observed and expected feed consumption. Acta Agric. Sri. 33: 289-304. Bentsen, H. B., 1983b. Genetic variation in feed efficiency of laying hens at constant body weight and egg production, n. Sources of variation in feed consumption. Acta Agric. Sci. 33:305-320. Bordas, A., and P. Merat, 1984. Correlated responses in a selection experiment on residual feed intake of adult Rhode-Island Red cocks and hens. Ann. Agric. Fen. 23:233-237. Bordas, A., M. Tixier-Boichard, and P. Merat, 1992. Direct and correlated responses to divergent selection for residual food intake in RhodeIsland Red laying hens. Br. Poult. Sci. 33: 741-754. Byerly, T. C , 1941. Feed and other costs of producing market eggs. Bulletin A. I. Maryland Agricultural Experimental Station, College Park, MD. Byerly, T. C , J. W. Kessler, R. M. Gous, and O. P. Thomas, 1980. Feed requirements for egg production. Poultry Sri. 59:2500-2507. Falconer, D. S., 1989. Introduction to Quantitative Genetics. 3rd ed. Longman, England. Fairfull, R. W., and J. R. Chambers, 1984. Breeding for feed efficiency: Poultry. Can. J. Anim. Sci. 64: 513-527. Gowe, R. S., A. Robertson, and B.D.H. Latter, 1959. Environment and poultry breeding problems. 5. The design of poultry control strains. Poultry Sci. 38:462-471. Graser, H.-U., S. P. Smith, and B. Tier, 1987. A derivative-free approach for estimating variance components in animal models by restricted maximum likelihood. J. Anim. Sci. 64:1362-1370. Hagger, C , and H. Abplanalp, 1978. Food consumption records for the genetic improvement of income over food costs in laying flocks of white leghorns. Br. Poult. Sci. 19:651-667. Katie, J., 1986. Results and information from the Norwegian selection experiment on feed efficiency in laying hens. 103rd Nordisk Jordbrugsforskning seminar, Foulum, Denmark.
1994 Poultry Science 73:1479-1484
INTRODUCTION Feed expenses are the main cost in egg production. Therefore, efficient layers are important for the economy of the farm. Traditionally, efficiency has been improved by selecting on egg mass production and body weight and getting a correlated response in feed efficiency (Luiting, 1990). Recently the use of feed consumption data or parameter estimates including feed consumption in the selection have been suggested to be good tools for improving feed efficiency further (Arboleda et al, 1976; Wing and Nordskog, 1982b; Nordskog et al, 1991). Residual feed consumption (RFC) is feed consump-
Received for publication December 27, 1993. Accepted for publication May 30, 1994. ^ o whom correspondence should be addressed.
tion corrected for egg mass production, body weight, and body weight gain (Byerly, 1941; Nordskog et al, 1972; Arboleda et al, 1976; Byerly et al, 1980; Bentsen, 1983a). It measures the remaining part of the variation in feed consumption not accounted for by these three traits. Studying RFC can give valuable information on the effects of using feed consumption data in the selection of layers (Bentsen, 1983b; Luiting and Urff, 1991a). The first estimated heritabilities of RFC were very small (Nordskog et al, 1972; Arboleda et al, 1976). Later, much higher values have been found, ranging from .30 to .60 (Hagger and Abplanalp, 1978; Bentsen, 1983b; Katie, 1986; Luiting and Urff, 1991b). Heritabilities are usually higher at the beginning of the laying period than later (Katie, 1986; Luiting and Urff, 1991b).
1479
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
ABSTRACT The purpose of the study was to estimate the heritability of residual feed consumption (RFC) and the genetic correlations between RFC and economically important traits. The genetic progress after four generations of selection for RFC and the changes in economically important traits were also investigated. A selection experiment for RFC was carried out from 1983 to 1987. The total data consisted of 3,750 birds and 2,661 records. The (co)variance components were calculated using derivative-free bivariate animal model restricted maximum likelihood (REML). Breeding values were estimated for calculating genetic progress in RFC and correlated responses in the other traits. The heritability of RFC calculated from the whole recorded period (16 to 42 wk) and using all 2,661 records was .46 (± .04). The genetic correlations between RFC and egg mass, number of eggs, egg weight, and body weight were not significant. The genetic correlation between RFC and feed consumption was .50 (± .04). The breeding value estimates indicated a moderate genetic progress in RFC due to selection. Feed consumption was decreased and body weight gain showed reduction in the last two generations. No change could be found in egg mass, number of eggs, egg weight, age at first egg, or body weight. (Key words: heritability, residual feed consumption, egglayer, feed conversion efficiency, restricted maximum likelihood)
1480
SCHULMAN ET AL.
