J. Anim Sci.
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stalder, K. J.
Right arrow Articles by Goodwin, R. N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stalder, K. J.
Right arrow Articles by Goodwin, R. N.
J. Anim. Sci. 2003. 81:611-616
© 2003 American Society of Animal Science

A comparison of hydrogen ion concentration and pH genetic predictions and fixed effect estimations when assessing pork quality1

K. J. Stalder*,2, A. M. Saxton*, R. K. Miller{dagger} and R. N. Goodwin{ddagger}

* Department of Animal Science, Tennessee Agricultural Experiment Station, University of Tennessee, Knoxville 37996; and {dagger} Department of Animal Science, Texas A&M University, College Station 77843; and the and {ddagger} National Pork Board, Clive, IA 50306

2 Correspondence:
West Tennessee Exp. Stn., 605 Airways Boulevard, Jackson, TN 38301, (phone: 731/425-4705; E-mail:
stalder{at}utk.edu).


   Abstract
Top
 Abstract
Introduction
 Experimental Procedures
 Results and Discussion
 Implications
 Literature Cited
 
An evaluation of ultimate pH (pH) and hydrogen ion concentration (H+) was conducted to determine if the mathematical conversion of H+ to pH could affect 1. fixed and random effect inferences and 2. prediction of genetic merit of animals when either pH or H+ is used as an indicator in the assessment of pork quality. Data from 4,262 purebred animals and 577 sires having complete three-generation pedigrees from the 1991 to 2001 National Barrow Show Progeny Tests were utilized in this study. Existing loin muscle pH data were converted to their original H+. Numerical changes in values occurred across all fixed effects and in the case of breed and test, changes in differences among subclasses occurred. These changes may result in differing inferences that can be made depending on whether pH or H+ is used as the dependent variable. Heritability estimates for pH and H+ were 0.52 ± 0.074 and 0.62 ± 0.078, respectively. The Pearson correlation between pH and H+ breeding values was -0.92. Spearman Rank correlation of -0.91 between pH and H+ breeding values was calculated and indicates that sires do not rank the same when ordered by breeding values for each trait. When pH is the selection objective, the selection differential reduction in H+ from these data ranges from 3.8 to 9.1%. Additionally, only 77.7% of the estimated genetic progress per generation in H+ is realized when selection (5% selected) is based on pH. The genetic correlation between pH and H+ was -0.96. Changes in the absolute values of the genetic correlations between various pork quality indicator traits and pH or H+ concentration were 0.04 or less. Differences in pH and H+ results could impact decisions made by swine breeders and meat processors who are concerned about pork quality. This, combined with the greater heritability and biochemical reality for H+, indicates that H+ rather than pH is the more appropriate trait breeders and processors should focus on when attempting to improve pork quality.

Key Words: Breeding Value • Concentration • Hydrogen Ions • pH


   Introduction
Top
 Abstract
 Introduction
Experimental Procedures
 Results and Discussion
 Implications
 Literature Cited
 
The pork industry is currently emphasizing quality in order to increase consumer acceptance and consumption of meat products. Meat scientists and geneticists have focused on various traits and their indicators in an attempt to improve the quality of pork that is produced on commercial operations. One of the indicator traits receiving much attention has been ultimate pH. The pork harvesting and processing industries are focused on identifying environmental factors that can improve pH and other quality traits so more of their pork products can be sold as premium products. The breeding stock industry is placing selection emphasis on pork quality traits and/or its indicators through index or direct breeding value selection.

When conducting genetic evaluations (Goodwin, 1994; Gibson et al., 1996) and when studying various factors associated with and/or affecting pork quality (Christian and Rothschild, 1981; Stalder et al., 1998), pH has been used as an indicator of pork muscle quality. Biologically, pH is defined as the negative log base 10 of the hydrogen ion concentration (pH = -log10 [H+]) of an item being measured (Zubay, 1988). Due to the nonlinearity of the log base 10 transformation, the mean of the log base 10 values (mean pH) will not mathematically equal the log base 10 of the mean hydrogen ion concentration (H+) values (Table 1Go). Few investigations have examined the statistical differences and the effects on the inferences (genetic parameters, prediction of breeding values, or estimation of fixed effects) made when H+ or pH is evaluated.


View this table:
[in this window]
[in a new window]
  		Table 1. An example of two sires with three progeny and each having identical pH averages, but differing hydrogen ion concentrations
 
The objectives of this study are to determine if the mathematical conversion of H+ to pH could affect 1) fixed and random effect inferences and 2) prediction of genetic merit of animals when either pH or H+ is used as an indicator in the assessment of pork quality.


   Experimental Procedures
Top
 Abstract
 Introduction
 Experimental Procedures
Results and Discussion
 Implications
 Literature Cited
 
The project utilized data from the National Barrow Show Progeny Test from 1991 to 2001. Data from 4,262 animals (Table 2Go) from 577 sires were utilized in this study. In addition, the eight breeds, 217 herds, 87 off-test dates within 11 test groups, two sexes and two Hal 1843 genotypes were included as fixed effects in all models. Complete three-generation pedigrees, fixed and random classifications used for analyses of various traits, and muscle quality information evaluated on pork carcass longissimus (pH, objective color, tenderness evaluation, and lipid content) were obtained from the National Pork Board.


View this table:
[in this window]
[in a new window]
  		Table 2. Distribution of animals by breed and test group from the 1991 to 2001 National Barrow Show Progeny Tests
 
Carcass data collection procedures followed those outlined in Stalder et al. (1998). A three-rib sample (10th through12th) of the longissimus muscle was removed from one side of the carcass. All ultimate pH measures were obtained from a 10th-rib section. The muscle sample was homogenized until finely ground using a food processor and divided into two portions. Ultimate pH (referred to hence forward as pH) was measured by inserting a surface probe electrode (Accumet pH/Ion Meter, Fisher Scientific, Pittsburgh, PA) directly into one of the ground sample portions. Total lipid content was determined in the remaining portion of the ground 10th-rib sample following procedures outlined by Bligh and Dyer (1959). Objective measures of muscle color were made using a Minolta Chromameter CR-310 (Minolta Corp., Ramsey, NJ) on the 10th-rib samples prior to their homogenization. Two measures of muscle color, Minolta L* (an index that is a lightness variable and more closely represents human sensitivity to color), and Hunter Color (an additional lightness variable, where higher values indicate more pale colors or lighter color) were collected. The 12th-rib section was broiled and evaluated for mechanical tenderness using a star probe (Oltrogge-Hammernick and Prusa, 1987) attached to an Instron Universal Testing Machine Model 4502 (Instron Corp., Canton, MA).

Existing muscle pH data was converted to its original H+ using 15-decimal place accuracy (SAS Inst. Inc., Cary, NC). Fixed-model, statistical analyses of the two dependent variables, pH and H+ (multiplied by 107 for numerical stability in analyses), were conducted implementing identical models using PROC MIXED of SAS (SAS Inst. Inc.). The model utilized in the analyses of each trait included the fixed effects of breed, test group, off-test date within test group, gender, and Hal 1843 genotype. The notation for the model used in this analysis is as follows:


where

Yijklmno =the pH or H+ concentration of oth individual with the nth Hal 1843 genotype of mth sex from lth off-test date within the kth test group from jth herd and from the ith breed,

µ=overall mean,

Bi=fixed effect common to the ith breed,

Hj=fixed effect common to the jth herd,

Tk=fixed effect common to the kth test group,

ODl(k)=fixed effect common to the lth off-test date within kth test group,

Sm=fixed effect common to the mth sex,

Pn=fixed effect common to the nth Hal 1843 genotype, and

eijklmno=random residual error.

Fixed effect differences and associated P value changes were evaluated for biological interpretation and accuracy to complete objective 1 of the study.

To accomplish objective 2, breeding values, heritabilities, and genetic gains were individually estimated for pH and H+ using ASREML software (Gilmour et al., 2001). Genetic correlations among pork quality traits were estimated in groups of three. All genetic correlation calculations included pH, H+, and an additional objectively measured pork quality trait.

An animal model that incorporated the full relationship matrix of individuals and the fixed effects from objective one was implemented. Heritability was calculated by multiplying the sire variance by four and dividing this value by total phenotypic variance. Sire pH and H+ breeding values were used to examine theoretical genetic gains and differences when selection is based on either trait. The notation of the model used in this analysis is as follows:


where

y =vector of pH and H+ concentration phenotypic values,

X1,2,3,4,5, and 6=incidence matrices for fixed effects related to the phenotypic values,

b1=unknown vector of a fixed breed effect associated with the record in y,

b2=unknown vector of a fixed herd effect associated with the record in y,

b3=unknown vector of a fixed test group effect associated with the record in y,

b4=unknown vector of an off-test date within a test group fixed effect associated with the record in y,

b5=unknown vector of a fixed sex effect associated with the record in y,

b6=unknown vector of a fixed Hal 1843 effect associated with the record in y,

Z=incidence matrix for random effects related to the phenotypic performance values,

u=unknown vector of random animal effects associated with the records in y, and

e=vector of random residual effects.

Spearman rank correlations were calculated to evaluate breeding value estimates obtained for two traits. Relative differences among estimated breeding values and the ranking order of sires were evaluated.


   Results and Discussion
Top
 Abstract
 Introduction
 Experimental Procedures
 Results and Discussion
Implications
 Literature Cited
 
Measurement of pH in pork carcasses has occurred at various times after harvesting has taken place. Ultimate pH has generally been measured 24 h after harvesting and after the carcass has been thoroughly chilled (National Pork Board, 2000). Ultimate pH has been a trait of interest because of its relationship to pork muscle water-holding capacity (Stalder et al., 1998). Improved water-holding capacity has the potential to increase the pork value because of increased carcass, primal cut, and retail product weights.

The need to investigate the effect of pH and H+ on fixed estimates and random predictions arises because of the mathematical properties of log functions. In addition, H+ may be the preferred value for genetic improvement because it is the biologically active component. That is, living cells have H+ concentrations, not pH values. Hence, the true value swine breeders are attempting to improve through selection is hydrogen ion concentration and not pH.

Originally, the pH transformation was developed to modify the extremely small values obtained when measuring H+ as shown in Table 1Go. The transformation was not performed to distribute the data normally for statistical analysis. The small H+ values could be corrected just as easily by a linear transformation, for example multiplying by 107, as was done in the present study. Another argument has been that the log transform provides better normality for statistical analysis. Plots of PROC MIXED (SAS Inst. Inc.) residuals for pH and H+ (Figure 1Go) both showed slight negative skew, but either is acceptable for statistical analysis.



View larger version (26K):
[in this window]
[in a new window]
  		Figure 1. Residual distribution of pork carcass longissimus pH and hydrogen measures from the 1991 to 2001 National Barrow Show progeny tests.

 
The results for the fixed effect model for breed, test, sex, and Hal 1843 genotype are shown in Table 3Go. All model effects were significant (P < 0.03 or smaller) sources of variation in the analysis of pH and H+ concentration. The hydrogen ion concentration has been converted back to pH in order to make direct comparisons between the two methods of analysis. Numerical changes in values occurred across all effects and in the case of breed and test, changes in differences among subclasses occurred. These changes result in differing inferences that can be made depending on whether pH or H+ concentration is used in the analysis. Changes in fixed effect results from pH and H+ analyses could impact decisions that breeders and pork processors make when examining various environmental factors (day, plant, etc.) affecting pork quality variation (Kauffman et al., 1992).


View this table:
[in this window]
[in a new window]
  		Table 3. Pork carcass longissimus hydrogen ion concentration and pH least squares means for breed, test, sex, and Hal 1843 from the 1991 to 2001 National Barrow Show Progeny Tests
 
The relative gilt-barrow and Hal 1843 NN-Nn differences are the same in the analysis of both pH and H+ (Table 3Go). Resulting differences in fixed effect results depending on whether pH or H+ is utilized are similar to those reported by Murphy (1982) but differ from the results previously reported by Boutilier and Shelton (1980).

Greater precision of fixed effect estimation is an important issue when predicting breeding values. Solutions for the H+ fixed effects generally had lower standard errors of the estimates when compared to the same pH values (Table 3Go). When fixed effects are more precisely estimated, the effective contemporary group size is larger; hence, improved accuracy and greater genetic progress can be expected (Rothschild et al., 1987; Lofgren and Stewart, 1994).

The heritability estimates for pH and H+ concentration were 0.52 ± 0.074 and 0.62 ± 0.078, respectively (Table 4Go). Both estimates would be considered highly heritable; however, greater genetic progress would be expected if selection were based upon H+ concentration rather than its transformed pH value when attempting to improve pork quality. The Pearson correlation coefficient between pH and H+ breeding values concentration was -0.92. The highly negative correlation between the two traits was expected, as pH is a transformed value of H+.


View this table:
[in this window]
[in a new window]
  		Table 4. Heritability estimates, genetic gain estimations, and breeding value correlations of pork carcass longissimus pH and hydrogen ion concentrations from the 1991 to 2001 National Barrow Show Progeny Tests
 
To better understand the impact of selection based upon pH or H+, Spearman Rank coefficient of correlation was calculated. While still relatively high, the -0.91 rank correlation between pH and H+ breeding values indicates that some differences in ranking do occur when comparing the breeding values obtained for pH and H+. Selection programs attempting to improve pork quality could be impacted by selecting on pH or H+, as shown in an example of truncation selection on the two traits in Figure 2Go. In this example, the observed loss in H+ selection differential by selecting on pH instead of H+ ranges from 3.8% to 9% depending upon selection rate. This is a consequence of not selecting individuals in the lower left quadrant of the graph, instead choosing those in the upper right. The graph shows the 1%, 5%, and 25% selected levels. In addition, Table 5Go shows the loss in H+ in selection differential when selection is based on pH instead of H+ at the top 1, 5, and 25% truncation levels. The observed selection differential loss increases as percentage selected decreases from 25 to 5% from the data evaluated in this study (Table 5Go). The observed selection differential loss decreased to 3.8% when the top 1% of animals is selected (only six sires would be selected, 1% of 577 sires). This decrease does not reflect what could happen when the population is larger or fewer progeny per sire are evaluated.



View larger version (31K):
[in this window]
[in a new window]
  		Figure 2. Example truncation selection of sires for pH and hydrogen ion concentration based on top 1, 5, and 25% breeding values from the 1991 to 2001 National Barrow Show progeny tests. The lines .........., __________, and –––––––––– represent 1%, 5%, 25% selection intensities.

 

View this table:
[in this window]
[in a new window]
  		Table 5. Percentage selection differential loss when sires are selected on breeding values for pH or hydrogen ion concentration based on top 1, 5, and 25% truncation selection from the 1991 to 2001 National Barrow Show progeny testsa
 
To gain an understanding of the impact of selection on pH or H+ concentration on various other pork quality indicator traits, genetic gain (Table 4Go) and genetic correlations (Table 6Go) were estimated. It is expected that only 77.7% (0.1880/0.2330) of the genetic progress per generation would occur in improving H+ if selection (5% selected) were based on pH rather than H+ (Table 4Go).


View this table:
[in this window]
[in a new window]
  		Table 6. Genetic correlations (±SE) between hydrogen ion concentration, pH, and indicators of pork quality measured on pork carcass longissimus from the 1991 to 2001 National Barrow Show Progeny Tests
 
The genetic correlation between pH and H+ concentration was -0.96. A high value was expected because the traits are identical, only pH is a transformed value of the composite genetic effects on H+ concentration. Changes in the absolute values of the genetic correlations between various pork quality indicator traits and pH or H+ concentration were 0.04 or less (Table 5Go). These indicate that selection on pH or H+ concentration would yield similar correlated responses in other pork quality traits.

Since pH has become a common trait to the pork industry, a plausible solution to the illustrated problem may be to perform statistical analyses using the H+ values and make the transformation to pH after all evaluations have taken place as has been done in Table 3Go. In this manner, the industry can continue to utilize pH as an indicator of pork quality.


   Implications
Top
 Abstract
 Introduction
 Experimental Procedures
 Results and Discussion
 Implications
Literature Cited
 
When statistical analyses of pH and H+ of pork or other substances are conducted using identical models, changes in ranking of fixed effect subclasses can occur. These changes can occur when small numbers of observations and/or unequal variances exist within fixed or random subclasses. Small observation numbers could be problematic when young sires are evaluated. This potentially could alter interpretations of various effects depending on whether pH or H+ as the unit of measure of pork quality. If selection is based on pH breeding values, the ranking differences could potentially result in inaccurate sire selection in order to improve H+, the biologically active trait. These types of changes reduce the H+ genetic progress that a seedstock producer could make to improve pork quality. As indicated by the estimated genetic correlations, similar rates of genetic improvement in other pork quality indicator traits could be expected when selection is based on H+ rather than pH.


   Footnotes
 
1 This project was supported with funds from the Pork Checkoff of the National Pork Board. Back

Received for publication April 25, 2002. Accepted for publication September 20, 2002.


   Literature Cited
Top
 Abstract
 Introduction
 Experimental Procedures
 Results and Discussion
 Implications
 Literature Cited
 


Bligh, E. G., and W. J. Dyer. 1959. A rapid method for total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911–917.

Boutilier, R. G., and G. Shelton. 1980. The statistical treatment of hydrogen ion concentration and pH. J. Exp. Biol. 84:335–339.[Free Full Text]

Christian, L. L., and M. F. Rothschild. 1981. Performance and carcass characteristics of normal, stress-carrier, and stress-susceptible swine. Publication AS-528-F. Iowa State University Cooperative Extension Service, Ames.

Gibson, J. P., C. A. Aker, and R. O. Ball. 1996. Proceedings of the Ontario Pork Carcass Appraisal Project Symposium. Ontario Swine Improvement Inc. Guelph, Ontario, Canada.

Gilmour, A. R., B. R. Cullis, S. J. Welharn, and R. Thompson. 2001. ASREML Reference Manual. NSW Agriculture, Orange, Australia.

Goodwin, R. N. 1994. Genetic parameters of pork quality traits. Ph.D. Thesis. Iowa State University, Ames.

Kauffman, R. G., R. G. Cassens, A. Scherer, and D. L. Meeker. 1992. Variations in Pork Quality. National Pork Producers Council publication. Des Moines, IA.

Lofgren, D. L., and T. S. Stewart. 1994. Optimal contemporary group structure to maximize genetic progress through genetic evaluation of swine. J. Anim. Sci. 72:2254–2259.[Abstract]

Murphy, M. R., 1982. Analyzing and presenting pH data. J. Dairy Sci. 65:161–163.[Abstract/Free Full Text]

National Pork Board. 2000. Pork Composition and Quality Assessment Procedures. National Pork Board, Des Moines, IA.

Oltrogge-Hammernick, M., and K. J. Prusa. 1987. Research Note: Sensor analysis and Instron measurements of variable-power microwave-heated baking hen breasts. Poult. Sci. 66:1548–1551.

Rothschild, M. F., M. E. Einstein, D. L. Lofgren, and T. S. Stewart. 1987. Accurate estimation of breeding values in swine. How big should contemporary groups be? Pages 40–43 Proc. Nat. Swine Imp. Fed. Conf. and Ann. Mtg., St. Louis, MO.

Stalder, K. J., J. Maya, L. L. Christian, S. J. Moeller, and K. J. Prusa. 1998. Effects of preslaughter management on the quality of carcasses from porcine stress syndrome heterozygous market hogs. J. Anim. Sci. 76:2435–2443.[Abstract/Free Full Text]

Zubay, G. 1988. Biochemistry 2nd ed. MacMillian Publishing Company, New York.


This article has been cited by other articles:


Home page
 J ANIM SCIHome page
D. W. Newcom, T. J. Baas, K. J. Stalder, and C. R. Schwab
Comparison of three models to estimate breeding values for percentage of loin intramuscular fat in Duroc swine
J Anim Sci, April 1, 2005; 83(4): 750 - 756.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stalder, K. J.
Right arrow Articles by Goodwin, R. N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stalder, K. J.
Right arrow Articles by Goodwin, R. N.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS