Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Genetics

Estimation of genetic parameters for traits associated with reproduction, lactation, and efficiency in sows1

 

This article in JAS

  1. Vol. 94 No. 11, p. 4516-4529
     
    Received: Dec 28, 2015
    Accepted: Aug 24, 2016
    Published: October 27, 2016


    2 Corresponding author(s): jdekkers@iastate.edu
 View
 Download
 Share

doi:10.2527/jas.2015-0255
  1. D. M. Thekkoot*,
  2. R. A. Kemp,
  3. M. F. Rothschild*,
  4. G. S. Plastow and
  5. J. C. M. Dekkers 2*
  1. * Department of Animal Science, Iowa State University, Ames
     Genesus Inc., Oakville, MB, Canada
     Department of Agriculture, Food and Nutritional Science, University of Alberta, Edmonton, AB, Canada

Abstract

Increased milk production due to high litter size, coupled with low feed intake, results in excessive mobilization of sow body reserves during lactation, which can have detrimental effects on future reproductive performance. A possibility to prevent this is to improve sow lactation performance genetically, along with other traits of interest. The aim of this study was to estimate breed-specific genetic parameters (by parity, between parities, and across parities) for traits associated with lactation and reproduction in Yorkshire and Landrace sows. Performance data were available for 2,107 sows with 1 to 3 parities (3,424 farrowings total). Sow back fat, loin depth and BW at farrowing, sow feed intake (SFI), and body weight loss (BWL) during lactation showed moderate heritabilities (0.21 to 0.37) in both breeds, whereas back fat loss (BFL), loin depth loss (LDL), and litter weight gain (LWG) showed low heritabilities (0.12 to 0.18). Among the efficiency traits, sow lactation efficiency showed extremely low heritability (near zero) in Yorkshire sows but a slightly higher (0.05) estimate in Landrace sows, whereas sow residual feed intake (SRFI) and energy balance traits showed moderate heritabilities in both breeds. Genetic correlations indicated that SFI during lactation had strong negative genetic correlations with body resource mobilization traits (BWL, BFL, and LDL; −0.35 to −0.70), and tissue mobilization traits in turn had strong positive genetic correlations with LWG (+0.24 to +0.54; P < 0.05). However, SFI did not have a significant genetic correlation with LWG. These genetic correlations suggest that SFI during lactation is predominantly used for reducing sow body tissue losses, rather than for milk production. Estimates of genetic correlations for the same trait measured in parities 1 and 2 ranged from 0.64 to 0.98, which suggests that first and later parities should be treated as genetically different for some traits. Genetic correlations estimated between traits in parities 1 and 2 indicated that BWF and BWL measured in parity 1 can be used as indicator traits for SFI and SRFI measured in parities 1 and 2. In conclusion, traits associated with lactation in sows have a sizable genetic component and show potential for genetic improvement.



INTRODUCTION

Lactation, an integral part of reproduction in mammals, is one of the most energy demanding processes. The modern sow is not an exception but provides additional challenges. Litter size in pigs has increased in the recent past and will continue as an important goal in breeding programs (Baxter et al., 2013). Larger litters mean more suckled mammary glands, which results in increased milk output (Auldist et al., 1998). In modern sows, feed consumed during lactation is not sufficient to sustain this high milk production (Noblet et al., 1998). Lactating sows mobilize body reserves in an attempt to maintain milk production when nutrient intake is insufficient (Eissen et al., 2000). However, excess mobilization of body reserves during lactation can lead to fertility complications in future parities (Lundgren et al., 2014). Selection for increased efficiency and leanness during finishing can result in young sows consuming insufficient feed during lactation (Cameron et al., 2002; Gilbert et al., 2012), thereby having negative side effects on subsequent reproductive performance. This can lead to early culling of sows and reduced profitability of pork production. Estimates of the genetic correlation of feed intake during the grow-finish phase with feed intake during lactation have been small but positive (Lewis and Bunter, 2011; Bergsma et al., 2013).

To improve lactation performance along with other traits, lactation traits have to be included in the breeding goal. To optimize the breeding scheme and to predict responses to selection, it is essential to have accurate estimates of genetic parameters. Thus, the main objective of this study was to estimate breed-specific genetic parameters for traits associated with lactation and reproduction in Yorkshire and Landrace sows. An additional objective was to identify indicator traits for predicting economically relevant traits associated with lactation.


MATERIALS AND METHODS

Animals were subjected to standard production conditions in a commercial breeding program, and no additional recordings were made; hence, no approval from the Iowa State University animal care and use committee was required.

Animals

This study was conducted on the basis of data collected from 2 commercial pure-bred maternal lines of Yorkshire and Landrace sows, which were housed together in the same nucleus breeding facility in Canada. Complete information on 3,424 farrowings, recorded between August 2011 and January 2014, was used for this analysis. Numbers of records by breed and parity are in Table 1. As animals were part of a nucleus breeding program, most sows were kept for only 2 parities, and few were kept for a third farrowing. All remaining sows were culled after the third farrowing. Culling of sows was not based on lactation traits.


View Full Table | Close Full ViewTable 1.

Number of farrowing records by breed and parity

 
Parity
Breed No. of sows 1 2 3
Yorkshire 1,075 888 642 237
Landrace 1,032 836 564 257

During gestation, sows were housed in individual stalls and were fed 1.9 to 2.5 kg of gestation diet per day, containing at least 12.6 MJ ME/kg. Around 3 to 5 d before farrowing, sows were moved to the farrowing house and fed a lactation diet containing 13.51 MJ ME/kg. In the farrowing room, sows were housed in individual farrowing pens for the entire period of lactation. Farrowings were planned such that all litters in 1 farrowing room of 28 pens were weaned each week.

Lactation Performance Traits

Bergsma et al. (2009) described in detail the energy metabolism in a lactating sow on the basis of the work of Noblet et al. (1990) and Everts and Dekker (1994a). Major sources of energy for a lactating sow are feed intake and body reserves mobilized during lactation, which are considered energy inputs during lactation. The available energy is used for growth and maintenance of the sow and for producing milk. Energy that is utilized for producing milk, which in turn is used for the growth and maintenance of piglets, is considered as output from the sow during lactation. Quantification of these energy sources allows the energy efficiency of the sow during lactation to be assessed. Most traits defined for this study were related to energy partitioning during lactation and were broadly divided into prefarrow traits, energy input traits, energy output traits, and efficiency traits. We also included several reproductive traits in this analysis. All traits studied, along with abbreviations used and units, are summarized in Table 2.


View Full Table | Close Full ViewTable 2.

Descriptive statistics of the traits studied by breed

 
Yorkshire
Landrace
Trait Trait abbreviation, unit n Mean (SD) n Mean (SD)
Body weight at farrowing BWF, kg 1,758 228.3 (24.0) 1,647 223.5 (23.5)
Back fat at farrowing BFF, mm 1,759 22.4 (4.7) 1,652 22.4 (4.9)
Loin depth at farrowing LDF, mm 1,759 64.0 (7.0) 1,652 65.0 (6.9)
Body weight loss BWL, kg 1,681 11.9 (13.6) 1,580 7.3 (13.5)
Back fat loss BFL, mm 1,688 2.5 (3.9) 1,590 2.4 (3.7)
Loin depth loss LDL, mm 1,686 2.3 (6.9) 1,590 2.6 (7.0)
Sow feed intake SFI, kg 1,767 103.0 (27.0) 1,657 93.8 (25.8)
Energy input EIP, MJ ME/d2 1,611 55.8 (16.4) 1,503 50.6 (14.7)
Litter weight gain LWG, kg 1,760 46.7 (13.01) 1,649 44.9 (12.6)
Energy output EOP, MJ ME/d2 1,760 34.3 (7.2) 1,649 33.7 (7.0)
Sow lactation efficiency SLE, % 1,611 67.2 (21.4) 1,503 73.0 (23.2)
Energy balance 1 EB1, MJ ME/d2 1,611 −11.8 (14.1) 1,516 −10.2 (13.8)
Energy balance 2 EB2, MJ ME/d2 1,681 −14.8 (10.9) 1,580 −17.1 (9.8)
Live born piglets LBP 1,767 13.0 (3.3) 1,657 12.0 (3.0)
Stillborn piglets SBP 1,767 1.6 (1.7) 1,657 1.4 (1.6)
Litter size at birth LSB 1,767 15.0 (3.7) 1,657 13.6 (3.3)
Litter size at weaning LSW 1,767 10.4 (2.5) 1,657 10.0 (2.4)

Prefarrow traits.

Body weight of the sow at farrowing (BWF), adjusted for weight of piglets and placenta, was estimated from BW of the sow at transfer to the farrowing pen (3 to 5 d before farrowing), BW of live-born and stillborn piglets at birth, and the estimated weight of placenta and intrauterine fluids on the basis of equations of Noblet et al. (1985) and Bergsma et al. (2009). Body weight gain from the day of measurement to farrowing was assumed to be negligible. Back fat (BFF) and loin depth (LDF) at farrowing were obtained using a single ultrasound scan taken when the sow was transferred to the farrowing pen on the left side of sow’s body above the last 4 ribs, parallel to the backbone, using an Aloka SSD-500V (Hitachi Aloka, Tokyo, Japan) ultrasound scanning machine.

Energy Input Traits.

Feed intake and body tissue mobilization are the 2 major sources of energy for a sow during lactation. Sow feed intake (SFI) over the lactation period was measured using the Gestal FM computerized feeding system from JYGA Technologies (Saint-Nichols, QC, Canada) that was installed in each farrowing pen. This equipment fed sows multiple times per day in a precise and continuous manner. Each feeder was mounted on the tubing of the feed delivery system and accurately dispensed a measured portion of feed using a motorized auger when the sow stimulated the electronic feed activator at the bottom of the feed trough with her snout. As the activator was not accessible when it was covered with feed, the sow had to empty the trough to activate the feeder, which avoided wastage and overestimation of feed consumption. Each feeder recorded real-time feed disposal and communicated the data to a central server. The system allowed restriction of feed intake, and accordingly, sows were fed on an ascending scale up to d 4 after farrowing and were then fed ad libitum until weaning.

Body energy mobilized by the sow during lactation was estimated using BW and back fat at farrowing and weaning. Sow BW at weaning was corrected for water content in the mammary gland on the basis of experiments conducted by Kim et al. (1999a,b, 2000) and equations derived by Bergsma et al. (2009). Body weight loss (BWL) of the sow during lactation was calculated by subtracting the corrected BW at weaning from BWF. Back fat loss (BFL) and loin depth loss during lactation (LDL) were calculated by subtracting ultrasound back fat and loin depth at weaning from those at farrowing. A positive value for these tissue mobilization traits indicates a loss during lactation.

Energy available for milk production from SFI and body resource mobilization constitutes the energy input (EIP) for the sow during lactation and was calculated following Bergsma et al. (2009) asEnergy mobilized from body fat and protein mass during lactation was calculated from BWL and BFL (Bergsma et al., 2009). The maintenance energy requirement of the sow was derived from its metabolic BW (average BW0.75; Noblet et al., 1990).

Energy Output Traits.

A lactating sow utilizes energy inputs for milk production, which in turn is used for growth and maintenance of piglets. Since direct measurement of milk yield is not practical in sows, litter weight gain (LWG) from birth to weaning was used as an indicator. All nonmummified piglets born were weighed at birth, death, weaning, and the time of fostering to quantify the weight gain of each piglet for each sow. Litter weight gain for a litter was the increase in weight of all piglets nursed by that sow.

As another method to estimate the energy output of a sow per day of lactation (EOP), Bergsma et al. (2009) derived the following equation based on the work of Everts and Dekker (1994a,b) based on piglet body fat and protein mass gain from birth to weaning:

Efficiency Traits.

Energetic efficiency of a sow during lactation was estimated in 4 ways. Sow lactation efficiency (SLE) was calculated as defined by Bergsma et al. (2008):Energy balance (EB1), as defined by Young et al. (2016) on a per day basis, was used to measure the body energy loss of a sow during lactation:Energy retained at farrowing and that retained at weaning were calculated from the estimated fat mass and protein mass in the body, which in turn were estimated from the back fat and BW measured at farrowing and weaning. Fat and protein mass at weaning and farrowing were calculated following Bergsma et al. (2009) and Young et al. (2016). Note that sows that lose (gain) body reserves during lactation have a negative (positive) value for EB1.

Another way of estimating the energy balance (EB2), as used in dairy cattle (Spurlock et al., 2012), is based on the difference between the dietary energy consumed and energy expenditure required for milk production and maintenance and was estimated using the following equation, with all component traits expressed on a megajoule of ME per day basis and constants reflecting energy efficiencies, following Noblet et al. (1990):

Gilbert et al. (2012) defined sow residual feed intake during lactation (SRFI) as the difference between the observed and predicted daily feed intake for maintenance and production of a sow during lactation. In this study, prediction of daily feed intake was based on multiple regression of SFI on BWL, BFL, LDL, LWG, EB1, sow metabolic midweight, piglet load (explained later), and lactation length, along with the fixed class effects of parity and year-season of farrowing.

Reproduction Traits.

Traits included in this category were number of live-born piglets (LBP), defined as all piglets that were alive at the time of birth; stillborn piglets (SBP), defined as all nonmummified piglets that were born dead; litter size at birth (LSB), which was the sum of born alive, born dead, and mummies; and litter size at weaning (LSW), which was the number of live piglets weaned by a sow, regardless of whether she farrowed them or not.

Piglet Load.

For studies involving lactation in pigs, most researchers include the number of piglets weaned by the sow as a covariate to account for nutrient demands on the sow, but this may not reflect all events that occur during lactation. For example, 2 sows with the same number of piglets born alive, numbers fostered on and off, and preweaning mortality will have the same number of piglets weaned at the end of lactation. But if the fostering and preweaning mortalities occurred at different time points in lactation, the nutrient demands will differ between these 2 sows. In this study, piglet load was introduced to address this by summing the number of piglets suckled on each day across the lactation period.

Statistical Analyses

For each breed, variance and covariance components were estimated separately using ASReml (Gilmore et al., 2009) on the basis of univariate and bivariate mixed linear animal models. Complete pedigrees up to 3 generations were used, and there was no sharing of pedigree between sows from the 2 breeds. Trait heritabilities were estimated separately for each breed across all 3 parities, using a single-trait repeatability animal model. For each trait, heritabilities were also estimated separately for parity 1 and 2 measurements but not for parity 3 because of insufficient numbers of records. In the repeatability models, for all traits, fixed effects included were parity (3 levels) and year-season of farrowing (12 levels). The direct additive genetic value of the sow, the permanent environmental effect of the sow, and contemporary group based on year and week of weaning were included as random effects for all traits. The average contemporary group size (year and week of weaning) was 13.4 for Landrace and 14.4 for Yorkshire. Traits BWL, BFL, LDL, SFI, EIP, LWG, EOP, SLE, EB1, and EB2 depend on measurements recorded on sows or piglets during lactation or at weaning. For these traits, to account for nongenetic variation associated with lactation, additional covariates were used such as lactation length, average birth weight of weaned piglets, the number of piglets in each of 5 classes of starting weights (to account for differences in the growth potential of fostered piglets), sow BW, back fat and loin depth measured at farrowing, the proportion of male piglets in the litter, and piglet load. For LSW, lactation length was added as a covariate. To estimate the heritability of traits within a parity (parity 1 or 2), the same models were used but without the permanent environmental effect of the sow and parity. Not all covariates included were statistically significant, but they were retained in the model to account for the nongenetic variation that was expected to be associated with them.

A series of bivariate analyses were run to estimate genetic and phenotypic correlations between traits across parities. Model terms included were the same as in the repeatability models described previously. Genetic and phenotypic correlations were also estimated between traits within parities 1 and 2, using the same models as described previously for single-parity analyses. Bivariate analyses for some pairs of traits, e.g., BFL and BFF, where 1 trait (BFF) is used both as a trait and as a covariate, resulted in inconsistent estimates of genetic parameters when solved by maximum likelihood. To avoid this, 1 trait (BFL in this example) was preadjusted for the second trait (BFF in this example) before it was analyzed in the 2-trait animal model, following Cai et al. (2008). Regression coefficients used for preadjustment were obtained from the respective single-trait models. Genetic correlations were also estimated between the same trait measured in parities 1 and 2, using a series of bivariate analyses, assuming that the parity 1 trait is genetically different from the parity 2 trait, following Oh et al. (2006). Models used were the same as described before, except that the effects of permanent environment and parity were excluded.

Indicator Traits for Economically Relevant Traits

Among the lactation traits described above, some directly influence the cost or income from production and are hence considered as economically relevant traits (ERT) and are breeding objective traits. This includes LSB, SFI, LWG, and SRFI. Measuring these ERT on a regular basis on all sows in a commercial setting may not be possible because some are costly to measure (SFI, SRFI, and LWG), and some can only be recorded later in life (second-parity LSB). An alternative is to use indicator traits as selection criteria for these breeding goal traits, which include any trait that has a sizeable genetic correlation with an ERT. Traits evaluated as indicator traits for ERT in this study were BWF, BFF, LDF, BWL, BFL, and LDL in parity 1, which can be measured easily and economically compared to the ERT. Genetic correlations of all indicator traits with all ERT were estimated using bivariate animal models. Fixed and random effects used in the model were the same as described previously, except for the effects of permanent environment and parity.


RESULTS AND DISCUSSION

Descriptive Statistics

The number of observations available for each trait varied as not all sows had complete information for all required parameters (Table 2). The efficiency trait SLE requires information on most component traits and hence had the lowest number of observations. Average lactation length was 20.6 ± 3.3 d for Yorkshire sows and 19.9 ± 3.4 d for Landrace sows. In general, Yorkshire sows were heavier at farrowing and weaned larger and heavier litters but consumed more feed and lost more body tissue during lactation than Landrace sows; that is, they had greater energy inputs. Outputs (LWG and EOP) were only slightly higher for Yorkshire than for Landrace sows, resulting in slightly lower SLE for Yorkshire compared to Landrace sows. The average SLE (67% and 73% for Yorkshire and Landrace sows, respectively) were higher than those reported by Young et al. (2016) for experimental lines of Yorkshire sows divergently selected for residual feed intake (RFI) during finishing (61% and 58% for the low- and high-RFI lines, respectively) but comparable to figures reported by Bergsma et al. (2008; 68%) for commercial Yorkshire and Landrace crossbred sows.

Traits EB1 and EB2 both measure the sow’s body energy mobilized during lactation. The negative average values for EB1 and EB2 (Table 2) indicate that, on average, sows mobilized body resources during lactation, resulting in a negative energy balance. Theoretically, EB1 and EB2 should produce comparable values, but the averages for EB2 were slightly lower than those for EB1, likely because the equations for estimation of EB1 and EB2 were derived on the basis of works from various authors on different groups of pigs from different geographical locations. Young et al. (2016) reported a larger average negative energy balance (EB1) of −19.9 MJ ME/d for sows selected for low RFI, whereas the high-RFI line had an average EB1 that was slightly lower than the averages observed here (−8.0 MJ ME/d; Table 2). No literature on lactating sows was found to compare the values obtained for EB2.

Heritabilities

The prefarrow traits BFF and LDF showed moderate heritabilities (0.21 to 0.38), whereas traits associated with mobilization of back fat and loin depth (BFL and LDL) had low heritabilities (0.10 to 0.16) in both breeds (Tables 3 and 4). Estimates of heritability for BFL were comparable to those reported by Grandinson et al. (2005) and Gilbert et al. (2012). Our estimate of heritability for BWF was near the lower end of the range of previously reported values, which ranged from 0.16 (Gilbert et al., 2012) and 0.19 (Grandinson et al., 2005) to 0.45 (Bergsma et al., 2008) and 0.66 (Young et al., 2016). Among the tissue mobilization traits, BWL showed higher heritability than BFL and LDL. All previous studies, except Gilbert et al. (2012), also found BWL to have higher heritability than BFL.


View Full Table | Close Full ViewTable 3.

Estimates of heritability (parity1, parity 2, and overall), permanent environment effects, and genetic correlations for the same trait measured in parities 1 and 2 in Yorkshire sows

 
Variance
Heritability (SE)
Trait1 Sow Permanent environment Phenotypic3 Overall Heritability (SE) Permanent environment effects (SE) Parity 1 Parity 2 Genetic correlation between parities 1 and 2 (SE)2
BWF 67.28 104.10 285.10 0.24 (0.05) 0.36 (0.04) 0.26 (0.07) 0.27 (0.09) 0.74 (0.13)
BFF 5.83 4.56 20.04 0.29 (0.06) 0.23 (0.05) 0.33 (0.08) 0.22 (0.08) 0.98 (0.11)
LDF 12.36 4.22 40.06 0.31 (0.05) 0.11 (0.05) 0.35 (0.08) 0.29 (0.09) 0.65 (0.16)
BWL 37.46 21.43 150.72 0.25 (0.05) 0.14 (0.05) 0.34 (0.08) 0.25 (0.08) 0.91 (0.10)
BFL 0.92 0.80 9.44 0.10 (0.04) 0.09 (0.05) 0.03 (0.05) 0.00 (0.00) NE
LDL 3.77 3.10 29.25 0.13 (0.04) 0.11 (0.05) 0.12 (0.06) 0.19 (0.09) 0.90 (0.30)
SFI 65.73 36.46 236.02 0.28 (0.05) 0.15 (0.05) 0.31 (0.08) 0.23 (0.08) 0.86 (0.16)
EIP 2.05 32.77 161.82 0.01 (0.03) 0.20 (0.05) 0.00 (0.00) 0.03 (0.07) NE
LWG 10.78 4.86 67.19 0.16 (0.04) 0.07 (0.05) 0.22 (0.08) 0.17 (0.08) 0.85 (0.17)
EOP 3.57 1.85 21.22 0.17 (0.05) 0.09 (0.05) 0.21 (0.08) 0.20 (0.09) 0.83 (0.18)
SLE 0.00 33.76 394.42 0.00 (0.00) 0.09 (0.03) 0.00 (0.00) 0.00 (0.00) NE
EB1 18.86 16.88 143.34 0.13 (0.04) 0.12 (0.05) 0.11 (0.07) 0.08 (0.08) NE
EB2 23.07 10.50 77.28 0.30 (0.05) 0.14 (0.05) 0.34 (0.08) 0.19 (0.08) 0.92 (0.16)
SRFI 49.79 30.15 193.38 0.26 (0.05) 0.16 (0.05) 0.22 (0.08) 0.19 (0.08) 0.87 (0.16)
LBP 0.58 0.87 9.24 0.06 (0.03) 0.09 (0.05) 0.06 (0.05) 0.15 (0.08) 0.70 (0.45)
SBP 0.35 0.20 2.99 0.12 (0.04) 0.07 (0.05) 0.17 (0.07) 0.09 (0.06) 0.88 (0.32)
LSB 1.48 1.85 13.49 0.11 (0.04) 0.14 (0.05) 0.04 (0.05) 0.25 (0.09) 0.86 (0.56)
LSW 0.23 0.66 5.21 0.04 (0.03) 0.13 (0.05) 0.00 (0.05) 0.12 (0.08) NE
1BWF = body weight at farrowing; BFF = back fat at farrowing; LDF = loin depth at farrowing; BWL = body weight loss; BFL = back fat loss; LDL = loin depth loss; SFI = total feed intake; EIP = energy input; LWG = litter weight gain; EOP = energy output; SLE = sow lactation efficiency; EB1 = energy balance 1; EB2 = energy balance 2; SRFI = sow residual feed intake; LBP = live-born piglets; SBP = stillborn piglets; LSB = litter size at birth; LSW = litter size at weaning.
2NE = convergence not attained, and hence, estimates could not be obtained.
3Phenotypic variance was calculated from ASReml analyses : σ2 error + σ2 sow + σ2 permanent environment.

View Full Table | Close Full ViewTable 4.

Estimates of heritability (parity1, parity 2, and overall), permanent environment effects, and genetic correlations for the same trait measured in parities 1 and 2 in Landrace sows

 
Variance
Heritability (SE)
Trait1 Sow Permanent environment Phenotypic3 Overall Heritability (SE) Permanent environment effects (SE) Parity 1 Parity 2 Genetic correlation between parities 1 and 2 (SE)2
BWF 53.84 68.72 256.43 0.21 (0.05) 0.27 (0.05) 0.34 (0.08) 0.10 (0.07) 0.81 (0.15)
BFF 8.59 4.11 22.84 0.38 (0.06) 0.18 (0.05) 0.47 (0.08) 0.16 (0.08) 0.77 (0.28)
LDF 9.57 3.27 39.50 0.24 (0.05) 0.08 (0.05) 0.20 (0.07) 0.19 (0.10) NE
BWL 44.02 0.00 135.59 0.32 (0.04) 0.00 (0.00) 0.41 (0.08) 0.42 (0.11) 0.84 (0.11)
BFL 1.53 0.50 9.71 0.16 (0.04) 0.05 (0.05) 0.07 (0.06) 0.20 (0.09) 0.88 (0.49)
LDL 3.62 1.02 29.28 0.12 (0.04) 0.03 (0.05) 0.07 (0.06) 0.20 (0.10) NE
SFI 76.70 16.76 206.26 0.37 (0.06) 0.08 (0.05) 0.33 (0.08) 0.19 (0.10) 0.93 (0.21)
EIP 16.57 12.06 153.32 0.11 (0.04) 0.08 (0.05) 0.00 (0.05) 0.22 (0.09) NE
LWG 6.83 0.66 38.57 0.18 (0.05) 0.02 (0.05) 0.12 (0.06) 0.20 (0.10) 0.80 (0.23)
EOP 2.35 1.13 13.44 0.17 (0.05) 0.08 (0.05) 0.10 (0.06) 0.17 (0.10) 0.85 (0.22)
SLE 23.58 12.85 437.11 0.05 (0.03) 0.03 (0.05) 0.00 (0.00) 0.16 (0.08) NE
EB1 32.70 0.00 140.39 0.23 (0.05) 0.00 (0.00) 0.26 (0.08) 0.25 (0.09) 0.65 (0.25)
EB2 22.60 4.35 61.65 0.36 (0.06) 0.07 (0.05) 0.45 (0.09) 0.21 (0.10) NE
SRFI 50.59 12.31 170.17 0.30 (0.06) 0.07 (0.05) 0.34 (0.08) 0.11 (0.08) 0.93 (0.16)
LBP 0.59 0.84 7.93 0.07 (0.04) 0.11 (0.05) 0.21 (0.07) 0.04 (0.07) 0.64 (0.48)
SBP 0.28 0.32 2.52 0.11 (0.04) 0.13 (0.05) 0.09 (0.06) 0.19 (0.10) 0.88 (0.32)
LSB 1.76 0.80 10.54 0.17 (0.05) 0.08 (0.05) 0.27 (0.07) 0.06 (0.07) 0.65 (0.35)
LSW 0.08 0.75 4.44 0.02 (0.03) 0.17 (0.05) 0.00 (0.00) 0.09 (0.07) NE
1BWF = body weight at farrowing; BFF = back fat at farrowing; LDF = loin depth at farrowing; BWL = body weight loss; BFL = back fat loss; LDL = loin depth loss; SFI = Total feed intake; EIP = energy input; LWG = litter weight gain; EOP = energy output; SLE = sow lactation efficiency; EB1 = energy balance 1; EB2 = energy balance 2; SRFI = sow residual feed intake; LBP = live-born piglets; SBP = stillborn piglets; LSB = litter size at birth; LSW = litter size at weaning.
2NE = convergence not attained, and hence, estimates could not be obtained.
3Phenotypic variance was calculated from ASReml analyses: σ2 error + σ2 sow + σ2 permanent environment.

Heritability estimates of SFI were moderate for both breeds (0.28 for Yorkshire and 0.37 for Landrace) and comparable to estimates reported by Bergsma et al. (2008), Gilbert et al. (2012), and Young et al. (2016) but higher than the estimate reported by Hermesch (2007; 0.17). Sows in this study were fed ad libitum during lactation using electronic feeders, except for the first 4 d, whereas sows were either restricted or hand fed in the studies referenced. Heritability estimates for EIP were low (0.01 in Yorkshire and 0.11 in Landrace). The estimate for Landrace sows was comparable to values previously reported by Bergsma et al. (2008). Energy input is a complex trait and depends on SFI, BWL, and BFL. Of these, SFI and BWL had moderate heritability in both parities for both breeds, but BFL had very low heritability estimates in parity 1 and 2 Yorkshire sows and in parity 1 Landrace sows (Supplemental Tables 2 to 5), which likely explains the low heritability estimates for EIP. Of the traits that were used to calculate BFL, BFF showed moderate to high heritability for both breeds in both parities, whereas back fat measured at weaning showed near-zero heritability in both parities for Yorkshire and in parity 1 Landrace sows.

Traits LWG and EOP reflect the output by the sow based on piglet BW gains and had very similar heritability estimates for both breeds (Tables 4 and 5). The estimate of heritability for LWG was moderate (0.16 for Yorkshire and 0.18 for Landrace) and was close to the upper range of previously reported figures, ranging from 0.06 (Young et al., 2016) and 0.09 (Gilbert et al., 2012) to 0.18 (Bergsma et al., 2008).


View Full Table | Close Full ViewTable 5.

Estimates of genetic correlations between prefarrow, input, output, efficiency, and reproductive traits for Yorkshire (below diagonal) and Landrace sows (above diagonal)1

 
Traits2 BWF BFF LDF BWL BFL LDL SFI LWG EOP SLE EB1 EB2 SRFI LBP SBP LSB LSW
BWF 0.49 (0.13) 0.02 (0.17) 0.01 (0.16) −0.41 (0.18) −0.18 (0.21) 0.36 (0.15) 0.18 (0.18) 0.08 (0.19) 0.53 (0.23) 0.39 (0.14) 0.13 (0.17) 0.27 (0.16) 0.00 (0.22) −0.08 (0.17) −0.15 (0.17) NE
BFF 0.43 (0.13) 0.39 (0.13) −0.29 (0.09) −0.903 (0.08) −0.25 (0.20) −0.26 (0.12) 0.18 (0.13) 0.09 (0.13) −0.35 (0.36) 0.03 (0.16) 0.07 (0.13) −0.26 (0.12) −0.07 (0.18) −0.43 (0.17) −0.25 (0.16) −0.50 (0.66)
LDF 0.39 (0.14) 0.23 (0.14) −0.063 (0.15) −0.02 (0.18) −0.623 (0.17) −0.11 (0.15) −0.12 (0.18) −0.22 (0.18) −0.03 (0.28) 0.163 (0.16) 0.09 (0.15) −0.18 (0.15) 0.00 (0.21) −0.34 (0.17) −0.16 (0.17) −0.09 (0.36)
BWL −0.30 (0.18) −0.23 (0.15) −0.23 (0.14) 0.01 (0.18) 0.49 (0.16) −0.60 (0.09) 0.49 (0.16) 0.49 (0.17) −0.12 (0.25) −0.80 (0.07) −0.73 (0.08) 0.083(0.15) −0.16 (0.18) −0.28 (0.20) −0.21 (0.16) −0.08 (0.38)
BFL −0.66 (0.13) −0.65 (0.17) −0.44 (0.19) 0.51 (0.17) 0.31 (0.19) −0.63 (0.15) 0.39 (0.16) 0.53 (0.17) −0.54 (0.14) −0.90 (0.02) −0.23 (0.13) −0.51 (0.21) 0.17 (022) −0.11 (0.24) 0.12 (0.19) 0.03 (0.45)
LDL −0.37 (0.17) 0.11 (0.19) −0.69 (0.15) 0.60 (0.15) 0.47 (0.22) −0.11 (0.19) 0.39 (0.21) 0.29 (0.22) −0.36 (0.34) −0.42 (0.16) −0.31 (0.17) 0.14 (0.19) 0.42 (0.26) 0.00 (0.27) 0.33 (0.22) −0.53 (0.57)
SFI 0.50 (0.15) −0.09 (0.15) −0.08 (0.14) −0.70 (0.09) −0.52 (0.21) −0.35 (0.17) 0.31 (0.18) 0.23(0.18) −0.48 (0.20) 0.58 (0.10) 0.90 (0.04) 0.92 0.01) −0.01 (0.19) 0.38 (0.21) 0.08 (0.17) 0.40 (0.38)
LWG −0.05 (0.19) 0.04 (0.19) −0.45 (0.16) 0.24 (0.17) 0.42 (0.21) 0.52 (0.19) 0.12 (0.17) NE −0.27 (0.23) −0.40 (0.14) −0.20 (0.11) 0.283 (0.16) −0.02 (0.22) 0.14(0.24) 0.03 (0.20) 0.33 (0.45)
EOP −0.04 (0.19) 0.01(0.18) −0.41 (0.16) 0.29 (0.16) 0.53 (0.20) 0.57 (0.18) 0.06 (0.18) NE −0.27 (0.29) −0.43 (0.14) −0.39 (0.14) 0.23 (0.16) −0.12 (0.22) 0.13 (0.25) −0.08 (0.20) 0.59 (0.38)
SLE NE NE NE NE NE NE NE NE NE NE NE −0.70 (0.11) −0.38 (0.31) 0.10 (0.37) −0.31 (0.27) −0.11 (0.71)
EB1 0.71 (0.13) 0.48 (0.15) 0.45 (0.16) −0.91 (0.07) −0.88 (0.06) −0.67 (0.17) 0.92 (0.14) −0.33 (0.19) −0.43 (0.18) NE 0.73 (0.11) 0.35 (0.17) −0.11 (0.20) 0.12 (0.22) −0.09 (0.18) −0.11 (0.42)
EB2 0.35 (0.16) −0.03 (0.15) 0.18 (0.14) −0.74 (0.09) −0.75 (0.15) −0.60 (0.15) 0.85 (0.06) −0.52 (0.13) −0.56 (0.12) NE 0.85 (0.08) NE 0.14 (0.19) 0.23 (0.20) 0.19 (0.16) 0.06 (0.39)
SRFI 0.42 (0.15) −0.10 (0.15) −0.01 (0.14) −0.66 (0.11) −0.43 (0.24) −0.27 (0.18) 0.95 (0.01) 0.13 (0.18) −0.07 (0.18) NE 0.69 (0.19) NE −0.06 (0.21) 0.41 (0.21) 0.08 (0.17) 0.33 (0.49)
LBP −0.30 (0.24) −0.31 (0.23) −0.44 (0.22) −0.23 (0.22) −0.32 (0.33) −0.01 (0.27) −0.06 (0.24) 0.03 (0.28) −0.02 (0.28) NE 0.16 (0.28) −0.03 (0.22) −0.33 (0.23) 0.15 (0.27) NE 0.19 (0.55)
SBP −0.32 (0.19) −0.20 (0.19) −0.26 (0.18) 0.18 (0.18) 0.35 (0.25) 0.08 (0.23) −0.09 (0.19) 0.09 (0.22) 0.14 (0.21) NE −0.33 (0.21) −0.24 (0.19) −0.09 (0.20) −0.08 (0.29) 0.59 (0.18) 0.26 (0.64)
LSB −0.39 (0.20) −0.41 (0.20) −0.49 (0.17) −0.13 (0.19) −0.01 (0.29) 0.01 (0.24) −0.09 (0.20) 0.01 (0.24) −0.01 (0.24) NE −0.06 (0.24) −0.13 (0.20) −0.31 (0.21) NE 0.54 (0.18) 0.30 (0.62)
LSW 0.21 (0.31) −0.23 (0.29) −0.12 (0.29) 0.20 (0.26) 0.26 (0.40) 0.26 (0.32) −0.04 (0.27) 0.86 (0.21) 0.80 (0.16) NE −0.35 (0.30) −0.61 (0.18) −0.34 (0.24) −0.13 (0.43) 0.28 (0.36) 0.17 (0.37)
1SE are in parentheses. Bold genetic correlations differ from zero (P < 0.05). NE = Convergence not attained and hence estimates could not be obtained.
2BWF = body weight at farrowing; BFF = back fat at farrowing; LDF = loin depth at farrowing; BWL = body weight loss; BFL = back fat loss; LDL = loin depth loss; SFI = total feed intake; LWG = litter weight gain; EOP = energy output; SLE = sow lactation efficiency; EB1 = energy balance 1; EB2 = energy balance 2; SRFI = sow residual feed intake; LBP = live-born piglets; SBP = stillborn piglets; LSB = litter size at birth; LSW = litter size at weaning.
3Convergence attained using preadjusted traits.

Among the efficiency traits, the heritability estimate of SLE was zero in Yorkshire sows (Table 3) and 0.05 in Landrace sows (Table 4). For Landrace sows, our estimate was lower than the previously reported values of 0.12 (Bergsma et al., 2008) and 0.09 (Young et al., 2016). Lactation efficiency is a ratio trait of EIP and EOP. Trait EOP showed moderate heritability for each breed (0.17), but EIP showed very low heritability for Yorkshire sows (0.01) compared to Landrace sows (0.11), which might be a reason for the very low heritability estimate of SLE for Yorkshire sows.

The heritability estimate obtained for SRFI was 0.24 for Yorkshire sows and 0.25 for Landrace sows (Tables 3 and 4). These estimates were higher than previously reported values of 0.14 (Gilbert et al., 2012) and 0.16 (Young et al., 2016). Heritability estimates of EB1 were low (0.13 for Yorkshire and 0.19 for Landrace) but were similar to those reported by Young et al. (2016) for Yorkshire sows selected for high and low RFI. No literature was available on sows to compare our estimate of heritability for EB2, but our estimates of 0.30 and 0.36 were higher than the range of values reported for a similarly computed trait in dairy cows (Spurlock et al., 2012). Of the 4 efficiency traits studied, SRFI and EB2 exhibited higher heritabilities than SLE and EB1.

Heritability estimates of reproductive traits were low for both breeds (Tables 3 and 4) and were comparable to estimates previously published by Bergsma et al. (2008), Bidanel (2011), and Gilbert et al. (2012). Reproductive traits such as litter size and litter survival are complex traits resulting from the interaction of many factors, such as the genetics of the sow, boar, and piglet, and hence can have low heritabilities (Bidanel, 2011). Another reason for the differences in heritabilities between our study and previously published results might be differences in the statistical models used.

For each trait, heritabilities were also estimated separately for parities 1 and 2 using single-trait animal models. In Yorkshire sows, for all traits except reproductive traits, estimates of heritability by parity were close to estimates obtained with the repeatability model across all 3 parities (Tables 4 and 5). In Landrace sows, parity 1 estimates of heritability were different from those of parity 2, but there was no clear pattern across traits. Among the reproductive traits, parity 2 heritability estimates for LSB and LBP for Yorkshire sows were higher than parity 1 estimates and also higher than estimates reported by Alfonso et al. (1997) and Oh et al. (2006) for parity 2 sows. Conversely, for Landrace sows, parity 1 estimates of LSB and LBP were higher than parity 2 estimates, and these estimates were higher than those reported by Alfonso et al. (1997) but comparable to estimates reported by Oh et al. (2006). Within-breed differences in heritability estimates between parities 1 and 2 can also be the result of limited numbers of animals with parity 1 and 2 records.