MATERIALS AND METHODS Data from a selection experiment for RFC in a Finnish Leghorn population was
used. The selection experiment had been carried out in the years 1983 to 1987 and lasted five generations (Tuiskula-Haavisto, 1986; Liuttula, 1988). The total data consisted of 3,750 birds and 2,661 records. Since 1984, the birds were divided into a selection line and a control line (Table 1). The parents of the base population were known without records. Results of hens that had died or stopped laying in the recorded period were excluded from the analysis (4% of all hens). There were two hatches each in Generations 1, 2, and 3 and only one hatch in Generations 0 and 4. The chicks were reared on the floor for Generations 0, 1 (both hatches), and 3 (first hatch); otherwise, chicks were in cages. At the age of 16 wk, the hens were moved into individual cages. The hens consumed feed ad libitum. The feed used was commercial concentrate supplemented by barley, oats, and calcium. The mixture contained in average 29 g / k g Ca and 6.3 g / k g P. Calculated energy content of the feed was 2.53 kcal/kg and the protein content 16.5% of dry matter. The average temperature during the experiment was about 20 C. The lighting started with 10 h and was gradually increased to 14 h light/d except for the last generation, for which the lighting was increased to 16 h light/d. The measuring period was 26 wk over the period from 16 to 42 wk in age. The whole period was divided into six 4-wk periods and one 2-wk period. The traits measured were age at first egg, number of eggs, egg weight, adult body weight, and feed consumption.
TABLE 1. Number of females and males in the selection and control lines and numbers selected Generation Selection line 0 1 2 3 4 Control line 0 1 2 3 4 iFull brothers were not selected.
Number of females
Selected females
Number of males
Selected males1
485 466 255 561 239
78 55 79 80
292 141 70 156
26 19 19 22
485 168 259 119 109
184 143 132 109
292 51 84 88
86 43 43 41
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
There are not many reports on genetic correlations between RFC and other economically important traits. Luiting and Urff (1991b) found a positive genetic correlation between RFC and feed consumption. The genetic correlations between RFC and egg mass, number of eggs, egg weight, age at first egg, body weight, and body weight gain are all reported to be close to zero (Bentsen, 1983b; Katie and Kolstad, 1991; Luiting and Urff, 1991b). The response to selection for RFC and the effects of selection on other economically important traits have been studied by selection experiments (Bordas and M6rat, 1984; Katie and Kolstad, 1991; Bordas et al., 1992). Residual feed consumption was changed due to selection in these experiments. The aim of this study was to estimate the heritability of RFC in a Finnish Leghorn population and to see whether there are significant genetic correlations between RFC and other economically important traits (e.g., egg mass, number of eggs, egg weight, age at first egg, body weight, body weight gain, and feed consumption). The genetic progress in RFC and the genetic changes in the other traits over four generations of selection for RFC were also investigated. The (co)variance component estimation was carried out using bivariate animal model REML (restricted maximum likelihood).
GENETIC VARIATION OF RESIDUAL FEED CONSUMPTION
E(FC) = a + ba AW + b 2 EM + b 3 BW75
where E(FC) = expected feed consumption (grams per day); AW = body weight gain (grams per day); EM = egg mass production (grams per day); BW75 = metabolic body weight (grams); bi, b2, b 3 = regression coefficients; and a = constant. The hens were selected using their own record for RFC. In the base generation, hens with less than the population mean 30.4 g egg mass/d were excluded from the selection. At the base generation, the males were selected on their full sibs's RFC. Four males were culled because of the extremely low egg mass production of their full sibs. In the subsequent genera-
tions, a selection index including full- and half-sib records was used (Becker, 1967). The heritability used in the index was .24, calculated from the half-sib correlation in the base population. The numbers selected are shown in Table 1. An unselected control line was kept during the experiment. In the control line the aim of the mating system was to increase the effective number of parents and thus to avoid inbreeding by choosing as sires one male offspring from each sire and one female offspring from each dam in the previous generation (Gowe et al, 1959) (Table 1). When producing the second generation, 26 female offspring from as many different full-sib groups were brought from the selection line to the control line because of a failure in the incubation of the control. The (co)variance components were estimated using a REML procedure (Patterson and Thompson, 1971), and the derivativefree method was applied as the computing procedure (Graser et al, 1987; Meyer, 1991). The following animal model was used: y = Xb + Zu + e where y = the vector of n observations of one or two different traits, of which one is always residual feed consumption; b = the vector of fixed generation-hatch classes; u = the vector of random animal effects; e = the vector of residuals; X = the incidence matrix for fixed effects; Z = the incidence matrix for animal effects;
TABLE 2. Regression coefficients used for calculation of expected feed consumption [E(FQ]. The equation where the coefficients were used is E(FC) = a + bj AW + b 2 EM + b 3 BW-75 1 Period
Age
1 2 3 4 5 6 7
(wk) 16 to 20 to 24 to 28 to 32 to 36 to 40 to
20 24 28 32 36 40 42
Constant a 6.48 5.57 7.00 2.04 10.10 7.35 1.91
^
EM b2
BW-75 b3
R2
.75 1.00 .69 .91 1.10 .74 .53
1.00 .46 .47 .62 .64 .83 .90
50.7 52.7 36.4 56.5 43.8 36.1 40.5
58.8 59.5 51.3 56.8 58.6 65.0 68.9
AW
1AW = body weight gain; EM = egg mass; BW-75
=
metabolic body weight.
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
The selection was based on total RFC expressed as a percentage of expected feed consumption. The RFC was calculated at each period as the difference between observed and expected feed consumption (Bentsen, 1983a). Finally, the results for each period were combined to give the total RFC for each hen. The expected feed consumption was derived from multiple regression equations. Every period had its own regression coefficients. The coefficients were calculated from the base population and used unchanged in all the other generations (Table 2), except in the base population, for which only one regression equation covering the whole laying period was used. The regression equation was:
1481
1482
SCHULMAN ET Ah.
and the (co)variances V(u) = G = A <8> G 0 , V(e) = R = I ® RQ, cov (u,e') = 0, and V(y) = V
RESULTS AND DISCUSSION Heritability of Residual Feed Consumption The heritability of total RFC (from 16 to 42 wk of age) estimated for the base population from the whole data (base generation, selection, and control line) was .46 with a standard error of .04. When only animals from the base population and the selection line were used the estimated heritability was .41 (± .06), and when the base population and control line were used it was .48 (± .06). These values are in good agreement with those found in the literature (Hagger and Abplanalp, 1978; Bentsen, 1983b; Luiting and Urff, 1991b). The heritability estimated by Bentsen (1983b) was .50
Genetic Correlations Between Residual Feed Consumption and Economically Important Traits The genetic correlations between RFC and egg mass, number of eggs, egg weight, and body weight were not significant. This is in agreement with the results obtained by Bentsen (1983b) and Luiting and Urff (1991b). However, negative genetic correlations have been also reported between RFC and egg mass, egg production, and egg weight (Fairfull and Chambers, 1984; Katie and Kolstad, 1991). A very small positive genetic correlation between RFC and body weight gain was found (.09 ± .06). Also, Bentsen (1983b) found a very small genetic correlation between body weight gain and RFC. However, nonsignificant correlations have been also reported (Katie and Kolstad, 1991; Luiting and Urff, 1991b). There seemed to be a small negative genetic correlation between RFC and age at first egg (-.11 ± .06). Negative correlations have also been found by Bentsen (1983b). However, a positive genetic correlation between RFC and age at first egg has been also reported (Fairfull and Chambers, 1984). The genetic correlation between RFC and feed consumption was .50 ± .04. This is well in agreement with the literature values reviewed by Luiting (1991). However, Luiting and Urff (1991b) reported a smaller positive value.
Genetic Progress in Residual Feed Consumption The average breeding value estimates showed a moderate genetic progress in RFC from 1983 to 1987 due to selection for RFC as a percentage of expected feed consumption (RFC%) (Figure 1). The genetic progress was in total -6.13 g / h e n per d. This is approximately -1.53 g / h e n per d per generation. In the control line, the RFC stayed close to zero. When the genetic progress was estimated by comparing the
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
where <8> is the Kronecker product; G 0 is the genetic (co)variance matrix between the traits; A is the relationship matrix between animals; and RQ is the residual (co)variance matrix between traits. The variance components were first estimated using univariate procedures and then with multivariate procedures. Only two traits were considered simultaneously in the multivariate procedure. The standard errors of the genetic correlations were calculated according to Falconer (1989). Breeding values for all animals were estimated to compute the genetic progress in RFC and a correlated change in the other traits. The breeding values were estimated using animal model best linear unbiased prediction (AM-BLUP). The same model as for (co)variance component estimation was used. The heritability used was .46. Inbreeding coefficients were obtained as diagonal elements of the numerator relationship matrix computed using the procedure presented by Quaas (1976) and averaged within lines and generations.
(from 16 to 22 wk of age), and the one estimated by Luiting and Urff (1991b) was .48 (from 20 to 44 wk of age). A lower estimate has been reported by Wing and Nordskog (1982a) and a slightly higher estimate by Katie (1986).
GENETIC VARIATION OF RESIDUAL FEED CONSUMPTION
1483
Generation
results of the selection and the control line the progress was estimated as -1.43 g/hen per d per generation. The final inbreeding coefficient was 1.9% in the selection line and .8% in the control. Because the coefficients were low in both lines, there was no reason to assume that inbreeding would have invalidated the use of the control line to estimate genetic trend or reduced the rate of response in the selection line. Moderate response in RFC has been found also in other selection experiments (Bordas and M6rat, 1984; Katie and Kolstad, 1991; Bordas et al, 1992). Genetic Changes in Economically Important Traits The average breeding value estimates showed that feed consumption was decreased due to the positive genetic correlation between RFC and feed consumption (Figure 2). The same kind of response was reported by Bordas et al (1992). In the present experiment, body weight gain tended to decrease slightly in the selection line in the last two generations, which was in agreement with the estimated genetic correlation. In the selection experiment of Bordas et al. (1992) no changes in body weight gain were observed. No genetic changes were found in egg mass, number of eggs, egg weight, age at first egg, or body weight. This is in agreement with the estimated genetic correlations except that the small negative genetic correlation found between RFC and age at first egg did not appear as a change in the age at first egg after selection for RFC as percentage of expected feed consumption.
FIGURE 2. The genetic changes in feed consumption over four generations of selection on residual feed consumption as a percentage of expected feed consumption.
In some selection experiments for RFC, no genetic changes in egg mass production, number of eggs, or body weight gain were found (Bordas and M6rat, 1984; Bordas et al, 1992). Katie and Kolstad (1991) did not observe changes in low lines, but in the high line egg mass was decreased. Inconsistent changes were found in age at first egg and egg weight (Bordas et al, 1992). In the experiment of Katie and Kolstad (1991), body weight was found to increase due to selection for RFC as a percentage of expected feed consumption in low lines, whereas Bordas et al (1992) did not find any differences in body weight between their high and low lines. When the genetic changes were investigated by comparing the phenotypic values of the different traits in the selection and control lines, the results were similar to those above except for body weight gain. This method did not show any change in body weight gain due to selection on RFC as a percentage of expected feed consumption. ACKNOWLEDGMENTS We are very grateful to Anne Nylander for help in doing the REML analysis and to Anni Jarvinen and Laura Lauttamaki for excellent technical assistance.
REFERENCES Arboleda, C. R., D. L. Harris, and A. W. Nordskog, 1976. Efficiency of selection in layer-type chick-
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
FIGURE 1. The genetic progress in residual feed consumption over four generations of selection.
