Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Genetics

µ-Calpain, calpastatin, and growth hormone receptor genetic effects on preweaning performance, carcass quality traits, and residual variance of tenderness in Angus cattle selected to increase minor haplotype and allele frequencies123

 

This article in JAS

  1. Vol. 92 No. 2, p. 456-466
     
    Received: Aug 26, 2013
    Accepted: Dec 15, 2013
    Published: November 24, 2014


    5 Corresponding author(s): gary.bennett@ars.usda.gov
 View
 Download
 Share

doi:10.2527/jas.2013-7075
  1. R. G. Tait Jr.,
  2. S. D. Shackelford,
  3. T. L. Wheeler,
  4. D. A. King,
  5. E. Casas44,
  6. R. M. Thallman,
  7. T. P. L. Smith and
  8. G. L. Bennett 5
  1. USDA, ARS, U.S. Meat Animal Research Center, Clay Center, NE 68933-0166

Abstract

Genetic marker effects and interactions are estimated with poor precision when minor marker allele frequencies are low. An Angus population was subjected to marker assisted selection for multiple years to increase divergent haplotype and minor marker allele frequencies to 1) estimate effect size and mode of inheritance for previously reported SNP on targeted beef carcass quality traits; 2) estimate effects of previously reported SNP on nontarget performance traits; and 3) evaluate tenderness SNP specific residual variance models compared to a single residual variance model for tenderness. Divergent haplotypes within µ-calpain (CAPN1), and SNP within calpastatin (CAST) and growth hormone receptor (GHR) were successfully selected to increase their frequencies. Traits evaluated were birth BW, weaning BW, final BW, fat thickness, LM area, USDA marbling score, yield grade, slice shear force (SSF), and visible and near infrared predicted slice shear force. Both CAPN1 and CAST exhibited additive (P < 0.001) modes of inheritance for SSF and neither exhibited dominance (P ≥ 0.19). Furthermore, the interaction between CAPN1 and CAST for SSF was not significant (P = 0.55). Estimated additive effects of CAPN1 (1.049 kg) and CAST (1.257 kg) on SSF were large in this study. Animals homozygous for tender alleles at both CAPN1 and CAST would have 4.61 kg lower SSF (38.6% of the mean) than animals homozygous tough for both markers. There was also an effect of CAST on yield grade (P < 0.02). The tender CAST allele was associated with more red meat yield and less trimmable fat. There were no significant effects (P ≥ 0.23) for GHR on any of the traits evaluated in this study. Furthermore, CAST specific residual variance models were found to fit significantly better (P < 0.001) than single residual variance models for SSF, with the tougher genotypes having larger residual variance. Thus, the risk of a tough steak from the undesired CAST genotype is increased through both an increase in mean and an increase in variation. This work confirms the importance of CAPN1 and CAST for tenderness in beef, provides a new effect of CAST on beef tenderness, and questions the utility of GHR as a selection marker for beef quality.



INTRODUCTION

Genetic markers are important tools for selecting traits that are expensive to measure, measured postmortem, or measured in older beef cattle. Koohmaraie (1996) identified the protease µ-calpain (CAPN1) and its inhibitor calpastatin (CAST) as major factors affecting post mortem tenderization in meat. Genetic tests based on genetic variation in genes coding for these proteins have been identified and are available commercially in beef cattle (Page et al., 2004; White et al., 2005; Casas et al., 2006). Genetic variants in the growth hormone receptor (GHR) gene have also been associated with effects on BW in beef cattle (Sherman et al., 2008).

Low frequency of an allele is a potential reason for limited ability to detect an effect for a genetic marker (Page et al., 2002). Studies that survey industry populations often encounter small numbers of the rare homozygous genotype class and do not include that genotype in their analyses (e.g., CAPN1White et al. [2005]; CASTMorris et al. [2006]; GHRWhite et al. [2007]). Selection to increase minor allele frequency is one method to ensure representation of all genotypes in data for analysis.

Discovery of genetic markers associated with a trait leads to a selection emphasis on those marker effects for that trait. However, characterization of genetic marker effects on nontarget traits is also important. Knowledge of marker effects on all traits allows incorporation of those markers into the multi-trait breeding objective.

The objectives of this study were to utilize a population of cattle selected to increase minor haplotype or allele frequency with increased observations of rare homozygous genotypes to 1) estimate effect size and mode of inheritance for previously reported SNP on targeted beef carcass quality traits; 2) estimate effects of previously reported SNP on nontarget performance traits; and 3) evaluate tenderness SNP specific residual variance models for tenderness compared to single residual variance models.


MATERIALS AND METHODS

The U.S. Meat Animal Research Center (USMARC) Animal Care and Use Committee approved the procedures used in this experiment.

Cattle Population

From 1992 to 1999, an Angus population at USMARC was used in a selection experiment comparing calving ease selected and control animals (Bennett, 2008). Subsequently, the calving ease selection and control line cows were bred to the same bulls with cows and their progeny treated as a single population for utilization in this experiment. About 210 calves were born each year. Approximately 15 sires born within the herd were used each year for AI and natural service. The pedigree used for this study included 7,433 animals recorded, with the oldest animal born in 1963.

Genetic Markers

Markers chosen for this experiment were SNP haplotypes in CAPN1 and single SNP within CAST and GHR. The CAPN1 haplotype evaluated in this study was based on 2 previously identified SNP: CAPN1-316 (BTA 29; rs17872000; Page et al., 2002) and CAPN1-4751 (BTA 29; rs17872050; White et al., 2005). CAPN1-316 segregates C and G alleles, whereas CAPN1-4751 segregates C and T alleles. CAPN1-316 and CAPN1-4751 SNP were used to define haplotypes within the CAPN1 gene. Haplotypes of interest in this study were: CAPN1-316 allele C with CAPN1-4751 allele C (CAPN1hCC) and CAPN1-316 allele G with CAPN1-4751 allele T (CAPN1hGT). Haplotypes not of interest in this study were: CAPN1-316 allele C with CAPN1-4751 alllele T (CAPN1hCT) and CAPN1-316 allele G with CAPN1-4751 allele C (CAPN1hGC). Amongst CAPN1 haplotypes with >1% frequency, White et al. (2005) reported the largest difference for 14 d Warner-Bratzler shear force (WBSF) between CAPN1hCC haplotype and CAPN1hGT haplotype, neither of which were the major haplotype in that study. Therefore, divergent haplotypes were selected to increase their frequency in this population. Additionally, a SNP in CAST (BTA7; rs109221039; Casas et al., 2006) segregating C (CASTaC) and T (CASTaT) alleles was selected to increase the frequency of CASTaC in this population. Third, a SNP in GHR (BTA20; rs385640152) segregating T (GHRaT) and A (GHRaA) alleles (White et al., 2007) causing a phenylalanine to tyrosine substitution at the 279th amino acid (Blott et al., 2003) was selected to increase the frequency of GHRaT in this population.

Samples of DNA were extracted from blood or semen. Extraction of DNA was done using a Qiagen QIAmp DNA mini blood kit (Qiagen, Valencia, CA). Blood samples were collected in 10 mL syringes with 4% EDTA. Blood was frozen until DNA was extracted. Genotyping was performed using a primer extension method with mass spectrometry-based analysis of the extension products on a MassArray system as suggested by the manufacturer (Sequenom, Inc., San Diego, CA) and described by Stone et al. (2002). When necessary, genotype assays were repeated to reduce missing genotypes.

Base, Selection, and Evaluation Phases

This experiment was a complement to the marker assisted selection experiment described by Bennett et al. (2012). Briefly, the experiment consisted of 3 phases: base, selection, and evaluation. Angus was selected from the four candidate populations (Angus and 3 stable composite [MARC I, MARC II, and MARC III] populations) for increasing both CAPN1hCC and CAPN1hGT haplotypes, CASTaC, and GHRaT alleles because of greater base population frequencies for these markers. Base population (birth years 2003 to 2005) haplotype and allele frequencies (Fig. 1) were: CAPN1hCC = 0.456; CAPN1hGT = 0.242; CASTaC = 0.080; and GHRaT = 0.182.

Figure 1.
Figure 1.

Haplotype or allele frequency by birth year in an Angus population selected to increase CAPN1hCC haplotype, CAPN1hGT haplotype, CASTaC allele, and GHRaT allele using marker assisted selection.

 

In the selection phase (birth years 2006 to 2008), the goal was to increase frequencies of CAPN1hCC, CAPN1hGT, CASTaC, and GHRaT towards 0.50. Calves were genotyped before weaning so that replacement bulls and heifers could be selected soon after weaning. Replacements and USMARC bred AI sires were selected to increase the frequencies of the desired genotypes in the population.

In the evaluation phase (birth years 2009, 2010, and 2011), sires mostly heterozygous for CAPN1, CAST, and GHR markers were bred to heifers and cows whose frequencies were near 0.50. Haplotype and allele frequencies achieved in the evaluation phase (Fig. 1) were: CAPN1hCC = 0.530; CAPN1hGT = 0.363; CASTaC = 0.348; and GHRaT = 0.421. After data edits (described later) this study utilized 465 bulls and heifers born in the evaluation phase. These cattle were progeny of 25 sires and 252 dams. During the evaluation phase, replacement bulls were randomly selected post-weaning from bulls heterozygous for CAPN1 haplotypes and SNP in CAST and GHR. Dams ranged in age from 2 to 12 yr, however, for analysis they were defined as 2, 3, 4, or ≥5 yr.

Phenotype Collection

All bulls and heifers had BW collected at birth (n = 465) and weaning (n = 464, average age = 170 d, SD = 22.5 d). After weaning, unselected bulls (n = 199) were castrated and fed diets based on corn and corn silage until harvest to characterize carcass quality attributes. For steers, a final BW was collected on a single day (5 to 11 d, depending on year) before harvest. All steers were harvested on a single day within each year at a commercial abattoir (average age = 433 d, SD = 21.5 d). Genotype frequencies for all 3 markers by trait category (preweaning or carcass traits) are reported in Table 1.


View Full Table | Close Full ViewTable 1.

Evaluation phase genotype frequency by trait category for CAPN1, CAST, and GHR in an Angus population selected to increase minor genotype and haplotype frequencies

 
Genotype Preweaning traits, n = 465 (%) Carcass quality traits, n = 199 (%)
CAPN11
CAPN1hCC:CAPN1hCC 166 (35.7) 77 (38.7)
CAPN1hCC:CAPN1hGT 228 (49.0) 94 (47.2)
CAPN1hGT:CAPN1hGT 71 (15.3) 28 (14.1)
CAST2
CASTaC:CASTaC 49 (10.5) 22 (11.1)
CASTaC:CASTaT 238 (51.2) 82 (41.2)
CASTaT:CASTaT 178 (38.3) 95 (47.7)
GHR3
GHRaA:GHRaA 153 (32.9) 78 (39.2)
GHRaA:GHRaT 240 (51.6) 88 (44.2)
GHRaT:GHRaT 72 (15.5) 33 (16.6)
1Haplotype of 2 SNP within the CAPN1 gene. CAPN1hCC = CAPN1-316 C allele with CAPN1-4751 C allele and CAPN1hGT = CAPN1-316 G allele with CAPN1-4751 T allele.
2CASTaC = C allele of CAST SNP and CASTaT = T allele of CAST SNP
3GHRaA = A allele of GHR SNP and GHRaT = T allele of GHR SNP

For carcass traits, carcasses were weighed hot, electrically-stimulated, and chilled using the commercial facility’s proprietary system. At 36 h postmortem, carcasses were ribbed between the 12th and 13th ribs and an image analysis based grading system (VBG2000; Shackelford et al., 2003) assessed adjusted fat thickness, LM area, USDA marbling score, and calculated vision yield grade. Meat tenderness was predicted at the abattoir using visible and near-infrared reflectance spectroscopy at 36 h postmortem (VISNIR; Shackelford et al., 2012a,b). A LM steak from the 13th rib region was returned to USMARC to evaluate slice shear force (SSF) at 14 d postmortem (Shackelford et al., 1999).

Statistical Analysis

All genetic marker and pedigree information was analyzed with GenoProb, a software implementation of algorithms described by Thallman et al. (2001a,b). In cases where GenoProb identified genotyping errors, GenoProb indicated genotypes were used for analyses. GenoProb provides ordered genotypes (identifying which allele came from the sire and which from the dam) for each individual at each SNP. The GenoProb ordered genotypes were used to create the CAPN1 haplotypes used for analysis. Candidate evaluation phase animals were removed from analyses if 1) marker genotype information was missing for CAST, GHR, or both CAPN1 markers (n = 6 and 2 for preweaning and carcass traits, respectively); or 2) Genoprob ordered genotypes indicated an unintended haplotype present for CAPN1 (CAPN1hGC or CAPN1hCT; n = 97 and 12 for preweaning and carcass traits, respectively).

Genotype Effects on Means.

Genetic markers were recoded to be numeric class effect terms. The CAPN1hCC homozygote was assigned 1, CAPN1hCC:CAPN1hGT heterozygote was assigned 2, and CAPN1hGT homozygote was assigned 3. The CASTaC homozygote was assigned 1, CASTaC:CASTaT heterozygote was assigned 2, and CASTaT homozygote was assigned 3. The GHRaA homozygote was assigned 1, GHRaA:GHRaT heterozygote was assigned 2, and GHRaT homozygote was assigned 3.

All traits were analyzed with a mixed model using MTDFREML (Boldman et al., 1995). The model for preweaning traits was:where Yi,j,k,l,m,n,o is the observation for animal o, µ is the mean, Yeari is the birth year (2009, 2010, or 2011), Aod5j is the age of dam (2, 3, 4, or ≥ 5 yr), Sexk is the sex of the calf from birth through weaning (bull or heifer), b0 is the estimated covariate effect of age, Ageo is the age (d) of animal o at the time of phenotype collection, CAPN1l is the CAPN1 genotype (1, 2, or 3), CASTm is the CAST genotype (1, 2, or 3), GHRn is the GHR genotype (1, 2, or 3), Ao is the additive polygenic effect of animal o, and ei,j,k,l,m,n,o is the residual effect of the i,j,k,l,m,n,o-th observation. The distribution of polygenic effects was assumed to be proportional to the pedigree relationship matrix and the residual effects were assumed independent with constant variance. Variance estimates for additive polygenic effects were expected to be imprecise because of the limited number of observations so h2 = σ2a/(σ2a + σ2e) was constrained to 0.20 ≤ h2 ≤ 0.70, consistent with the limits imposed by Bennett et al. (2012). Carcass traits were modeled with a similar model, excluding the Sexk effect because only steers were evaluated for carcass traits.