Genetic Correlations

Genetic correlations for the same trait measured in parities 1 and 2 ranged from 0.64 to 0.95 for Landrace sows and from 0.65 to 0.98 for Yorkshire sows (Tables 4 and 5). These genetic correlations were not estimable for some breed by trait combinations, either because of near-zero additive genetic variances in one parity or because of a high genetic correlation between the same trait measured in 2 parities. Estimates of genetic correlations between traits, across the 3 parities, from bivariate repeatability models are in Table 5. Corresponding phenotypic correlations are in Supplemental Table 1 (see online version of journal for Supplemental material). Estimates of genetic correlations between various trait categories are further described in the following.

Genetic Correlations of Prefarrow Traits with Body Resource Mobilization Traits.

The physiological urge of a lactating sow to produce milk at the expense of other body functions is a key component of the metabolic state of the lactating sow and is controlled by factors such as genetics, parity, stage of lactation, and litter size (Pettigrew et al., 1993). Body weight and body composition at farrowing are important factors that regulate lactation performance in sows (Eissen et al., 2000). Results from this study (Table 5) show that body reserves at the start of lactation have a negative genetic correlation with body tissue mobilization during lactation (−0.29 to −0.90); that is, sows that genetically have greater body reserves at farrowing mobilized fewer body reserves during lactation. Bergsma et al. (2008) also observed a similar trend in a population of commercial crossbred sows, but their estimates were not significantly different from zero.

Genetic Correlations of Prefarrow Traits with Lactation Feed Intake.

In our data, BWF showed a significant positive genetic correlation with SFI (0.50 for Yorkshire and 0.36 for Landrace; Table 5) but a negligible correlation at the phenotypic level (0.07 for Yorkshire and 0.04 for Landrace; Supplemental Table 1). Bergsma et al. (2008) also reported a positive genetic correlation between BWF and SFI. In a review, Eissen et al. (2000) concluded that SFI depends on the rate of turnover of body fat tissue, insulin, and leptin levels in blood and cerebrospinal fluid; the presence of insulin resistance and glucose intolerance; and the level of milk production.

Genetic Correlations of Prefarrow Traits with Output Traits.

Grandinson et al. (2005) reported that there was no genetic association between BWF and the maternal effect of preweaning growth of piglets. Similarly, Bergsma et al. (2008) reported nonsignificant negative genetic correlations (−0.08 to −0.13) between body resources (weight, body fat, and protein mass) at farrowing and LWG. In our study, for all prefarrow traits except LDF, genetic correlations with output traits were also not significantly different from zero (Table 5). Loin depth at farrowing showed significant negative genetic correlations with LWG and EOP for Yorkshire sows (−0.45 and −0.41). These genetic correlations were also negative for Landrace sows but with high SE. Thus, sows that had genetically high LDF weaned lighter litters, which was supported by the negative genetic correlation of LDF with tissue mobilization traits (BWL, BFL, and LDL); that is, sows with genetically high LDF mobilized fewer body reserves and weaned piglets with lower weights. Fat sows have fewer protein reserves to supply substrates for milk production compared to lean sows of similar weight (Revell et al., 1998), which might be a reason for their lower milk production and low LWG.

Genetic Correlations of Output Traits with Body Resource Mobilization Traits.

Estimates of genetic correlations of output traits (LWG and EOP) with body tissue mobilization traits (BWL, BFL, and LDL) were significantly positive for both breeds (ranging from 0.24 to 0.57 in Yorkshire and from 0.29 to 0.53 in Landrace; Table 5). Thus, sows with a high genetic predisposition to use body reserves during lactation also had the ability to wean heavier piglets, in agreement with the literature: Bergsma et al. (2008) reported positive (nonsignificant) genetic correlations of BWL, fat mass loss, and protein mass loss with LWG. Grandinson et al. (2005) reported significant positive genetic correlations of changes in sow BW and back fat during lactation with maternal genetic effects of piglet survival and growth to weaning; Valros et al. (2003) reported that a greater sow BWL during the third week of lactation was strongly associated (genetically) with higher piglet growth rate.

Genetic Correlations between Output Traits, Body Resource Mobilization Traits, and Feed Intake.

Estimates of genetic correlations of SFI with output traits (LWG and EOP) were slightly positive for both breeds, ranging from 0.06 to 0.31 (Table 5), but not significantly different from zero and lower than genetic correlations observed between tissue mobilization and output traits. Thus, output traits were more strongly correlated (genetically) with body tissue mobilization traits than with feed intake. At the same time, genetic correlations of feed intake with body tissue loss traits were strongly negative (−0.35 to −0.70), indicating that sows that had the genetic ability to eat more during lactation had less body tissue loss. A similar trend was observed by Bergsma et al. (2008) for crossbred sows, by Eissen et al. (2003) in primiparous Landrace and crossbred sows, and by Koketsu et al. (1997) in primiparous sows. This pattern of association, that is, strong genetic correlations of SFI with tissue mobilization traits and a weak genetic correlation of SFI with LWG, reflects how dietary energy is partitioned in the sow’s body. Pluske et al. (1998), in a phenotypic study using overfed and ad libitum fed primiparous lactating sows, observed similar growth for piglets in both groups. They concluded that the extra energy received by overfed sows through extra feed intake is partitioned more toward the sow’s body maintenance than to milk production.

This genetic independence of milk production from feed consumed during lactation supports the evolutionary importance of lactation. The mammary gland is hypothesized to have evolved from apocrine-like glands associated with hair follicles, and later, these glands evolved from providing primarily moisture and antimicrobials to providing parchment-shelled eggs to the role of supplying nutrients for offspring (Oftedal, 2002). Although converting food to body reserves and milk is less efficient than delivering nutrients directly to offspring, lactation as a process evolved because lactating females can draw on their body reserves for milk production, and this ability offers mothers and their young ones a reasonable level of independence from fluctuations in food supply (Pond, 1977). Dall and Boyd (2004) showed that animals that lactate have a dramatic advantage in terms of success of reproduction during energetic shortfalls compared to animals that nurse their young by regurgitating a portion of their food. This evolutionary adaptation, which makes lactation independent of fluctuations in energy supply during lactation, might explain the low genetic correlation of SFI with LWG.

Genetic and phenotypic correlations by parity for all traits are given in Supplemental Tables 2 to 5. In both parities, genetic correlations of SFI with traits associated with tissue mobilization (BWL, BFL, and LDL) were strongly negative (−0.54 to −0.80). Eissen et al. (2003) and Koketsu et al. (1997) reported similar results for primiparous sows. The pattern of genetic correlations of SFI with output traits (LWG and EOP) was slightly different between parity 1 and 2 sows. In parity 1, genetic correlations of SFI with output traits were very low and ranged from −0.31 to 0.13, but in parity 2, estimates were slightly higher and positive, ranging from 0.29 to 0.34. For both parities, SE of the estimates were very large, and hence, none of the estimates were significantly different from zero. Nevertheless, these trends suggest that in parity 1, when sows are still in a growing phase, more of the energy from feed is prioritized toward body growth or maintenance than to milk production than in parity 2. In parity 2, sows have almost reached mature body size, and hence, energy requirements for growth are lower, which may explain the slightly higher genetic correlations between feed intake and LWG in parity 2.

Estimates of genetic correlations of SFI in parity 1 with output traits (LWG and EOP) in parity 2 (Supplemental Tables 6 and 7) were not significantly different from zero. At the same time, SFI in parity 1 was strongly correlated (genetically) with tissue mobilization traits in both parity 1 (Supplemental Tables 2 and 3) and parity 2 (Supplemental Tables 6 and 7). All these findings affirm that the feed consumed by the sow during lactation is predominantly used for reducing sow body tissue losses, rather than for milk production (litter growth).

Genetic Correlations with Lactation Efficiency Traits.

In this study, 4 efficiency traits were considered, SLE, EB1, EB2, and SRFI. Genetic correlations of SLE with other traits were estimated only for Landrace sows, as the estimates did not converge for Yorkshire sows (Table 5) because of low estimates of additive genetic variance for SLE. In Landrace sows, SLE showed negative genetic correlations with all input traits (significant for BFL, P < 0.05), indicating that sows with genetically high SLE (efficient sows) tended to eat less and minimized mobilization of body resources; that is, they had lower input components. Trait SLE also had nonsignificant (P > 0.05) negative genetic correlations with output traits for Landrace sows. Thus, SLE is unfavorably correlated (genetically) with both input and output traits, but the correlations were stronger for input traits than for output traits; that is, genetic selection of sows for higher SLE will tend to reduce inputs rather than increase outputs. Bergsma et al. (2008) reported a similar trend in crossbred sows: a significant negative genetic correlation of SLE with input traits and positive (but not significantly different from zero) genetic correlations with output traits.

For both breeds, SRFI showed a strong significant (P < 0.05) positive genetic correlation with SFI (0.95 in Yorkshire and 0.92 in Landrace) and a significant (P < 0.05) negative genetic correlation with BFL (−0.43 in Yorkshire and −0.51 in Landrace; Table 5). Thus, sows with genetically low SRFI (efficient sows) ate less and mobilized more body reserves. We also observed a strong negative genetic correlation between SRFI and SLE in Landrace sows (−0.70).

Estimates of genetic correlations between EB1 and EB2 were positive for both breeds (0.85 in Yorkshire and 0.73 in Landrace, Table 5). Genetic correlations of EB1 and EB2 with traits associated with body energy mobilization (BWL, BFL, and LDL) were negative for both breeds; that is, sows that had a genetically higher net energy balance mobilized fewer body resources during lactation. The latter estimates ranged from −0.60 to −0.91 for Yorkshire sows and from −0.23 to −0.90 for Landrace sows. Traits EB1 and EB2 were also positively correlated (genetically) with SFI (0.92 and 0.85 in Yorkshire and 0.58 and 0.90 in Landrace) and negatively correlated (genetically) with output traits (−0.33 to −0.56 in each breed) and with body resource mobilization traits (−0.60 to −0.91 in Yorkshire and −0.31 to -0.90 in Landrace sows). The strong positive genetic correlations of EB1 and EB2 with SFI and their negative genetic correlations with output and body mobilization traits support our hypothesis that the milk production potential of a sow is strongly associated with its ability to mobilize body resources during lactation and that dietary energy is mostly used to minimize body tissue loss during lactation.

Genetic Correlations with Reproductive Traits.

Among the reproductive traits, SBP and LSB had a strong positive genetic correlation in both breeds (Table 5). A similar trend was observed by Canario et al. (2006) for Large White sows and by Imboonta et al. (2007) for Landrace sows. The genetic correlation between LSB and LSW was also positive for both breeds but had large SE. Chen et al. (2003) also reported a positive genetic correlation between the number born alive and the number weaned for U.S. Yorkshire, Landrace, Hampshire, and Duroc sows.

In both breeds, sow body composition at the time of farrowing had a favorable genetic and phenotypic correlation with SBP (Table 5); that is, sows with genetically high body reserves at farrowing (indicated by BWF, BFF, and LDF) had lower SBP. At the same time, body reserves at farrowing did not have favorable genetic or phenotypic correlations with LBP but had a negative genetic correlation (significant for Yorkshire) with LSB; that is, sows with genetically greater body reserves at farrowing had fewer total born piglets. This suggests that the favorable genetic correlation of body composition at farrowing with SBP might be due to an overall reduction in litter size for sows with high body reserves at farrowing. Bergsma et al. (2008) also reported that genetically high fat mass, protein mass, and BW at the beginning of lactation were associated with reduced total litter size at birth.

As expected, the output traits LWG and EOP were strongly positively correlated (genetically and phenotypically) with LSW in both breeds (Table 5 and Supplemental Table 1). We did not observe significant genetic correlations between efficiency traits and reproductive traits, other than a significant negative genetic correlation of EB2 with LSW in Yorkshire sows (−0.61; Table 5), which indicates that sows that mobilized more body resources weaned larger litters. The genetic correlation of EB2 with LSW was negative but not significantly different from zero in Landrace sows. Similarly, EB1 also had negative genetic correlations with litter size at birth and weaning in both breeds, but estimates were not significantly different from zero because of high SE. Some estimates of genetic correlations between efficiency traits and reproductive traits had opposite signs in both breeds, but none of these correlations were significantly different from zero (P > 0.05).

Impact on Future Reproductive Performance

Genetic correlations of parity 1 lactation traits with parity 2 reproductive traits were estimated to evaluate the influence of lactation performance in parity 1 on reproductive performance in parity 2 (Table 6). Estimates of genetic correlations for all traits across parities are in Supplemental Tables 6 and 7. In both breeds, traits associated with body resource mobilization (BWL and BFL) in parity 1 had negative genetic correlations with reproductive traits (LBP and LSB) in parity 2, but none were significantly different from zero because of large SE, except for the genetic correlation between BFL and LBP in Yorkshire sows (Table 6). These trends suggest that sows with genetically high BWL and BFL in parity 1 may have low LSB and fewer LBP in parity 2. Similar trends were also reported by Eissen et al. (2003), Vinsky et al. (2006), and Schenkel et al. (2010). A lower litter size in parity 2 for such sows may be due to fewer medium-sized follicles or follicles with less follicular fluid at weaning (Clowes et al., 2003). Poor embryonic survival might also be a reason for this reduction of litter size in parity 2 (Willis et al., 2003).


View Full Table | Close Full ViewTable 6.

Estimates of genetic correlations between lactation traits in parity 1 and reproductive traits in parity 21

 
Parity 2 reproduction traits, Yorkshire
Parity 2 reproduction traits, Landrace
Parity 1 lactation traits LBP LSB LSW LBP LSB LSW
BWL −0.28 (0.32) −0.25 (0.22) −0.04 (0.30) −0.02 (0.50) −0.29 (0.40) 0.13 (0.35)
BFL −0.84 (0.47) −0.53 (0.39) −0.33 (0.63) −0.14 (0.34) −0.30 (0.38) 0.21 (0.54)
SFI 0.05 (0.27) −0.01 (0.23) −0.25 (0.28) 0.31 (0.53) 0.42 (0.43) 0.24 (0.36)
SRFI −0.09 (0.30) −0.22 (0.26) −0.16 (0.32) 0.48 (0.60) 0.48 (0.46) 0.29 (0.38)
EB1 0.55 (0.39) 0.30 (0.34) 0.06 (0.47) −0.40 (0.56) −0.11 (0.46) −0.36 (0.41)
EB2 0.34 (0.26) 0.35 (0.22) −0.13 (0.30) 0.29 (0.48) 0.29 (0.39) −0.11 (0.34)
1SE are in parentheses. Bold correlations differ from zero (P < 0.05). BWL = body weight loss; BFL = back fat loss; SFI = total feed intake; SRFI = sow residual feed intake; EB1 = energy balance 1; EB2 = energy balance 2; LBP = live-born piglets; LSB = litter size at birth; LSW = litter size at weaning.

Estimates of the genetic correlations of SFI in parity 1 with reproductive traits in parity 2 were positive for Landrace sows (Table 6), whereas in Yorkshire sows, some of these genetic correlations were negative, but none were significantly different from zero (P > 0.05) in either breed. In a study of 3 breeds, Eissen et al. (2003) reported that the size of the second litter was not influenced by feed intake of the sow during first lactation.

In general, BFL of Yorkshire sows in parity 1 had strong negative genetic correlations with reproductive output traits in parity 2 (−0.84 and −0.53). Corresponding estimates were closer to zero in Landrace sows (−0.14 and −0.30; Table 6). In contrast, SFI in parity 1 had no or negative genetic correlations with reproductive output traits in Yorkshire sows and positive genetic correlations in Landrace sows. This suggests that the mechanism for utilizing body resources may differ between Yorkshire and Landrace sows, which needs further investigation.

Indicator Traits for Economically Relevant Traits

Estimates of genetic correlations of potential indicator traits with ERT suggest that BW at first farrowing and BWL in parity 1 can be used as indicator traits for SFI and SRFI in parities 1 and 2 (Tables 7 and 8). These indicator traits have moderately high heritability (0.26 to 0.41) and are easy to measure. Weighing sows is an easy procedure, and weighing piglets at birth is a standard procedure in most nucleus herds. However, these results must be validated with larger data sets. Back fat loss in parity 1 was also found to be significantly correlated with LWG in parities 1 and 2 (Table 8), but the low heritability of BFL in parity 1 (0.03 and 0.07) prevents its consideration as an indicator trait.


View Full Table | Close Full ViewTable 7.

Estimates of genetic correlations between economically important traits and indicator traits in Yorkshire sows1

 
Parity 1 economically important traits
Parity 2 economically important traits
Indicator traits in parity 1 SFI SRFI LWG SFI SRFI LWG LSB
BWF 0.49 (0.19) 0.56 (0.20) −0.04 (0.24) 0.55 (0.22) 0.48 (0.22) −0.08 (0.30) 0.06 (0.24)
BFF −0.21 (0.19) −0.12 (0.21) −0.01 (0.22) 0.04 (0.23) 0.07 (0.23) 0.51 (0.24) 0.01 (0.22)
LDF −0.16 (0.19) −0.14 (0.20) −0.67 (0.17) 0.04 (0.23) −0.05 (0.22) −0.01 (0.27) −0.11 (0.22)
BWL −0.80 (0.11) −0.85 (0.09) 0.22 (0.21) −0.42 (0.20) −0.15 (0.22) 0.39 (0.24) −0.25 (0.22)
BFL −0.86 (0.53) −0.81 (0.52) 0.44 (0.36) −0.90 (0.75) −0.35 (0.49) 0.74 (0.41) −0.53 (0.39)
LDL −0.49 (0.25) −0.23 (0.30) 0.87 (0.32) −0.19 (0.32) −0.16 (0.33) 0.76 (0.51) −0.76 (0.35)
1SE are in parentheses. Bold correlations differ from zero (P < 0.05). BWF = body weight at farrowing; BFF = back fat at farrowing; LDF = loin depth at farrowing; BWL = body weight loss; BFL = back fat loss; LDL = loin depth loss; SFI = total feed intake; LWG = litter weight gain; SRFI = sow residual feed intake; LSB = litter size at birth.

View Full Table | Close Full ViewTable 8.

Estimates of genetic correlations between economically relevant traits and indicator traits in Landrace sows1

 
Parity 1 economically relevant traits
Parity 2 economically relevant traits2
Indicator traits in parity 1 SFI SRFI LWG SFI SRFI LWG LSB
BWF 0.43 (0.18) 0.41 (0.19) 0.56 (0.24) 0.46 (0.27) 0.14 (0.28) 0.20 (0.28) −0.46 (0.42)
BFF −0.23 (0.16) −0.29 (0.16) 0.42 (0.23) 0.16 (0.25) −0.60 (0.22) −0.22 (0.24) −0.36 (0.32)
LDF −0.01 (0.23) −0.24 (0.23) 0.07 (0.32) NE NE −0.36 (0.30) 0.28 (0.48)
BWL −0.74 (0.10) −0.75 (0.11) 0.38 (0.26) −0.86 (0.24) −0.73 (0.28) 0.07 (0.27) −0.29 (0.40)
BFL −0.66 (0.32) −0.28 (0.35) 0.88 (0.34) NE NE 0.91 (0.48) −0.30 (0.71)
LDL −0.04 (0.35) 0.33 (0.38) 0.47 (0.45) −0.69 (0.52) −0.29 (0.50) 0.12 (0.49) NE
1SE are in parentheses. Bold correlations differ from zero (P < 0.05). BWF = body weight at farrowing; BFF = back fat at farrowing; LDF = loin depth at farrowing; BWL = body weight loss; BFL = back fat loss; LDL = loin depth loss; SFI = total feed intake; LWG = litter weight gain; SRFI = sow residual feed intake; LSB = litter size at birth.
2NE = convergence not attained, and hence, estimates could not be obtained.

Inclusion of Lactation Traits in Breeding Programs

The usefulness of including traits associated with sow lactation in a breeding program depends on economic values, trait heritabilities, and genetic relationships with other traits of interest. The results from this study show that the traits associated with lactation can be considered to be included in the breeding goal and index. Using simulation, Bergsma et al. (2008) reported that including SLE in the breeding goal and as a selection index trait resulted in an increase in SLE, a decrease in SFI, and an increase in sow body resource mobilization during lactation. The sows also became leaner at the beginning of lactation because of reduction in body fat mass, and the negative energy balance during lactation became more severe because of excess mobilization of resources from the body. For this simulation, Bergsma et al. (2008) assumed the economic values of SLE to be equal to that of total born piglets. The results from our study also show that SLE was unfavorably correlated (genetically) with both input and output traits during lactation. The correlations were stronger with input traits than with output traits; that is, a genetic selection on SLE will reduce inputs rather than increasing outputs. We expect the results from inclusion of sow lactation traits in breeding programs to be similar to those reported by Bergsma et al. (2008). The unfavorable correlated responses observed in both input and output traits can, however, have negative effects on the long-term reproductive performance of sows. An alternative for this will be to design breeding programs that yield genetic progress in some of the economically important component traits, rather than SLE itself. Among the component traits, SFI and LWG (weaning weight) are traits with the largest economic impact and can be considered for inclusion in the breeding goal.

A breeding goal for dam lines with emphasis on total number born, survival at birth and at weaning, weaning weight of the piglet, age of puberty of sow, ADFI of the sow during lactation, weaning to estrus interval, and sow longevity was developed to predict the consequences of including sow lactation traits in a modern breeding program (results not shown). The relative contributions of ADFI of the sow during lactation and average weaning weight of a piglet to the selection index when the economic weights were weighted by the respective genetic SD were 1% and 3%, respectively. A reduction in SFI will result in excess mobilization of body reserves during lactation, and the sow will compensate for this loss in the following cycle by increasing feed intake during gestation. This is the reason for the low economic weight for ADFI of sows during lactation. No literature was available to directly compare our estimate of economic values for ADFI of sows during lactation. The relative economic weight of piglet weaning weight was less than estimated by Quinton et al. (2006; 5.7%) for a breeding goal developed for the feeder pig market, which did not include sow ADFI or longevity in the breeding goal.

Including SLE or component traits in the index will result in increasing the accuracies of the index, but the cost of the breeding program itself can increase because this involves additional data collection. As discussed above, an alternative is to use indicator traits for ERT or breeding goal traits in the index that are easy to measure and economically less challenging.

Conclusions

This study produced a set of breed-specific genetic parameters for Yorkshire and Landrace sows for lactation traits, estimated across parities, by parity, and between parities. These estimates can be used as a starting point for developing breeding programs that include lactation traits. Results from this study show that, in general, heritabilities for most traits associated with lactation were low to moderate. Milk production in sows was found to depend on the sows’ ability to mobilize body resources and was found to be genetically uncorrelated with feed intake during lactation, which is consistent with the evolutionary advantage of lactation. Larger studies involving lactation traits, reproductive traits, and growth traits, along with modeling of breeding programs, are required to extensively evaluate the effect of incorporating lactation traits in breeding programs.

 

References

Footnotes



Files:

Comments
Be the first to comment.



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