1484
SCHULMAN ET AL. Katie, J., and N. Kolstad, 1991. Selection for efficiency of food utilisation in laying hens: Direct response in residual food consumption and correlated responses in weight gain, egg production and body weight. Br. Poult. Sci. 32:939-953. Liuttula, M., 1988. Selection experiment for feed efficiency in egg laying hens. 144th Nordisk Jordbrugsforskning seminar, Wik, Sweden. Luiting, P., 1990. Genetic variation of energy partitioning in layer hens: causes of variation in residual feed consumption. World's Poult. Sci. 46:133-152. Luiting, P., 1991. The Value of Feed Consumption Data for Breeding in Laying Hens. Ph.D. Thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Luiting, P., and E. M. Urff, 1991a. Optimization of a model to estimate residual feed consumption in the laying hen. Livest. Prod. Sci. 27:321-338. Luiting, P., and E. M. Urff, 1991b. Residual feed consumption in laying hens. 2. Genetic variation and correlations. Poultry Sci. 70:1663-1672. Meyer, K., 1991. Estimating variances and covariances for multivariate animal models by restricted maximum likelihood. Genet. Sel. Evol. 23:67-83. Nordskog, A. W., H. L. French, C. R. Arboleda, and D. W. Casey, 1972. Breeding for efficiency of egg production. World's Poult. Sci. J. 28:175-188. Nordskog, A. W., Y. H. Hou, H. Singh, and S. Cheng, 1991. Selecting for efficiency of egg production using food consumption records in layer-type chickens. Br. Poult. Sci. 32:87-101. Patterson, H. D., and R. Thompson, 1971. Recovery of interblock information when block sizes are unequal. Biometrika 58:545-554. Quaas, R. L., 1976. Computing the diagonal elements of a large numerator relationship matrix. Biometrics 32:949-953. Tuiskula-Haavisto, M., 1986. Resultater og information fra finske forsag med selektion for foderudnyttelse hos verpelwms. 103rd Nordisk Jordbrugsforskning seminar, Foulum, Denmark. Wing, T. L., and A. W. Nordskog, 1982a. Use of individual feed records in a selection program for feed efficiency. I. Heritability of the residual component of feed efficiency. Poultry Sci. 61: 226-230. Wing, T. L., and A. W. Nordskog, 1982b. Use of individual feed records in a selection program for feed efficiency. II. Effectiveness of different selection indexes. Poultry Sci. 61:231-235.
Downloaded from http://ps.oxfordjournals.org/ at University of New Orleans on May 27, 2015
ens by using supplementary information on feed consumption. II Application to net income. Theor. Appl. Genet. 48:75-83. Becker, W. A., 1967. Manual of Quantitative Genetic Procedures. 2nd ed. Washington State University, Puyallup, WA. Bentsen, H. B., 1983a. Genetic variation in feed efficiency of laying hens at constant body weight and egg production. I. Efficiency measured as a deviation between observed and expected feed consumption. Acta Agric. Sri. 33: 289-304. Bentsen, H. B., 1983b. Genetic variation in feed efficiency of laying hens at constant body weight and egg production, n. Sources of variation in feed consumption. Acta Agric. Sci. 33:305-320. Bordas, A., and P. Merat, 1984. Correlated responses in a selection experiment on residual feed intake of adult Rhode-Island Red cocks and hens. Ann. Agric. Fen. 23:233-237. Bordas, A., M. Tixier-Boichard, and P. Merat, 1992. Direct and correlated responses to divergent selection for residual food intake in RhodeIsland Red laying hens. Br. Poult. Sci. 33: 741-754. Byerly, T. C , 1941. Feed and other costs of producing market eggs. Bulletin A. I. Maryland Agricultural Experimental Station, College Park, MD. Byerly, T. C , J. W. Kessler, R. M. Gous, and O. P. Thomas, 1980. Feed requirements for egg production. Poultry Sri. 59:2500-2507. Falconer, D. S., 1989. Introduction to Quantitative Genetics. 3rd ed. Longman, England. Fairfull, R. W., and J. R. Chambers, 1984. Breeding for feed efficiency: Poultry. Can. J. Anim. Sci. 64: 513-527. Gowe, R. S., A. Robertson, and B.D.H. Latter, 1959. Environment and poultry breeding problems. 5. The design of poultry control strains. Poultry Sci. 38:462-471. Graser, H.-U., S. P. Smith, and B. Tier, 1987. A derivative-free approach for estimating variance components in animal models by restricted maximum likelihood. J. Anim. Sci. 64:1362-1370. Hagger, C , and H. Abplanalp, 1978. Food consumption records for the genetic improvement of income over food costs in laying flocks of white leghorns. Br. Poult. Sci. 19:651-667. Katie, J., 1986. Results and information from the Norwegian selection experiment on feed efficiency in laying hens. 103rd Nordisk Jordbrugsforskning seminar, Foulum, Denmark.