Markers were initially evaluated using the significance of genotype effects for each trait. When the genotype effect was significant (P < 0.05), linear contrast statements for additive (-0.5, 0, 0.5) and dominance (-0.5, 1, -0.5) modes of inheritance were evaluated for significance and estimation of effect. In the case where CAPN1 and CAST were both significant (only SSF), the interaction between CAPN1 and CAST was added to the model and tested for signficance.

Genotype Effects on Residual Variance.

We further analyzed SSF for evidence of CAPN1 or CAST genotype specific residual variance. Additive genetic relationships amongst animals with phenotypes were calculated using the full pedigree and the INBREED procedure of SAS version 9.3 (SAS Inst. Inc., Cary, NC). These relationships were scalar multiplied by the additive genetic variance for SSF from the converged MTDFREML analysis to calculate 2g. A single residual variance model was analyzed with the MIXED procedure of SAS. This model fit the same model as MTDFREML, with the 2g matrix provided as the animal random effect (using the GDATA = option). Genotype specific residual variance models were analyzed by adding heterogeneous residual variances based on CAPN1 or CAST genotypes (using the REPEATED statement with GROUP = genotype option) to the MIXED analysis. A likelihood ratio test was calculated between the genotype specific residual variance and the single residual variance model, and compared to a χ2 distribution with 2 degrees of freedom, because 2 additional parameters were estimated under the heterogeneous variance models.


RESULTS AND DISCUSSION

Means and SD for performance and carcass characteristics are presented in Table 2. In comparison to typical U.S. beef industry harvest endpoints (Gray et al., 2012), these cattle had less HCW, less LM area, more fat, and slightly more marbling.


View Full Table | Close Full ViewTable 2.

Angus cattle evaluation phase performance and carcass quality trait means and SD

 
Trait n Mean SD
Birth BW, kg 465 34.7 4.8
Weaning BW, kg 464 196.5 25.3
Yearling BW, kg 199 474.6 50.5
Final BW, kg 199 567.2 43.1
Dressing percent, % 199 61.6 1.7
HCW, kg 199 349.8 30.4
Adjusted fat thickness, mm 199 15.6 4.2
Marbling score1 199 5.67 0.85
LM area, sq. cm 199 73.0 6.1
Vision yield grade 199 3.77 0.65
Slice shear force, kg 199 11.94 2.39
VISNIR2 predicted shear force, kg 199 13.03 1.65
14.00 = Slight00; 5.00 = Small00; 6.00 = Modest00 (USDA, 1997; BIF, 2010).
2Visible and near-infrared reflectance spectroscopy (VISNIR; Shackelford et al., 2012a,b).

Marker Effects on Trait Means

Effects on target traits.

Effects of CAPN1 and CAST on SSF were significant (P < 0.001 for both; Table 3). Furthermore, the inheritance pattern was additive (P < 0.001) for both CAPN1 and CAST with a lack of significance for dominance inheritance (P ≥ 0.19) on SSF (Table 4). The additive effect of the CAPN1hGT haplotype in comparison to the CAPN1hCC haplotype was 1.049 kg for SSF. A somewhat larger additive effect on SSF for CASTaC of 1.257 kg compared to the CASTaT was identified (Table 5). Direction of the effects on SSF for both markers was consistent with previous research (Page et al., 2002; White et al., 2005; Casas et al., 2006). The relative size of the CAST additive effect compared to CAPN1 haplotype additive effect on SSF in this study (119.8%) is larger than the relative effects reported by Van Eenennaam et al. (2007) on WBSF of 44.1% or 57.6%, dependent on CAST genetic marker evaluated.


View Full Table | Close Full ViewTable 3.

Significance (P-values) for sources of variation and heritability estimates for performance and carcass quality traitswith values ≤ 0.05 in bold

 
Trait Year Dam age Calf age1 Calf sex CAPN1 CAST GHR h2 ± SE
Birth BW < 0.001 < 0.001 < 0.001 < 0.001 0.19 0.31 0.54 0.54 ± 0.14
Weaning BW < 0.001 < 0.001 < 0.001 < 0.001 0.72 0.09 0.58 0.24 ± 0.14
Yearling BW 0.03 < 0.001 < 0.001 2 0.71 0.32 0.40 0.203 ± 0.18
Final BW 0.01 < 0.001 < 0.001 0.33 0.48 0.23 0.24 ± 0.17
Dressing percent < 0.001 0.72 < 0.01 0.43 0.24 0.82 0.34 ± 0.20
Hot carcass weight < 0.01 < 0.001 < 0.001 0.57 0.31 0.24 0.22 ± 0.17
Adjusted fat thickness < 0.001 0.14 < 0.001 0.59 0.06 0.35 0.26 ± 0.21
Marbling score < 0.001 0.52 < 0.01 0.31 0.91 0.53 0.26 ± 0.22
LM area < 0.001 0.38 0.21 0.69 0.16 0.82 0.703 ± 0.22
Vision yield grade 0.03 0.07 < 0.001 0.77 0.02 0.33 0.37 ± 0.22
Slice shear force < 0.01 0.43 0.52 < 0.001 < 0.001 0.56 0.32 ± 0.22
VISNIR4 predicted shear force 0.03 0.86 0.54 0.53 0.16 0.27 0.203 ± 0.20
1Julian birth day linear covariate used for birth BW; linear age (d) covariate used for all other traits.
2Postweaning traits were only measured on steers, so sex effect was not included in the model.
3Constrained to 0.20 ≤ h2 ≤ 0.70.
4Visible and near-infrared reflectance spectroscopy (VISNIR; Shackelford et al., 2012a,b)

View Full Table | Close Full ViewTable 4.

Significance (P-Values) for additive or dominance effects of CAPN1 or CAST on traits where either marker had significant effect with values ≤ 0.05 in bold

 
CAPN1
CAPN1
CAST
CAST
Trait Additive effect Dominance effect Additive effect Dominance effect
Vision Yield Grade cNS1 NS1 < 0.01 0.45
Slice shear force < 0.001 0.19 < 0.001 0.43
1NS = Nonsignificant (P = 0.77) effect.

View Full Table | Close Full ViewTable 5.

Significant estimated additive effects and SE for CAPN1 and CAST genetic markers on carcass quality traits

 
CAPN1
CAST
Trait additive effect/GT haplotype additive effect/T allele
Vision yield grade NS1 -0.198 ± 0.071
Slice shear force, kg 1.049 ± 0.246 -1.257 ± 0.261
1NS = Nonsignificant (P = 0.77) genetic effect.

Direct comparison of our results to other studies is challenging because most studies have used WBSF or tenderometer to measure tenderness instead of SSF. However, studies have reported genetic marker effects relative to the mean, genetic SD, or phenotypic SD of the trait. The CAPN1hGT haplotype additive effect on SSF was 8.7% of the mean, 85.0% of the genetic SD, and 43.9% of the phenotypic SD. Van Eenennaam et al. (2007) reported the largest CAPN1 haplotype differences to be between the same CAPN1 haplotypes that we evaluated and the effect was found to be 22.1% or 22.5% of the phenotypic SD, depending on subpopulation evaluated. Barendse et al. (2007) reported an allele substitution effect of the CAPN1-316 marker of 13% of the phenotypic SD for tenderness. Similar to our results, in a study utilizing Angus cattle, Gill et al. (2009) did not identify a dominant mode of inheritance for the CAPN1-316 marker. Furthermore, Gill et al. (2009) identified the effect of CAPN1-316 to be 5.9% of the mean for tenderometer and 3.3% of the mean for tenderness evaluated by taste panel. Gill et al. (2009) did not find an association between CAPN1-4751 and tenderness measures. The results of Gill et al. (2009) are not surprising, because CAPN1-316 has been more consistently associated with tenderness in cattle of Bos taurus decent, whereas the CAPN1-4751 SNP has been more consistently associated with tenderness in cattle of Bos indicus decent. When Gill et al. (2009) tried to add CAPN1 haplotype information to their model, there was no improvement in the prediction capabilities, again reinforcing the effectiveness of the CAPN1-316 allele in their population. Because our population was selected for the haplotypes rather than single SNP, we likely changed the allele frequency of the causative mutation more than if single SNP selection were applied.

The CASTaC additive effect on SSF was 10.5% of the mean, 101.9% of the genetic SD, and 52.6% of the phenotypic SD. Van Eenennaam et al. (2007) reported CAST allele effects to be 9.9% or 12.6% of phenotypic SD for WBSF, depending on which CAST genetic marker was evaluated. In comparison, Schenkel et al. (2006) found a CAST allele substitution effect to be 25% to 20% of phenotypic SD (and speculated the effect to be 14% to 29% of genetic SD) for WBSF with 2 to 21 d of aging. Barendse et al. (2007) indicated a CAST SNP allele substitution effect accounted for up to 14% of phenotypic SD for tenderness. The effects of CAPN1 and CAST genetic markers estimated in this study are likely increased when expressed on a percent of mean or SD basis for 2 reasons: 1) we have reported the additive genetic effects rather than allele substitution effects as reported by others, and 2) this population is more tender and less variable for SSF than previously reported populations (Shackelford et al., 2012a).

In this study, the interaction between CAPN1 and CAST genotypes for SSF was not significant (P = 0.55). While Casas et al. (2006) did report an interaction for WBSF between the CAPN1-4751 SNP selected as part of the haplotype in this study and the same CAST SNP, they readily admitted the relatively few number of observations within some of the genotype combinations as a possible cause for a spurious result. Even in the current study where rare alleles were selected to increase frequency, interaction testing between CAPN1 haplotypes and CAST genotypes is challenging (i.e., CAPN1hGT homozygote in combination with CASTaC homozygote only represented 1% of observations).

The VISNIR predicted shear force was not significantly affected by either CAPN1 or CAST genetic effects (Table 3). Predictions using VISNIR technology appear to identify tenderness characteristics associated with sarcomere length more than postmortem proteolysis (Shackelford et al., 2012a). Therefore, VISNIR tenderness prediction may not accurately detect proteolysis differences caused by CAPN1 and CAST.

There were no significant effects (P ≥ 0.23) for GHR on any of the traits in this study (Table 3). This is in contrast to the results of White et al. (2007) where GHR was associated with several carcass quality traits in 1 population, but not in a second population evaluated. However, White et al. (2007) excluded GHRaT homozygotes from their analysis for both populations. Consistent with our results, Gill et al. (2009) reported no association between the same GHR mutation and carcass quality traits in common between the studies (only 1 association with the taste panel trait of odor). Whereas, Sherman et al. (2008) evaluated several SNP within GHR and identified associations with BW traits, those associations were with SNP in the promoter and intron 4 regions. There were no associations between any of the 3 SNP evaluated by Sherman et al. (2008) in exon 10 (the SNP evaluated in this study was in exon 7) and any carcass or weight traits.

Effects on nontarget traits.

This study reveals one significant association between a genetic marker and a nontarget trait (Table 3): CAST and vision yield grade (P = 0.02). In this work, the effect of CAST on vision yield grade exhibits an additive mode of inheritance (P < 0.01) and not a dominance mode of inheritance (P = 0.45, Table 4). Furthermore, the tender allele of CAST (CASTaT) was associated with more red meat yield and less trimmable fat (Table 5). The vision yield grade effect is a combination of a suggestive (P = 0.06, Table 3) effect of adjusted fat thickness and a complementary numeric estimate (P = 0.16, Table 3) of LM area for the CAST tender allele. However, this relationship is in contrast to the work of Schenkel et al. (2006) where an association between CAST and fat yield was identified, but the tender allele was associated with higher fat yield.

Marker Effects on Residual Variance

In comparison to the single residual variance model where σe = 1.79 kg for SSF (Fig. 2A and 3A), the CAST genotype specific residual variance model fit significantly (P < 0.001) better with σe = 1.22 kg, 1.99 kg, and 2.82 kg for CASTaT homozygotes, CASTaT:CASTaC heterozygotes, and CASTaC homozygotes, respectively (Fig. 2B). The CAST genotype specific residual model implies genotype specific heritabilities of h2 = 0.50, 0.27, and 0.16 for CASTaT homozygotes, CASTaT:CASTaC heterozygotes, and CASTaC homozygotes, respectively. In comparison, there was less support (P = 0.05) for the CAPN1 specific residual variance model with σe = 1.98 kg, 1.43 kg, and 2.21 kg for CAPN1hCC homozygotes, CAPN1hCC:CAPN1hGT heterozygotes, and CAPN1hGT homozygotes, respectively (Fig. 3B). The CAPN1 genotype specific residual model implies genotype specific heritabilities of h2 = 0.28, 0.43, and 0.24 for CAPN1hCC homozygotes, CAPN1hCC:CAPN1hGT heterozygotes, and CAPN1hGT homozygotes, respectively. It is quite appealing that for the CAST genotype specific residual variance models, the most tender animals are also the least variable group, providing 2 ways CASTaT may reduce the risk of a tough eating experience for consumers. In comparison, for the CAPN1 genotype specific residual variance model, the heterozygous genotype has the smallest variance for SSF. Estimated additive genetic effects of CAPN1 and CAST genetic markers under either single residual variance, CAPN1 genotype specific residual variance, or CAST genotype specific residual variance models are reported in Table 6. Additive genetic effects under the genotype specific residual variance models are similar to the effects under the single residual variance model.

Figure 2.
Figure 2.

Additive effects of CAST genotype on slice shear force under: A) single residual variance model, B) CAST genotype specific residual variance model.

 
Figure 3.
Figure 3.

Additive effects of CAPN1 haplotype on slice shear force under: A) single residual variance model, B) CAPN1 haplotype specific residual variance model.

 

View Full Table | Close Full ViewTable 6.

Estimated additive effects and SE for CAPN1 and CAST genetic markers on slice shear force with either single residual variance or genotype specific residual variance models

 
Type of residual variance model CAPN1
CAST
additive effect/GT haplotype additive effect/T allele
Single
Slice shear force, kg 1.049 ± 0.246 -1.257 ± 0.261
CAPN1 genotype specific
Slice shear force, kg 1.040 ± 0.279 -1.209 ± 0.251
CAST genotype specific
Slice shear force, kg 1.080 ± 0.224 -1.240 ± 0.341

Because log transformation is a common approach to analyze data where the variance increases with the mean, we conducted these analyses on log transformed SSF data as well. Estimated CAPN1 and CAST genotype effects were similar to the untransformed SSF analyses (Appendix I, Table 7). Our results of CAST genotype specific residual variance being a better model were further supported when the log transformed SSF models were compared. The CAST specific residual variance model (Appendix I, Fig. 4B) again fit significantly better (P = 0.03) than the single residual variance model (Appendix I, Fig. 4A and Fig. 5A). However, the CAPN1 specific residual variance model (Appendix I, Fig. 5B) did not fit significantly better (P = 0.10) than the single residual variance model. Genetic factors have been previously reported to affect the variability of carcass traits, and tenderness in particular. Crews and Franke (1998) reported better fit of models accounting for heterogeneous variance based on percentage Brahman inheritance for several carcass traits. Also, Freking et al. (1999) identified an effect of the callipyge gene on variation of shear force in lambs. However, to our knowledge this is the first study to evaluate the effect of CAPN1 or CAST genotypes on variance parameters for beef tenderness.


View Full Table | Close Full ViewTable 7.

Back transformed average additive effects and SE for CAPN1 and CAST genetic markers on log transformed slice shear force with either single residual variance or CAPN1 genotype specific residual variances or CAST genotype specific residual variances

 
Type of residual variance model CAPN1
CAST
additive effect/GT haplotype additive effect/T allele
Single
Slice shear force, kg 1.081 ± 0.242 -1.218 ± 0.255
CAPN1 genotype specific
Slice shear force, kg 1.068 ± 0.268 -1.181 ± 0.242
CAST genotype specific
Slice shear force, kg 1.105 ± 0.229 -1.204 ± 0.305
Figure 4.
Figure 4.

Back transformed additive effects of CAST genotype on log transformed slice shear force under: A) single residual variance model, B) CAST genotype specific variance model.

 
Figure 5.
Figure 5.

Back transformed additive effects of CAPN1 haplotype on log transformed slice shear force under: A) single residual variance model, B) CAPN1 haplotype specific variance model.

 

Struchalin et al. (2010) proposed that heterogeneity of variance based on each marker genotype may be an indicator of a SNP interaction with an unobserved factor (e.g., another genetic marker). They also propose that another plausible interpretation is that the SNP affects the variance without a specific interaction. Because CAST has been identified as an inhibitor for multiple calpain proteases, it may be possible that different genotypes of CAST have different affinity to inhibit calpains not modeled in this study and this may be the source of the effects on residual variance observed in this study.

A large amount of research has been applied to the CAPN1 and CAST system (Koohmaraie, 1996) and the effects of genetic markers in these genes on beef tenderness (Casas et al., 2006; Barendse et al., 2007; Van Eenennaam et al., 2007; Gill et al., 2009). This work suggests a reevaluation of the potential mechanism by which CAST influences tenderness via the increasing residual variance associated with genotypes having the increased mean tenderness values should be investigated in other resource populations.

Implications

A tough steak can be a very negative experience for a consumer. This study indicates that CAPN1 and CAST genetic markers both have large effects on beef tenderness and CAST also influences the variation as well as the mean tenderness of a group of steaks. Both differences in means and variances of CAST alleles and CAPN1 haplotypes can be modeled to estimate the probability of a tough steak. The larger variance associated with the tough CAST allele should increase the proportion of tough meat more than the tough CAPN1 haplotype which does not increase variance. This will be important for estimating the economic effect for the beef industry of including these tenderness markers in the beef cattle breeding objective. Because CAPN1 and CAST genetic markers have been investigated extensively, other research groups may choose to also investigate the genotype specific residual variance models in their populations.

APPENDIX I

Average additive genetic effects of CAPN1 and CAST markers on log transformed SSF are reported in Table 7. Figures 4 and 5 show back transformed distributions of log transformed SSF analysis results under single residual variance and CAST genotype specific variance (Fig. 4) or CAPN1 genotype specific variance (Fig. 5) models. Estimates of genetic marker effects on target and nontarget traits, whether significant or not, may be of interest to future researchers for comparison purposes. Table 8 reports the estimated additive and dominance effects for the CAPN1hGT haplotype relative to the CAPN1hCC haplotype on all traits evaluated in this study. Table 9 reports the estimated additive and dominance effects for the CASTaT allele relative to the CASTaC allele on all traits evaluated in this study. Table 10 reports the estimated additive and dominance effects for the GHRaT allele relative to the GHRaA allele on all traits evaluated in this study.


View Full Table | Close Full ViewTable 8.

Estimated effect of CAPN1 GT haplotype relative to CAPN1 CC haplotype for additive and dominance modes of inheritance

 
Trait Additive Dominance
Birth BW, kg -0.04 ± 0.29 -0.63 ± 0.37
Weaning BW, kg -0.27 ± 1.40 1.50 ± 1.84
Yearling BW, kg 2.49 ± 3.65 1.20 ± 4.87
Final BW, kg 5.57 ± 4.47 2.37 ± 5.96
Dressing percent, % -0.06 ± 0.15 -0.22 ± 0.21
Hot carcass weight, kg 3.00 ± 3.03 0.34 ± 4.05
Adjusted fat thickness, mm 0.31 ± 0.42 0.25 ± 0.56
Marbling score1 0.125 ± 0.084 -0.088 ± 0.112
LM area, sq. cm 0.41 ± 0.65 0.33 ± 0.85
Vision yield grade 0.041 ± 0.067 0.017 ± 0.089
Slice shear force, kg 1.049 ± 0.246 -0.432 ± 0.327
VISNIR2 predicted shear force, kg -0.205 ± 0.187 0.144 ± 0.250
14.00 = Slight00; 5.00 = Small00; 6.00 = Modest00 (USDA, 1997; BIF, 2010).
2Visible and near-infrared reflectance spectroscopy (VISNIR; Shackelford et al., 2012a,b).

View Full Table | Close Full ViewTable 9.

Estimated effect of CAST T allele relative to CAST C allele for additive and dominance modes of inheritance

 
Trait Additive Dominance
Birth BW, kg -0.10 ± 0.32 -0.55 ± 0.39
Weaning BW, kg -1.90 ± 1.58 2.26 ± 1.97
Yearling BW, kg -3.77 ± 3.88 3.31 ± 5.32
Final BW, kg −5.23 ± 4.75 -0.14 ± 6.51
Dressing percent, % -0.27 ± 0.16 -0.09 ± 0.22
Hot carcass weight, kg −4.70 ± 3.22 -0.99 ± 4.42
Adjusted fat thickness, mm -1.05 ± 0.44 -0.80 ± 0.61
Marbling score1 0.034 ± 0.089 0.000 ± 0.122
LM area, sq. cm 1.09 ± 0.68 -0.26 ± 0.93
Vision yield grade -0.198 ± 0.071 -0.073 ± 0.097
Slice shear force, kg -1.257 ± 0.261 -0.281 ± 0.357
VISNIR2 predicted shear force, kg 0.006 ± 0.199 0.473 ± 0.273
14.00 = Slight00; 5.00 = Small00; 6.00 = Modest00 (USDA, 1997; BIF, 2010).
2Visible and near-infrared reflectance spectroscopy (VISNIR; Shackelford et al., 2012a,b).

View Full Table | Close Full ViewTable 10.

Estimated effect of GHR T allele relative to GHR A allele for additive and dominance modes of inheritance

 
Trait Additive Dominance
Birth BW, kg -0.07 ± 0.29 -0.36 ± 0.36
Weaning BW, kg -0.20 ± 1.42 -1.77 ± 1.83
Yearling BW, kg 4.71 ± 3.46 -1.52 ± 4.84
Final BW, kg 7.34 ± 4.25 -3.60 ± 5.92
Dressing percent, % 0.08 ± 0.15 0.02 ± 0.20
Hot carcass weight, kg 4.90 ± 2.88 -2.08 ± 4.02
Adjusted fat thickness, mm 0.57 ± 0.40 -0.35 ± 0.55
Marbling score1 -0.089 ± 0.080 0.053 ± 0.111
LM area, sq. cm 0.39 ± 0.62 -0.23 ± 0.84
Vision yield grade 0.095 ± 0.064 -0.046 ± 0.088
Slice shear force, kg 0.254 ± 0.234 -0.086 ± 0.324
VISNIR2 predicted shear force, kg 0.029 ± 0.177 -0.395 ± 0.248
14.00 = Slight00; 5.00 = Small00; 6.00 = Modest00 (USDA, 1997; BIF, 2010).
2Visible and near-infrared reflectance spectroscopy (VISNIR; Shackelford et al., 2012a,b).
 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.