HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

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

Genetics of length of productive life and lifetime prolificacy in the Finnish Landrace and Large White pig populations¹

T. Serenius^*,2 and K. J. Stalder^,3

^* MTT Agrifood Research Finland, Animal Production Research, Animal Breeding, Jokioinen, Finland; and and Department of Animal Science, Iowa State University, Ames 50011-3150

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
The objective of this study was to estimate direct and indirect selection potential for length of productive life and lifetime prolificacy in Finnish Large White and Landrace swine populations. To study the direct selection potential, the heritabilities of these traits were estimated. The genetic correlations of length of productive life and lifetime prolificacy with prolificacy traits and overall leg conformation were estimated to evaluate whether selection for these traits could indirectly improve measures of sow longevity. In addition, correlations between length of productive life, lifetime prolificacy, ADG, and backfat thickness were estimated. Records were used from Finnish purebred Landrace (n = 26,744) and Large White (n = 24,007) sows born on operations that perform on-farm production tests on all females. Heritabilities were estimated using both a survival analysis procedure and a linear model. Due to computational limitations, correlations were estimated with the linear model only. Estimated length of productive life heritabilities obtained from linear model analyses were less (0.05 to 0.10) than those obtained from survival analyses (0.16 to 0.19). This may be indicative of the superiority of survival analysis compared with linear model analysis methods when evaluating longevity or similar types of data. All the prolificacy traits were genetically correlated with length of productive life and lifetime prolificacy, and the correlations were greater than 0.13. These results indicate that selection for increased number of piglets weaned in the first litter and for short first farrowing interval is beneficial for sow longevity and also for sow’s lifetime prolificacy. The genetic correlations between length of productive life and leg conformation score also were favorable (0.32 in Landrace and 0.17 in Large White). The heritability estimates indicate that survival analysis is likely the most appropriate method of evaluating longevity traits in swine. Because of computational problems, simultaneous analysis of linear traits and longevity is not currently possible. More research is needed to develop methods for multiple linear and survival trait analyses.

Key Words: Genetic Correlation • Heritability • Lifetime Prolificacy • Longevity • Sow

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Prolificacy traits and longevity play an important role in efficient piglet production. Lifetime prolificacy (LTP) is related to longevity of the sow (the greater the number of parities the sow remains in the herd, the greater the likelihood of an increased number of piglets it will produce during its productive lifetime). Higher replacement rates due to poor longevity increase the number of replacement gilts needed and the associated expenses related to purchase or raising of those gilts.

Sow longevity can be defined as "stayability" or length of productive herd life (LPL). Stayability is a binary trait, measuring whether a sow has survived in a herd until some defined fixed parity or time. Similarly, LPL is the number of days between a beginning event, such as date of birth or date of first farrowing and culling of a sow. These traits continue to be breeding objectives for many seedstock suppliers; however, indirect selection for LPL and LTP has been traditionally employed in most breeding programs because productive life can be recorded only after a sow has been culled or death occurs. Indirect selection for this type of trait may be more useful because it can be carried out with a shorter generation interval.

To compare the efficiency of direct and indirect selection for sow longevity, the genetic parameters (heritabilities and genetic correlations) should be known. The effectiveness of indirect and direct selection for swine longevity traits is likely population-dependent and should be evaluated before selection is actually implemented. Information regarding the effectiveness of direct and indirect selection potential for sow longevity is sparse, particularly for the Finnish Landrace and Large White populations. Genetic correlations between sow longevity and other economically important swine traits are needed to determine the importance of their association.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Data
Data obtained from the Finnish Animal Breeding Association (Vantaa, Finland) were used to estimate the genetic parameters for LPL and LTP. Additionally, ADG (0 to 100 kg), body composition (backfat thickness) measured with ultrasound, age at first farrowing (AFF), first farrowing interval (FFI), number of weaned piglets at first farrowing (NW), and leg score data were captured to estimate genetic and phenotypic correlations of these traits with LPL and LTP. Records were used from purebred Landrace and Large White sows born between 1989 and 2001 on operations that perform on-farm production tests on all sows (Landrace sows were from 493 and Large White sows were from 437 operations). There were 26,744 Landrace and 24,007 Large White sows with at least one reproductive record in the data, and 32% of these records were censored (sow sold or still alive) from both datasets (Table 1). The sows were daughters of 1,366 and 1,221 sires in the Landrace and Large White, respectively. When all the informative relatives were used in the analysis, there were 1,686 Landrace and 1,576 Large White sires in the pedigree.

Statistical Analyses
In the analysis of survival data, the main challenge is to use information on animals that are still alive, and to model factors that are time-dependent (e.g., the effect of farm or some disease). Ducrocq and Sölkner (1998) developed the Survival Kit to analyze this type of data. Using proportional hazard models, it is possible to account for censored records, to model time-dependent factors, and to fit the distribution of longevity data (Ducrocq and Sölkner, 1998); however, it is possible to run only single-trait analysis with the Survival Kit. Therefore, a single-trait proportional hazard model and a multitrait linear model were fitted to the current data. The single-trait analyses were carried out with the Survival Kit (Ducrocq and Sölkner, 2001), and multitrait analyses with DMU package (Madsen and Jensen, 2000).

In survival analysis, the hazard function of a sow’s LPL, t days after first farrowing, was defined as:

where (t)^–1 is the Weibull baseline hazard function with location () and shape () parameters, ß is a vector of fixed and random effects, and x(t) is the corresponding incidence matrix. The suitability of the Weibull distribution was assessed by plotting ln{–ln[S(t)]} against ln(t), where S(t) is the Kaplan-Maier survivor function (Figure 1). The plot of this relationship produced a relatively straight line, which indicates that the Weibull distribution fits the data very well. Moreover, the lines of different stratums produced from the data used in this study were approximately parallel; hence, one baseline hazard function was assumed over the data.

View larger version (20K): [in this window] [in a new window]   Figure 1. Survivor functions, ln{–ln[S(t)]}, plotted against natural logarithm of days from first farrowing, ln(t), for various stratums (number of piglets weaned in first parity = NW; overall leg action score = Score). A straight line indicates that the Weibull distribution fits the data well, and parallel lines of different stratums indicate that one baseline hazard function over the data can be assumed.

where ²/6 is the variance of extreme value distribution.

In multitrait analysis, the following linear sire model, in matrix notation, was fitted for all of the traits studied using DMU-package (Madsen and Jensen, 2000):

where y is the vector of observations of the seven traits considered simultaneously (LPL, LTP, NW, AFF, FFI, ADG, and either backfat thickness or leg score), X and Z are the incidence matrices for fixed (b) and random sire (u) effects, and e is the vector of residuals. The censored records (sows sold or still alive) of LPL and LTP were treated as missing. The effect of farm and year interaction was included in vector b for all the traits studied. In addition, farrowing month (NW, AFF, ADG, backfat thickness, leg score), mating type (NW), age of litter at weaning (NW, FFI), and breeding consultant (ADG, fat percent, leg score) were the other effects included in vector b. Moreover, the fixed regression of test weight was included in the statistical models for ADG and fat percent. The (co)variance structure of random effects of u, and e were assumed to be var(u) = A * G₀, and var(e) = I * R₀, where A is additive relationship matrix among sires, I is the identity matrix, and G₀ and R₀ are the variance-covariance matrices for additive genetic sire and residual effects, respectively.

Two types of "genetic" correlations between LPL and the other traits studied are presented. From multitrait analysis, the genetic correlations were obtained directly. In addition, the correlations between estimated breeding values for LPL in survival analysis and the breeding values of other traits studied in linear model analysis were estimated.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Heritabilities
Three different heritability estimates for LPL for Finnish Landrace and Large White breeds are presented in Table 2. In general, all the estimates were very similar between the breeds. The heritabilities obtained from linear model analysis (0.05 in Landrace and 0.10 in Large White) were lower than those obtained from the survival analysis (0.16 to 0.17 in Landrace and 0.17 to 0.19 in Large White). In addition, the difference between the breeds was greater in the linear model analyses compared with those obtained from the survival analyses. In linear model analysis, the estimated heritabilities for LTP (0.09 in Landrace and 0.12 in Large White) were slightly higher than those for LPL (Tables 3 and 4).

View this table: [in this window] [in a new window]   Table 2. Additive sire variance () and heritability (h2) estimates for length of productive life (LPL) obtained from survival analysis and multi-trait linear model analysis in the Finnish Landrace and Large White populations

Correlations Between Breeding Values
Correlations of the LPL breeding values obtained from survival analysis with the other traits studied are presented in Tables 3 and 4. Although the correlations among LPL breeding values estimated using linear model and survival analysis were moderate or high (Pearson correlations ranged between 0.40 and 0.72, and Spearman rank correlations similarly ranged between 0.38 and 0.72), it should be noted that differences in sire ranking between these methods occurred. The correlations were higher in the Large White (Pearson correlations = 0.52 to 0.72; Spearman rank correlations = 0.48 to 0.72) compared with those from the Landrace (Pearson correlations = 0.40 to 0.57; Spearman rank correlations = 0.38 to 0.55) analysis. Similarly, survival analysis breeding values for true LPL (Spearman rank correlations was 0.55 in Landrace and 0.72 in Large White) had higher correlation with linear model breeding values than those for functional LPL (the corresponding correlations being 0.38 and 0.48).

In general, the correlations between the estimated breeding values were lower than the corresponding genetic correlations obtained from a multiple trait analysis (Tables 3 and 4). However, most of the estimates had the same sign (12 out of 14 estimates), and the tendency for changes in the magnitude of correlation estimates were similar between the two methods.

Phenotypic and Genetic Correlations
Genetic and phenotypic correlations are presented in Tables 3 and 4. In general, phenotypic correlations were very similar between the breeds, whereas some genetic correlations differed between breeds. Moreover, the phenotypic correlations were commonly very low. However, there was an indication that LPL and LTP are very closely associated, as both phenotypic and genetic correlations were greater than 0.90. Because of that, the correlations with the other traits studied are very similar between LPL and LTP.

All the prolificacy traits were genetically correlated with LPL and LTP and the correlations were generally greater than 0.13 (Tables 3 and 4). This indicates that selection for more piglets weaned in the first litter and for short first farrowing interval will have a beneficial indirect effect for LTP and for LPL. The absolute values of these genetic correlations ranged between 0.30 and 0.54 and were similar between the Landrace and Large White populations; however, genetic correlation differences for AFF with LTP and LPL were found between the breeds. In Landrace, the correlations were positive (0.17 with PLP and 0.13 with LTP), whereas they were negative in Large White (–0.28 with LPL and –0.29 with LTP).

In Landrace, the genetic correlations between overall leg score and longevity (0.32 with LPL and 0.28 with LTP; Table 3) were moderate but positive. The corresponding correlations in Large White were 0.17 and 0.19 (Table 4). Although the correlations of overall leg action with LPL and LTP were low to moderate, it may be said that selection for leg conformation, measured when the sow has reached 100 kg, is beneficial for improving sow longevity in an indirect manner.

Among the other traits studied, there were moderate genetic and phenotypic correlations between ADG and backfat thickness in both breeds (r_g = 0.32 and r_p = 0.40 in Landrace and r_g = 0.39 and r_p = 0.40 in Large White; Tables 3 and 4). There were genetic correlations between AFF and FFI (average over breeds = 0.40), between AFF and ADG (–0.40), and between backfat thickness and FFI (–0.29). In addition, there were a few moderate genetic correlations present only in the Landrace breed (NW and daily gain = 0.27; AFF and backfat = –0.18; AFF and Score = –0.37; Tables 3 and 4).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
The objective of the current study was to examine the potential for direct and indirect selection for LPL and LTP. To accomplish this, the heritabilities of these traits were estimated. Heritabilities were estimated both with a survival analysis procedure and with a linear model. The genetic correlations of LPL and LTP with prolificacy traits and overall leg conformation were estimated to determine whether selection for these traits could indirectly improve measures of sow longevity. In addition, correlations of LPL and LTP with ADG and backfat thickness were estimated to determine whether production traits are genetically and/or phenotypically correlated with LPL and LTP. The correlation analysis was carried out with multitrait linear sire model and, and thus the correlations between prolificacy traits, leg conformation, and production traits were also obtained.

Heritability Estimates for LPL
The heritability estimates for LPL were lower for the linear model (0.05 to 0.10) than for the survival analysis (0.16 to 0.19) in the current study. This is in agreement with the estimates presented in the literature (Tholen et al., 1996a,b; López-Serrano et al., 2000), which reported stayability heritability estimates ranging from 0.02 to 0.11 when analyzed using linear models. The heritability estimates have ranged between 0.11 and 0.31 when analyzed using survival analyses (Yazdi et al., 2000a,b). Although heritability estimates from linear model analyses had been generally lower than those from survival analyses, moderate heritabilities estimates were obtained when censored records were accounted in the linear model analyses (Guo et al., 2001). In their analyses, the estimated heritabilities for LPL and LTP ranged between 0.22 and 0.25, which were substantially higher than the current estimates when a linear model was applied and censored records were not considered (Guo et al., 2001).

Current heritability estimates indicate that environmental effects may be modeled more precisely in survival analysis compared with the linear model analysis (i.e., it may be assumed that the higher heritability estimates are due to the possibility to model farm-year effect as time-dependent). On the other hand, it should be remembered that the residual variance was not estimated in the survival analysis, and the heritability estimates were calculated assuming that the residual effects were following extreme value distribution with the variance ²/6. Therefore, one might argue that the different heritability estimates are not comparable.

Correlation Estimates
The results of the current study demonstrate that LPL and LTP are very similar, both genetically and phenotypically (r_g = 0.96 in Landrace and 0.97 in Large White; r_p = 0.94 in both the breeds). Additionally, the correlations of LPL and LTP with the other traits studied were very similar. In general, our correlation estimates between LPL and prolificacy traits are very similar to those reported by Tholen et al. (1996b), except that their genetic correlations between litter size at first parity and stayability were negative in one dataset, ranging from –0.04 to –0.25. However, the same study reported correlation estimates between litter size and stayability ranging from 0.25 to 0.45, which were of the same magnitude as the current estimates between number of piglets weaned and LPL (0.39 in Landrace and 0.30 in Large White). Similarly, the genetic correlations reported by Tholen et al. (1996b) for farrowing interval and stayability ranged between –0.24 and –0.54, compared with current estimates, which ranged between –0.40 and –0.43. Moreover, fixed effects of litter size and farrowing interval also significantly affected LPL in a Swedish pig population (Yazdi et al., 2000a,b).

Although the estimated genetic correlations of LPL and LTP with litter size and farrowing interval were very similar between the breeds, the correlations with AFF differed between the breeds (Tables 3 and 4). In Large White, negative genetic associations between AFF and longevity (–0.28 with LPL and –0.29 with LTP) were found, whereas the signs in corresponding correlations were positive in the Landrace (0.17 and 0.13) analyses. Similarly, the number of piglets weaned was positively correlated (0.21) with AFF in the Large White analysis, whereas the correlation in the Land-race population was approximately zero (Table 3). These results concur with our previously reported studies, where the genetic correlations between AFF and litter size, and piglet survival have been different between the two breeds (Serenius et al., 2004a,b). Thus, the genetic correlation differences associated with AFF among the Finnish Landrace and Large White breeds is likely due to differences in the genetic makeup of the two breeds.

In the current analysis, there was no clear genetic association of LPL and LTP with production traits. The only substantial genetic correlation exists between LPL (and LTP) and backfat thickness in the Large White population (0.22). These findings are in agreement with the literature estimates, where both zero and unfavorable genetic correlations of LPL and LTP with backfat thickness and ADG have been reported. In the study of López-Serrano et al. (2000), stayability was positively correlated with backfat thickness, ranging from 0.11 to 0.27 and negatively associated with ADG (ranging from –0.06 to –0.32). However, Tholen et al. (1996b) reported genetic correlation estimates between stayability and backfat thickness that ranged from –0.03 to 0.36, and between stayability and ADG from 0.02 to –0.13. Thus, genetic correlations among longevity and production traits seem to differ depending on the population from which they are estimated. Additionally, the differences could be the result of variation in the standardized weight at which backfat measures were obtained, back-fat measurement procedures, or other testing methodology.

In Landrace, a moderate genetic correlation (0.32) was found between LPL and leg conformation score. The genetic correlation between the same traits in the Large White breed was approximately half (0.17) that found in the Landrace breed. Similarly, López-Serrano et al. (2000) reported that there was a positive genetic correlation between stayability and leg score in the German Landrace population, and the corresponding correlation was approaching zero in the German Large White population. It may be concluded that leg conformation is genetically correlated with LPL. However, it should be recognized that the reported heritabilities of leg conformation traits are very low (0.06 in the current study); hence, one should not expect large nor rapid genetic improvement in LPL through selection for leg conformation.

Based on heritability estimates, it seems that survival analysis may be the most appropriate method of evaluating swine longevity traits compared with linear models. However, there is one major concern relating to the use of survival analysis in breeding value estimation: because of computational problems, multiple-trait analyses involving longevity and other economically important traits are not currently possible. As stated earlier, LPL and LTP have genetic correlations with litter size, farrowing interval, and leg conformation that are relatively high compared with genetic correlations among longevity and measures of leg soundness.

	Implications

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
The estimated heritabilities for length of productive life and lifetime prolificacy in sows determined from survival analyses were higher than those from linear models. It is likely that survival analysis is a more appropriate method for evaluating longevity traits in swine populations. When determining breeding values for longevity traits, use of survival analyses will provide more accurate estimates; however, simultaneous analysis of longevity and more traditional measures of sow productivity are not currently possible due to computational difficulties. Moderate genetic correlations among length of productive life, lifetime prolificacy, leg conformation, and prolificacy traits indicate the need for a multiple-trait analysis approach to accomplish simultaneous improvement of longevity and other economically important sow productivity traits. Thus, more research is needed to develop methods for multiple-linear and survival-trait analyses.

	Footnotes

 
¹ Funding of the project was supported by Ministry of Agriculture and Forestry of Finland, and ProAgria (Finnish Animal Breeding Association, Finnish AI—cooperatives).

³ Collaborator, via a fellowship under the OECD Cooperative Research Program: Biological Resource Management for Sustainable Agriculture Systems.

² Correspondence: 109 Kildee Hall (phone: 515-294-4103; fax: 515-294-5698; e-mail: timo.serenius{at}mtt.fi).

Received for publication April 24, 2004. Accepted for publication July 16, 2004.

	Literature Cited

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 


Ducrocq, V. P., and J. Sölkner. 1998. Implementation of routine breeding value evaluation for longevity of dairy cows using survival analysis technique. Pages 359–362 in Proc. 6th World Cong. Genet. Appl. Livest. Prod., Armidale, Australia.

Ducrocq, V. 2001. Survival Analysis Applied to Animal Breeding and Epidemiology. Mimeo. Inst. Natl. Rech. Agron., Jouyen-Josas, France.

Ducrocq, V., and J. Sölkner. 2001. The Survival Kit V3.12. User’s Manual. Available: http://www-sgqa.jouy.inra.fr/diffusions.htm. Accessed June 21, 2004.

Guo, S.-F., D. Gianola, R. Rekaya, and T. Short. 2001. Bayesian analysis of lifetime performance and prolificacy in Landrace sows using linear mixed model with censoring. Livest. Prod. Sci. 72:243–252.

López-Serrano, M., R. Reinsch, H. Looft, and E. Kalm. 2000. Genetic correlations of growth, backfat thickness and exterior with stayability in Large White and Landrace sows. Livest. Prod. Sci. 64:121–131.

Madsen, P., and J. Jensen. 2000. A User’s guide to DMU. A package for analysing multivariate mixed models. Mimeo. Danish Inst. Agric. Sci., Tjele, Denmark.

Serenius, T., M.-L. Sevón-Aimonen, A. Kause, E. A. Mäntysaari, and A. Mäki-Tanila. 2004a. Genetic associations of prolificacy with performance, carcass, meat quality and leg conformation traits in the Finnish Landrace and Large White pig populations. J. Anim. Sci. 82:2301–2306.[Abstract/Free Full Text]

Serenius, T., M.-L. Sevón-Aimonen, A. Kause, E. A. Mäntysaari, and A. Mäki-Tanila. 2004b. Selection potential of different prolificacy traits in the Finnish Landrace and Large White populations. Acta Agric. Scand., Sect. A. Anim. Sci. 54:36–43.

Tholen, E., K. L. Bunter, S. Hermesch, and H.-U. Graser. 1996a. The genetic foundation of fitness and reproduction traits in Australian pig populations. 1. Genetic parameters for weaning to conception interval, farrowing interval, and stayability. Aust. J. Agric. Res. 47:1261–1274.

Tholen, E., K., L. Bunter, S. Hermesch, and H.-U. Graser. 1996b. The genetic foundation of fitness and reproduction traits in Australian pig populations. 2. Relationships between weaning to conception interval, farrowing interval, stayability, and other common reproduction and production traits. Aust. J. Agric. Res. 47:1275–1290.

Yazdi, M. H., N. Lundeheim, L. Rydhmer, E. Ringmar-Cederberg, and K. Johansson. 2000a. Survival of Swedish Landrace and Yorkshire sows in relation to osteochondrosis: A genetic study. Anim. Sci. 71:1–9.

Yazdi, M., L. Rydhmer, E. Ringmar-Cederberg, N. Lundeheim, and K. Johansson. 2000b. Genetic study of longevity in Swedish Landrace sows. Livest. Prod. Sci. 63:255–264.

This article has been cited by other articles:

				J. Tarres, J. P. Bidanel, A. Hofer, and V. Ducrocq Analysis of longevity and exterior traits on Large White sows in Switzerland J Anim Sci, November 1, 2006; 84(11): 2914 - 2924. [Abstract] [Full Text] [PDF]

				T. Serenius, K. J. Stalder, T. J. Baas, J. W. Mabry, R. N. Goodwin, R. K. Johnson, O. W. Robison, M. Tokach, and R. K. Miller National Pork Producers Council Maternal Line National Genetic Evaluation Program: A comparison of sow longevity and trait associations with sow longevity J Anim Sci, September 1, 2006; 84(9): 2590 - 2595. [Abstract] [Full Text] [PDF]

				T. Serenius and K. J. Stalder Selection for sow longevity J Anim Sci, April 1, 2006; 84(13_suppl): E166 - E. [Abstract] [Full Text] [PDF]

				J. Arango, I. Misztal, S. Tsuruta, M. Culbertson, and W. Herring Study of codes of disposal at different parities of Large White sows using a linear censored model J Anim Sci, September 1, 2005; 83(9): 2052 - 2057. [Abstract] [Full Text] [PDF]

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