Relationship between shell color and incidence of speckles in brown egg lines

Transcription

1 Relationship between shell color and incidence of speckles in brown egg lines J. A. ARANGO*, P. SETTAR, J. A. ARTHUR and N. P. O SULLIVAN Hy-Line International, Dallas Center, Iowa 50263, USA. *Corresponding author: jarango@hyline.com Abstract Long term selection for dark and uniform shell color in brown egg lines has been successful. A side problem is the presence of speckles (SP) or pigment spots, which is undesirable. Studying the genetic determination of SP, and targeting a correlated continuous trait currently used in the selection criteria would be of interest to search for strategies for future selection against incidence of SP. Data were also available for egg shell color (CO). Incidence of SP was measured with an increasing score scale (0 to 5) in four brown-egg lines. On average, each hen had about five records taken during the early and late laying periods, respectively. Total numbers of animals with records ranged from 9,449 to 17,281 depending on the line. Corresponding total number of records ranged from 50,764 to 92,123. Three generations of pedigree information were used in the analyses. Three sets of animal models (repeatability, RE; simple average, AV; and early/late average, EL) were carried out for the largest population using average information restricted maximum likelihood (AIREML). Repeatability and AV models were bivariate SP-CO, while EL was a four-trait model defining early (E) and late (L) records as separate traits for SP and CO. Models included the fixed effect of hatch- generation and the random animal genetic effect; in addition, RE models included the animal s permanent environmental effect. Results from the EL analyses indicated high genetic correlation between early and late CO (0.95) and SP (0.94), indicating that combining information would be reasonable. Heritability and repeatability estimates were larger for CO (h 2 = 0.21 to 0.52; re = 0.47 to 0.70) than for SP (h 2 = 0.15 to 0.22; re = 0.33 to 0.42) in all lines. Genetic correlations between CO and SP were all positive, ranging from 0.35 to Corresponding residual correlation were about zero across lines. Extreme selection in favor of intense color may increase the incidence of SP in the long term; however, the low frequency of extreme SP scores indicate that the current selection intensity for CO is not creating a problem with incidence of SP. Selection for increased CO and against SP has been successful in producing dark uniform brown eggs relatively free of speckles. Keywords: laying hens; brown eggs; speckles; color. Introduction In nature, egg pigmentation serves as camouflage, and it may have a role in thermoregulation (Solomon, 1997). Most wild bird species show some pattern of speckling. Response to selection for shell color in brown-egg layer lines has been successful in providing darker and uniformly colored eggs; however, a side problem for marketing purposes is an increased incidence of spots or speckles. It is desirable to apply selection to produce darker but spot-free eggs. To do that, estimation of variance component and genetic parameters for the incidence of speckles is needed. It is important to

2 determine the extent to which the incidence of speckles is genetically related to shell color in order to establish an adequate selection strategy aiming to ensure steady progress for egg color while keeping an acceptable product for the public. The objectives of the research herein were to estimate variance components and other parameters for the incidence of speckles and to determine its genetic relationship with shell color in four established brown-egg lines. Materials and methods Data. Data and genealogical pedigree information were compiled from the research computing system and database at Hy-line International (Dallas Center, IA, USA) from four established brown-egg layer lines (herein coded as 1 to 4). Traits included: (1) shell color (CO) measured by light absorbance with a Minolta Chroma Meter CR400 (Minolta Inc.), which determines values of luminance or lightness (l*), chromatic component red to green (a*) and chromatic component blue to yellow (b*), and computes a color index L*a*b, CO, and (2) incidence of speckles (SP), measured on an increasing scale from 0 (clean) to 5 (intense spotting) by subjective scoring. Three generations of data and full pedigree information were used in the analysis. Total numbers of animals with records ranged from 9,449 to 17,281 depending on the line (Table 1). Corresponding total number of records ranged from 50,764 to 92,123. On average, each hen had about five records (2 to 14) taken during the early (25 wk of age) and late (40 wk of age) laying evaluations, respectively. Many hens with early evaluations did not have late evaluations, due to selection in the interim. Statistical analysis. Preliminary analyses were carried out to test the association between shell color and speckles at the phenotypic level using least square analysis. There were significant differences in shell color for different speckle scores across lines. A set of linear animal models were implemented to jointly evaluate the co-inheritance of SP and CO using average information restricted maximum likelihood (AIREML) in the family of BLUPF90 programs (Misztal et al., 2002). Models included: (1) Repeatability model to estimate the relative importance of additive genetic vs. animal permanent environmental effects for SP score, and to estimate repeatability. (2) Average-record models to explore a simplified implementation aiming to reduce number of records. (3) Within the former implementations (1 and 2), early and late evaluation were initially considered as being different traits to estimate genetic and residual correlations between the two evaluation periods. All models included the fixed effect of hatch-generation. For (1) and (2) separate analyses were carried out for early (CO_E, SP_E) and late (CO_L, SP_L) evaluation. A series of multiple trait analyses were carried out to explore combinations of traits. Here, only the full four-trait model (CO_E, SP_E, CO_L, SP_L) with averaged records, and the combined repeatability model, including a fixed factor to account for the period of evaluation, are presented due to length restrictions. Table 1. Data description and summary statistics for four lines (1 to 4) Line Animals with records 10,721 15,894 17,281 9,449 Total records 53,481 87,503 92,123 50,764 Sires Dams 1,454 1,635 1,268 1,032 Color average (SD) (7.05) (7.02) (6.93) (7.33) Speckles average (SD) 0.28 (0.68) 0.36 (0.85) 0.51 (1.10) 0.71 (1.08)

3 Results and discussion The estimates of heritability for SP score varied with statistical model and evaluation period. In general, heritability was of moderate magnitude, and tended to be larger for late evaluation than for early evaluation. Results for the full four-trait model for the averaged records in the largest population (Line 3) are summarized in Table 2. Genetic correlation between early and late SP scores was large (0.94), and similar to that for CO (0.95). For the purpose of practical selection, that favors the use of pooled data to implement a simpler repeatability model for breeding value prediction. The genetic correlations between CO and SP score were moderate and positive in both evaluation periods but were slightly lower for early (0.36) than for late (0.40) evaluation. The corresponding residual correlations tended to be low (0.09 to 0.16), as suggested in preliminary least square analysis. This indicates that intense selection for increased shell color may have negative long term effect on the incidence of speckles. Selection for shell color is more efficient than that for incidence of speckles; therefore, speckles should be monitored during selection to avoid extremes. Table 2. Estimates of (co)variance components and other parameters for the four-trait model for Line 3 (Co)variance a σ 2 a or σ a1,a2 h 2 or σ 2 e or σ e1,e2 Mean se r a1,a2 Mean se e 2 or r e1,e2 CO_E CO_L SP_E SP_L CO_E, CO_L CO_E, SP_E CO_E, SP_L CO_L, SP_E CO_L, SP_L SP_E, SP_L a σ 2 a = additive genetic variance; σ 2 e = residual variance; σ a1,a2 = genetic covariance; σ e1,e2 = residual covariance; h 2 = heritability; r a1,a2 = genetic correlation; e 2 = residual variance as proportion of total variance; r e1,e2 = residual correlation. Results from the combined (early-late) repeatability model for all lines are shown in Table 3. Part of the variance associated with animal additive effects in the average-record analysis corresponded to permanent environmental variance for both traits. Estimates of heritability for CO (0.21 to 0.52) tended to be greater than those for SP score (0.15 to 0.22) across lines. Likewise, repeatability estimates for CO were larger (0.47 to 0.70) that those for SP (0.33 to 042), as expected from a trait based on subjective scoring; that was previously indicated by a large within hen variation for SP scores in preliminary leastsquare analyses.

4 Variance components and parameter estimates for shell color are scarce in the literature. In a recent review (Szwaczkowski, 2003) there was only one study for shell color. It was the first report using REML animal model, and found heritability estimates of 0.27 to 0.53, using the average of three records at 39 wk of age of each hen, in three Catalan breeds (Francesh et al., 1997). A more recent study in brown-egg dwarf layers, used the same approach but at 40 wk of age, and reported a heritability estimate of 0.46 for shell color (Zhang et al., 2005). No reports were found for variance components and parameter estimates for the incidence of speckles or related traits in the literature. There is sizeable variation for SP scoring across lines. The heritability estimates for SP varied with line, and tended to be low to moderate (0.15 to 0.22); however, these values are acceptable, considering the skewed distribution of the trait. Selection against speckles could be implemented in situations with large SP incidence; however, the selection response would be slow. On the other hand, repeatability of SP scoring showed moderate values (0.33 to 0.42), indicating that there would be a considerable advantage of using repeated records if a selection strategy to control SP is to be applied. Table 3. Estimates of (co)variance components and other parameters by line for the repeatability model σ 2 a or σ a1,a2 h 2 or σ 2 e or σ e1,e2 e 2 or σ 2 c (Co)variance a Mean se r a1,a2 Mean se r e1,e2 Mean se re Line 1 CO SP CO, SP CO SP CO, SP CO SP CO, SP CO SP CO, SP a σ 2 a = additive genetic variance; σ 2 e = residual variance; σ a1,a2 = genetic covariance; σ e1,e2 = residual covariance; ; h 2 = heritability; r a1,a2 = genetic correlation; e 2 = residual variance as proportion of total variance; r e1,e2 = residual correlation; σ 2 c = permanent environmental variance; re = repeatability.

5 The genetic correlation between CO and SP was always positive and moderate (0.35 to 0.54) across lines. Therefore, when applying selection to either CO or SP the other trait has to be monitored to avoid an undesirable long term effect due to correlated response. Concerning the biology of the traits, it seems logical that overall shell color and the incidence of speckling should be correlated since both involve increased elaboration and deposition of pigment. However, since there are eggs with dark brown shells and no speckling, there must some characteristic of the shell gland which can cause uneven deposition. That is consistent with a cyclical pattern of synthesis and release of pigment in the surface epithelial cells of the shell gland pouch (Solomon, 1997). Perhaps if this flaw is present it may be more obvious in birds that produce high levels of the pigment. If that is true, it would contradict the low residual (0.00 to 0.05) and phenotypic (0.06 to 0.11) correlation between CO and SP (i.e., a non apparent association at level of trait measurements). That could be, in part, a consequence of the poor data structure since the genetic model tended to show the underlying genetic correlation quite consistently. Extreme selection in favor of intense color may increase the incidence of SP in the long term; however, the low frequency of extreme SP scores indicate that the current selection intensity for CO is not creating a problem with incidence of SP in the four lines under consideration in the present study. Selection for increased CO and against SP has been successful in producing dark uniform brown eggs relatively free of speckles. References Francesch, A., Estany, J., Alfonso, L. and Iglesias, M. (1997) Genetic parameters fro egg number, egg weight, and eggshell color in three Catalan poultry breeds. Poultry Science. 76: Misztal, I., Tsuruta, S., Strabel, T., Auvray, B., Druet, T. and Lee, D.H. (2002) BLUPF90 and related programs (BGF90). Proc. 7th World Congress on Genetics Applied to Livestock Production. Montpellier, France. Communication No Solomon, S.E. (1997) Egg & Eggshell Quality. Manson Publishing/The Veterinary Press Iowa State University Press/Ames. Szwaczkowski, T. (2003) Use of Mixed Model Methodology in Poultry Breeding: Estimation of genetic parameters. In: Poultry Genetics, Breeding and Biotechnology (W.M. Muir and S.E. Aggrey eds.), pp CABI Publishing. Zhang, L.-C., Hing, Z.-H., Xu, G.-Y., Hou, Z.-C. and Yang, N. (2005) Heritabilities and genetic and phenotypic correlations of egg quality traits in brown-egg dwarf layers. Poultry Science. 84: