Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Genetics

Genetic variation in efficiency to deposit fat and lean meat in Norwegian Landrace and Duroc pigs1

 

This article in JAS

  1. Vol. 93 No. 8, p. 3794-3800
     
    Received: Apr 07, 2015
    Accepted: June 03, 2015
    Published: July 24, 2015


    2 Corresponding author(s): kristine.martinsen@nmbu.no
 View
 Download
 Share

doi:10.2527/jas.2015-9174
  1. K. H. Martinsen 2*,
  2. J. Ødegård,
  3. D. Olsen and
  4. T. H. E. Meuwissen*
  1. * Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences, P.O. Box 5003, NO-1432 Ås, Norway
     AquaGen AS, P.O. Box 1240 Sluppen, NO-7462 Trondheim, Norway
     Topigs Norsvin, P.O. Box 504, NO-2304 Hamar, Norway

Abstract

Feed costs amount to approximately 70% of the total costs in pork production, and feed efficiency is, therefore, an important trait for improving pork production efficiency. Production efficiency is generally improved by selection for high lean growth rate, reduced backfat, and low feed intake. These traits have given an effective slaughter pig but may cause problems in piglet production due to sows with limited body reserves. The aim of the present study was to develop a measure for feed efficiency that expressed the feed requirements per 1 kg deposited lean meat and fat, which is not improved by depositing less fat. Norwegian Landrace (n = 8,161) and Duroc (n = 7,202) boars from Topigs Norsvin’s testing station were computed tomography scanned to determine their deposition of lean meat and fat. The trait was analyzed in a univariate animal model, where total feed intake in the test period was the dependent variable and fat and lean meat were included as random regression cofactors. These cofactors were measures for fat and lean meat efficiencies of individual boars. Estimation of fraction of total genetic variance due to lean meat or fat efficiency was calculated by the ratio between the genetic variance of the random regression cofactor and the total genetic variance in total feed intake during the test period. Genetic variance components suggested there was significant genetic variance among Norwegian Landrace and Duroc boars in efficiency for deposition of lean meat (0.23 ± 0.04 and 0.38 ± 0.06) and fat (0.26 ± 0.03 and 0.17 ± 0.03) during the test period. The fraction of the total genetic variance in feed intake explained by lean meat deposition was 12% for Norwegian Landrace and 15% for Duroc. Genetic fractions explained by fat deposition were 20% for Norwegian Landrace and 10% for Duroc. The results suggested a significant part of the total genetic variance in feed intake in the test period was explained by fat and lean meat efficiency. These new efficiency measures may give the breeders opportunities to select for animals with a genetic potential to deposit lean meat efficiently and at low feed costs in slaughter pigs rather than selecting for reduced the feed intake and backfat.



INTRODUCTION

Production efficiency is of importance in livestock production because of greater competition for feed resources due to growth in human food consumption and an increasing scarcity of feeds due to climate change (Åby et al., 2014). In addition, Shirali et al. (2012) showed that selection for feed efficiency could reduce total protein excretion, which is the greatest pollution factor in pig production. These factors make improvement of total pork production efficiency an important goal for future pig breeding to meet the likely prospective challenges. The profitability of pork production is dependent on feed requirements, as feed costs are the greatest costs in pork production (Niemi et al., 2010). Based on current market economy, feed efficiency is a trait of importance for genetic improvement of production efficiency (Kanis et al., 2005). Bernard and Fahmy (1970) proved that selection for changed body composition, such as carcass leanness and reduced backfat, lead to indirect selection for feed-efficient animals. Since then, this has been a common way to select for improved feed efficiency in commercial breeding companies (Patience, 2012). Other approaches as gross feed intake, residual feed intake, feed conversion ratio, and feed intake relative to growth rate in the breeding goal are also used (Korver, 1988; Gilbert et al., 2007; Do et al., 2013). These traits have moderate heritabilities, and selection for these traits has resulted in more cost-effective production (Suzuki et al., 2005; Chen et al., 2010; Saintilan et al., 2013). However, based on selection responses for these traits, a hypothesis might be that the traditional traits are related to allocation of nutritional resources to lean meat and fat growth rather than how efficient the animal converts feed into product (Rauw et al., 1998; Cai et al., 2008). Moreover, selection for reduced feed intake and reduced body reserves may cause problems for lactating sows that raise large litters (Eissen et al., 2000). The objective was, therefore, to develop a novel measure of feed efficiency expressed as feed consumed (kg) per 1 kg lean meat or fat deposited and to test whether genetic variation in efficiency to deposit fat and lean meat existed within Norwegian Landrace and Duroc pig populations.


MATERIALS AND METHODS

Data and Trait Recording

Data were provided by the Topigs Norsvin company in Norway and recorded on Norwegian Landrace and Duroc boars born from 2008 to 2014 at their boar testing station (“Delta”). Annually, about 3,500 boars from the Topigs Norsvin’s nucleus herds in Norway are tested, equally divided between the 2 breeds. Boars are housed in pens of 12. In each pen, there is 1 feed station (FIRE; Osborne Industries Inc., Osborne, KS), where individual amount of feed consumed at each visit, number of visits, and time spent per visit in the feed station is recorded. In addition, individual BW is recorded as the median of all weights registered at visits to the feeding station that day. Boars are fed ad libitum on conventional concentrate containing 194 and 164 g digestible protein and 9.79 and 9.61 MJ NE/kg before and after 65 kg live weight. Boars enter the test at approximately 40 kg live weight, with an average age of 90 d for Duroc and 85 d for Landrace, respectively. As a standard, the test is terminated and computed tomography (CT) scans performed when boars reach approximately 120 kg (approximately 100 kg before March 2012) live weight. Boars are sedated during CT scanning and do not eat on the scanning day; that is, feed intake recording is terminated the day before scanning. Through image analysis of the scans, CT provides information directly on the selection candidate boars for the traits lean meat (kg) and fat (kg) on the carcass. In total, 8,161 Norwegian Landrace and 7,202 Duroc boars had information on total feed intake (over the test period), lean meat (kg), and fat (kg) in the data set. Pedigree information for the boars was traced back 11 generations and included 18,843 and 13,901 animals for Norwegian Landrace and Duroc, respectively.

Estimation of Total Feed Intake in Test Period

The trait analyzed was total feed intake during the test period (FI). For both breeds, the trait was a summation of the feed intake from different stages of the growth curve. These stages were 40 to 60, 60 to 80, and 80 to 100 kg live weight. In addition, for boars that entered the test after March 1, 2012, the feed intake from 100 to 120 kg live weight was also included in the summation. Descriptive statistics for the data sets are shown in Table 1, and each boar had 1 record for FI, 1 for lean meat content, and 1 for fat content. Slaughter percentage was similar for the 2 breeds, but the Duroc had a generally greater average and variation in feed intake during the test period compared with the Norwegian Landrace. In addition, the Duroc had a lower average amount of lean meat and a higher fat content in the carcass compared with Norwegian Landrace.


View Full Table | Close Full ViewTable 1.

Number of boars, average, SD, minimum, and maximum values for total feed intake during the test period (FI), lean meat, fat, live weight (LW) and slaughter percentage (SP) for Norwegian Landrace (NL) and Duroc (D)

 
No. of boars
Average
SD
Minimum
Maximum
Breed
Parameter NL D NL D NL D NL D NL D
FI, kg 8,161 7,202 152.0 157.0 29.4 29.9 97.2 80.1 270.6 258.2
Lean meat, kg 8,161 7,202 52.3 48.6 3.6 3.7 40.5 35.8 68.2 64.8
Fat, kg 8,161 7,202 15.9 19.4 4.3 4.5 7.2 8.0 33.0 35.6
LW, kg 8,161 7,202 111.8 111.5 11.8 11.8 93.9 93.4 140.7 149.2
SP, % 8,161 7,202 69.2 69.9 2.1 1.7 56.3 48.0 82.9 86.7

Statistical Analysis

Records more than 4 SD from the mean within breed were discarded as outliers. Boars with missing values for at least 1 of the subtraits (i.e., FI from 40 to 60 or from 80 to 100 kg) were deleted from the data sets and all records were standardized as a deviation from the mean within breed.

The data were analyzed in a univariate animal model, and estimation of variance and covariance components was performed using the DMU software package (Madsen and Jensen, 2013). Lean meat and fat in carcass were included as both fixed and random regression cofactors in the model. In addition, the analysis included each boars’ maintenance requirement in the model as a fixed regression cofactor. The maintenance requirement was estimated as the integrated metabolic growth curve for each boar, with the assumption that the metabolic BW (MBW) was proportional to the BW raised to 0.75. The function integrated was MBW = (μ + bx)0.75, in which b was the linear regression coefficient of MBW, x was the age of the boar, and μ was the overall mean. The linear regression was used to estimate accumulated metabolic BW (AMW) for each boar. The lower limit (z) was age at 40 kg and upper limit (w) was age at 100 or120 kg:

The efficiency traits were expressed as the genetic regression coefficients that represented the extra feed needed to increase lean meat and fat deposition with 1 kg. This method was based on nutritional models with fixed regression earlier addressed by, for example, van Milgen and Noblet (1999) and also by Aggrey and Rekaya (2013), which used a random regression model for calculating residual feed intake (RFI) for maintenance and RFI for growth in broiler chicken.

The following model [1] was fitted separately for both breeds:

In the model, Yijknoq was total feed intake from 40 to 100 or 120 kg of live weight (kg), depending on when CT scanning occurred. Fixed effects included were herd–year (HY), birth month (BM), scanning time (ST), and section (SEC). In the model, ao and penq were the random effects of the breeding value of the boar and the pen they were housed in. Pen was included as a random effect because of small numbers of animals in each pen.

The regressions βlm × LMEATo and βfat × FATo were the fixed regression on lean meat (kg) and fat (kg), respectively. Lean meat and fat was estimated by the CT. The regression βamw × AMWo was the fixed regression on AMW. Random regressions were also included and × lmeato was the random regression on lean meat (kg), in which was the measure for the feed efficiency to deposit lean meat and represented the amount of feed used to produce 1 kg lean meat (lean meat efficiency of boar o).

The regression × fato expressed as was the random regression of fat (kg) for boar o, in which was a measure for the feed efficiency to deposit fat and represented the amount of feed used to produce 1 kg fat (fat efficiency of boar o). The residual variance in the model was eijknoq for boar o.

In the model, the animal intercept (ao) explained the variation in FI caused by other factors, such as the part of activity not related to the animals’ size (AMW). These factors could be the maintenance requirement part that is not explained by the MBW (e.g., the animal’s activity, heat production, disease status). In general, the effect includes all genetic variation in feed intake caused by the animal that is not explained by the animals’ MBW, deposition of lean meat and fat, or other effects included in the model.

After variance component estimation, the fraction of total genetic variance in FI due to lean meat and fat efficiency was defined as , in which was the average over all boars’ squared amounts of lean meat (kg), denoted by , (k = p) or fat (kg),, (k = f), with the corresponding variance, , estimated by model [1]. The variance () represented the variation in the regression coefficient for lean meat or fat. Estimation of total genetic variance in FI () was an average over all boars’ amounts of fat (Xf) and lean meat (Xp) and was estimated using the following formula:in which denotes average over all Xp and Xf. In the formula, was the genetic variation in FI that could not be explained by the other factors included in the model. To investigate the importance of lean meat and fat efficiency, variance components were also estimated with a simpler animal model [2], analyzing residual feed intake:

Model [2] included the same effects as model [1] but excluded the random effects of lean meat and fat deposition.


RESULTS

Fixed Effects

Table 2 includes the fixed regression coefficients for lean meat efficiency, fat efficiency, and AMW. There was no effect of lean meat deposition on total feed intake for any of the breeds, whereas fixed regression coefficients for fat efficiency and AMW were different from zero (Table 2). For Norwegian Landrace, the fixed regression coefficient for fat efficiency indicated that a boar, on average, used 2.24 ± 0.06 kg extra feed/kg fat growth. Duroc, on the other hand, needed slightly more additional feed (2.49 ± 0.07 kg feed/kg fat growth). The regression coefficient for AMW reflected the average amount of feed needed for maintenance per kilogram MBW; the estimates were both different from zero but lower for Duroc than Norwegian Landrace. This suggests that the Norwegian Landrace had a greater average maintenance requirement than Duroc per kilogram MBW (Table 2).


View Full Table | Close Full ViewTable 2.

Fixed regression coefficients (SE) for lean meat (βlm), fat (βfat), and accumulated metabolic BW (βamw) for Norwegian Landrace and Duroc

 
Regression coefficient Norwegian Landrace Duroc
βlm –0.027 (0.06) 0.073 (0.10)
βfat 2.241 (0.06) 2.495 (0.07)
βamw 0.050 (0.00) 0.046 (0.00)

Genetic Variance Components and Genetic Correlations

Genetic variance and covariance components (SE) estimated with model [1] for the effect of animal, lean meat efficiency, and fat efficiency are shown in Tables 3 and 4 for Norwegian Landrace and Duroc, respectively. All variance components for both breeds were greater than zero (Tables 3 and 4). For Norwegian Landrace, the genetic variation in fat efficiency was greater than for lean meat efficiency, whereas the opposite was true for the Duroc. Genetic variance components calculated with models [1] and [2] are shown in Table 5. Genetic variation was greater for both breeds when model [1] was used, whereas residual variation was lower.


View Full Table | Close Full ViewTable 3.

Variance components (SE) for the intercept of total feed intake during the test period (a), regression coefficients for lean meat in kilograms (ap) and fat in kilograms (af) for Norwegian Landrace on the diagonal and genetic correlations (SE) among a, ap, and af for Norwegian Landrace on the off-diagonal

 
a ap af
a 17.38 (1.42) 0.24 (0.06) 0.72 (0.05)
ap 0.23 (0.04) –0.17 (0.11)
af 0.26 (0.03)

View Full Table | Close Full ViewTable 4.

Variance components (SE) for the intercept of total feed intake during the test period (a), regression coefficients for lean meat in kilograms (ap) and fat in kilograms (af) for Duroc on the diagonal and genetic correlations (SE) among a, ap, and af for Duroc on the off-diagonal

 
a ap af
a 26.04 (2.15) 0.44 (0.05) 0.58 (0.07)
ap 0.38 (0.06) –0.24 (0.14)
af 0.17 (0.03)

View Full Table | Close Full ViewTable 5.

Variance components (SE) for the animal, pen, and residual for Norwegian Landrace and Duroc based on models [1] and [2]

 
Breed Model [1]
Model [2]
Norwegian Landrace Duroc Norwegian Landrace Duroc
Animal 24.93 34.19 13.69 (1.49) 21.83 (2.14)
Pen 5.66 (0.52) 5.41 (0.61) 6.49 (0.58) 5.74 (0.64)
Residual 17.83 (0.93) 23.69 (1.36) 26.18 (1.14) 33.12 (1.56)

The correlation between the random regression coefficients for fat and lean meat was close to zero and nonsignificant for both breeds. The genetic correlation between animal intercept for FI and fat and lean meat efficiencies were, respectively, 0.72 and 0.24 for the Norwegian Landrace and 0.58 and 0.44 for the Duroc. This indicates that those animals with a low feed intake are also likely to have lower feed requirements per unit fat deposited and are thus more fat/lean meat efficient.

Fraction of Total Genetic Variance

Table 6 summarizes that genetic variation in the lean meat and fat efficiencies contribute substantially to the total genetic variance in FI. In Norwegian Landrace, fat efficiency was more important than lean meat efficiency with respect to genetic variation in FI (20 and 12%, respectively), whereas the opposite was the case for Duroc (10% for fat efficiency and 15% for lean meat efficiency).


View Full Table | Close Full ViewTable 6.

Fraction of total genetic variation due to lean meat and fat efficiency in Norwegian Landrace and Duroc

 
Parameter Norwegian Landrace Duroc
Fraction of total genetic variance due to lean meat efficiency 0.12 0.15
Fraction of total genetic variance due to fat efficiency 0.20 0.10


DISCUSSION

Although approaches to improve feed efficiency through recording of feed intake, reduced backfat, increased carcass leanness, and daily gain exists, it is not obvious that these selection efforts result in pigs with more efficient fat and lean meat deposition. The increased feed efficiency may be due to nutrient resources increasingly being allocated to fat and lean meat growth and less to other processes (e.g., disease resistance). At some point, this reallocation reaches a biological limit and it would be necessary to breed for actual efficiency of fat and lean meat deposition instead of reallocation of resources. The current research investigated whether this was possible and 1) developed statistical methodology to perform the breeding value estimation and 2) found that there were genetic differences between pigs in efficiency of fat and lean meat deposition. Selection for growth rate remains important, next to a selection for lean meat efficiency, because it reduces the costs of housing of the animals and the maintenance requirements. If further reduction in backfat is not desired, as Norwegian Landrace are very lean (Gjerlaug-Enger et al., 2012), selection for fat efficiency may replace the current selection against backfat.

In practice, selection against feed intake is accompanied by selection for (lean meat) growth and is closely related to residual feed intake (Kennedy et al., 1993). In terms of the present study, Kennedy’s residual feed intake is similar to the residual feed intake modeled by model [2]. Model [1] in the current study extends model [2] residual feed intake model by splitting am into components that are due to the actual efficiency of deposition of fat and lean meat (af and ap). With model [1], it is thus possible to select for actual fat and lean meat efficiency without affecting the reallocation of feed resources. Therefore, the model for analyzing FI in the routine genetic evaluation for boars would be superior if fat and lean meat efficiency were included and it would be useful to get a better understanding of the components underlying overall feed efficiency.

Fixed Regression Coefficients

The fixed regression coefficients of lean meat deposition were not different from zero for Landrace and Duroc (Table 2). This could partly be due to the negative correlation between backfat and lean meat content on the carcass, which makes the lean meat regression coefficient difficult to estimate (Lo et al., 1992). Furthermore, the pigs were CT scanned at approximately the same live weight, which means that a pig with a high lean meat content typically has reduced noncarcass body mass (differences in fat deposition are corrected for in the model). Hence, these pigs may have reduced feed intake due to the lower costs of depositing noncarcass body mass. This suggests that the lean meat regression coefficient reflected the costs of depositing lean meat subtracted from the costs of depositing noncarcass body parts. The results in Table 2 implied that this difference was not different from zero. The same argument also holds for the regression on fat deposition, but the difference was positive due to the great costs of depositing fat compared with lean meat and noncarcass body parts.

Genetic Variance Components and Genetic Correlations

The fractions of total genetic variance due to fat and lean meat efficiency and fixed regression coefficients differed between breeds. Based on the present study, individual differences between the boar’s efficiency to deposit lean meat and fat and differences between breeds existed. A high regression coefficient for lean meat (ap) implies that a boar is expected to consume a large amount of feed to produce 1 kg lean meat and is inefficient. In Norwegian Landrace, a smaller fraction of the total genetic variance in FI was explained by lean meat than fat efficiency, but the opposite was true for Duroc. These breed differences may be caused by different breeding goals and different selection strategies in the past. The Norwegian Landrace has been selected for lower feed intake, increased lean meat percentage, and lower backfat thickness for many years (Kolstad, 2000). In Duroc, the selection has been more focused toward carcass and meat quality traits such as intramuscular fat. The fraction of genetic variance due to lean meat and fat efficiency were small, suggesting that the genetic variation in total feed intake at the test station was also due to other genetic factors.

Aggrey and Rekaya (2013) reported variance components for maintenance efficiency and growth efficiency in chickens estimated with the same method as this study. Our results could not be directly compared to these due to different efficiency measures, but their study proves that genetic variation in efficiency for growth exists between animals and is supporting our results. Sizeable estimates of genetic variation have been reported for lean meat and fat on the carcass and FI and in maintenance requirements (Cameron, 1990; Hermesch et al., 2000; Gjerlaug-Enger et al., 2012). The abovementioned components affect the new efficiency traits, fat and lean meat efficiency, and therefore, genetic variation was expected to exist in these new traits.

Genetic correlations between the animal intercept and the random regression coefficients were significantly different from zero. The results indicated that animals with low feed intake (intercept) were more efficient in deposition of fat compared with animals with a high feed intake (Tables 3 and 4). In agreement with this, Barea et al. (2010) found that a pig line selected for high RFI was energetically less efficient due to greater basal metabolism and higher physical activity, whereas there was no significant line effect on N retention (i.e., lean meat growth).

Statistical Analysis

When comparing model [1] with model [2] in Table 5, the variance components for residual and pen were greater with model [2], suggesting that the model [1] explained more of the heritable variation in FI, that is, due to the explicit modeling of fat and lean meat efficiency (Tables 3, 4, and 5). It also attempted to fit the individual boar’s maintenance requirement as a random regression coefficient in the model. However, these analyses did not converge (in DMU [Madsen and Jensen, 2013] or ASReml [Gilmour et al., 2009]). According to Kolstad and Vangen (1996), there are breed differences in maintenance requirements due to body composition, heat production, temperature, and activity, suggesting there could be individual variation in maintenance requirements. Individual differences in maintenance requirements can be partly due to differences in the same factors as mentioned above and could affect the regression coefficients for fat and lean meat depositions. Hence, the fat and lean meat deposition itself may also influence maintenance requirements, making these effects hard to disentangle. The latter could explain the convergence difficulties of a model with individual regression coefficients on fat deposition, lean meat deposition, and AMW, that is, maintenance requirement.

In general, variation in lean meat and fat efficiency may be caused by 1) actual differences in fat and lean meat efficiency; 2) differences in body composition, which may affect heat production and general activity of the animal; and 3) individual differences in which fraction of the fat is deposited around intestines or on the carcass. By including CT scans of the whole pig (and not just the carcass part), differences due to individual differences in which fraction of the fat is deposited around intestines or on the carcass could be eliminated by correcting for total fat and lean meat deposition.

Implications

The results indicated that significant genetic variation existed in Norwegian Landrace and Duroc in efficiency for deposition of lean meat and fat during the test period and that a significant part of the total genetic variance in feed intake was explained by efficiency in fat and lean meat deposition. A challenge in swine genetics is to make the slaughter pigs more feed effective and be lean but have fatty (juicy) meat and at the same time maintain sufficient feed intake and body condition on the sows. Lean meat and fat efficiency, as defined in model [1], gives the breeders opportunities to select animals that have genetic potential to efficiently deposit lean meat at low feed costs, rather than animals that eat less (due to reallocation of feed resources) or produce less fat. Our novel model enables selection for a feed-efficient pig, with a high fat and lean meat efficiency, without the aforementioned problems.

 

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.