2010
Genetic Improvement of Sow Longevity
By L. Engblom, K. Stalder, M. Nikkilä and J. Mabry, Department of Animal Science, Iowa State University, Ames, IA
Sow longevity represents the sow's ability to stay and remain productive at an acceptable level within commercial swine breeding herds. Inferior sow longevity is currently a problem with 50% or more of the sows removed in the US and other countries annually (Engblom et al., 2007; Rodriguez-Zas et al., 2003). Today, the majority of sows in the US commercial herds are removed in their early parities, mainly due to reproductive disorders. This high early removal is likely to reduce production efficiency since litter size typically increases up to parity five, but today fewer than 50% of the US sows produce five litters. Poor sow longevity represents substantial costs to the commercial pork operation. Studies have shown that it takes at least three litters before a sow gives a positive cash flow for the producer, i.e. sows have "paid for themselves" (Lucia et al., 2000; Stalder et al., 2003).
Today, 30% of the US sows are removed before parity 3. High removal rates may also be an indication for compromised animal well-being. Reducing high replacement costs due to high removal and poor longevity are especially important today when the pig production industry is operating on very slim if any profit margins. Increased sow longevity would reduce replacement rates, the costs for replacement gilts and thereby increase the net income. The input cost for a gilt is the same regardless of how many parities she produces. Gilts should therefore be considered as an investment, an investment which should be used as effectively as possible.
Sow longevity is a complex trait and it is determined by many factors. Not only is the sow's physical status important, but also genetics, season, management and housing. In addition, it is the herdsman's subjective decision which also involves cull sow market prices that determines whether a sow is removed or remains in the breeding herd. In this decision, the herdsman considers the sow's parity number, production, reproductive status, health status and herd structure, as well as access to replacement gilts of relevant reproductive status. Today, most removals of sows are unplanned (due to e.g. reproductive disorders and locomotor problems or what is often termed involuntary culling) and early, but if sow longevity would be improved, the proportion of sows getting removed due to old age and low production would increase. This would not only increase productivity, ease planning at farm level but also as stated above increase profitability in piglet production.
How to improve sow longevity?
Improvements in housing and management are likely to increase sow longevity. Selection for more robust animals which can produce well in existing production systems would also improve sow longevity as shown in several studies (Heusing et al., 2005; Serenius et al., 2006; Tarrés et al., 2006). The advantage of genetic improvement is that an improved genetic potential would give a higher upper limit regardless of the environment. Although sow longevity has received more attention during the last decade and it has been shown that selection can improve this trait, it is usually not included in genetic evaluations. This paper will show and discuss results from a couple of new studies where issues related to the inclusion of sow longevity in breeding evaluations have been investigated analyzing data from American swine farms.
The first data set consisted of 12,725 pedigreed crossbred sows (Landrace × Large White) from one farm. All these sows had entered the farm between December 2004 and March 2009. The second data set consisted of data from 57 farms with at least 100 sows with their first litter recorded from 2001 to 2008. The farms had either pure and/ or crossbred sows. The total data set contained records on 27,687 (12,199 Yorkshire, 6137 Landrace, 7188 Yorkshire × Landrace and 2163 Landrace × Yorkshire) sows. The two data sets were analyzed with two statistical programs; DMU6 (Madsen and Jensen, 2008) and GIBBS2CEN (personal communication Tsuruta, S.).
Genetic improvement of sow longevity
Selection for improved sow longevity can be performed either by selecting for the trait itself (direct selection) or for a genetically highly correlated trait (indirect selection). Direct selection is safer, i.e. we are more confident that we select the best animals but genetic progress will be slower since data collection may take long time. Indirect selection can give faster genetic progress but it is a balance with accuracy.
Direct selection for sow longevity can be performed by selecting sows with long life measured in days, high removal parity number or high lifetime production. The present study analyzed lifetime born alive as the main measurement for longevity instead of the more commonly used length of productive life (measured in days). The reason for using lifetime born alive is that it is the trait having economic importance for piglet production. Additionally, when sow longevity or productive lifetime is measured as herd life or true lifetime, represented by the number of days in the herd, it will include non-productive days which we do not want to select for since these are undesirable from a productivity standpoint.
The problem with direct selection for sow longevity is that it takes a long time interval (several years) to collect complete lifetime data. Long data collection is likely to result in increased generation interval which will slow down genetic progress. One way to handle this problem is to analyze longevity data with statistical methods which have the ability to include incomplete (censored) data. This means that also sows that still are alive can contribute information which is included in the analyses. We have investigated how sires of crossbred sows are ranked when longevity data are analyzed with programs accounting for censoring in comparison to the ranking based on analyses with a program that ignores censoring and either included removed sows only or all data (acting like all sows have complete records). The results show that it is better to include all data in the analyses even if the analyses do not account for censoring since the sires are ranked almost identically with both of these methods. These results, however, are based on one population only and further analyses utilizing data from several populations are needed before this methodology can be widely implemented.
The second way to select for sow longevity is by indirect selection, which means by selecting for a trait which can be measured earlier in life one would expect improvement in sow longevity too. Strong genetic correlation is required between the two traits. The advantage is that data can be collected earlier and the disadvantage is that the accuracy can be lower. Examples of possible indicator traits are reproductive, conformation or stayability traits (if a sow remained at the farm up to a certain parity or not). The present study investigated stayability traits and accumulated litter sizes from parity 1 to parity 4 as indicator traits. To determine how well these traits actually predict true lifetime productivity, it was examined how similarly they ranked sires in comparison to full lifetime data. The trait was considered to be a good indicator trait if it ranked sires similarly to lifetime born alive. Our results show that stayability in parity 1 ranks sires more similar to lifetime born alive than litter size in parity 1. But ranking sires on parity 2 to 4 accumulated born alive was more similar to lifetime born alive ranking than stayability to second, third and fourth parities, respectively. For both accumulated born alive and stayability, sire ranking became more similar to lifetime born alive ranking with each additional parity. The largest improvement was between the first and second parities followed by the increase between parity two and three. Based on the results from this study, accumulated born alive to parity three, or perhaps two may be useful as early indicators for lifetime born alive and could be considered for inclusion into breeding evaluations.
Sow longevity at different levels in the breeding pyramid
We have investigated the association of the "same" traits between purebred sows and crossbred sows in 57 American swine farms. The results show that among traits compared between pure- and crossbred sows, traits measured early in life are quite similar at the two levels. This means that selection for age at first farrowing and litter size in parity 1 performed in nucleus farms is likely to result in improvement in commercial crossbred sows also. But since longevity traits had lower genetic correlation between the two levels, it is not certain that selection for longevity trait in nucleus farms would result in genetic improvements among crossbred sows in commercial farms. This problem can, however, be solved by including data from crossbred sows in commercial farms to the breeding evaluation. To include crossbred data is often logistically difficult. For example, commercial farms do not typically keep as accurate records as needed in a breeding evaluation, including using single sire matings so that F1 females or commercial line females would have known parents. However, the improvement would occur based on the fact that animals would be ranked based on how their offspring would perform in the commercial farms, which is ultimately the breeding goal for any genetic supplier that is implementing an effective genetic improvement system.
Conclusions
Sow longevity needs to be improved and selecting for it to increase the genetic potential regardless of the environment is an effective way along with improvements in management. The goal is to identify sires which improve lifetime born alive among crossbred sows. The difficulties associated with breeding for sow longevity can be solved by analysing the data with different analysis methods and by including data from commercial farms also.
References
- Engblom, L., N. Lundeheim, A.-M. Dalin, and K. Andersson. 2007. Sow removal in Swedish commercial herds. Livestock Science 106: 76-86.
- Heusing, M., H. Hamann, and O. Distl. 2005. Genetic analysis of lifetime performance and fertility traits in the pig breeds large white, German landrace and pietrain. Züchtungskunde 77: 15-34.
- Lucia, T., G. D. Dial, and W. E. Marsh. 2000. Lifetime reproductive and financial performance of female swine. Journal of the American Veterinary Medical Association 216: 1802-1809.
- Madsen, P., and J. Jensen. 2008. A user's guide to DMU. A package for analysing multivariate mixed models, Danish Institute of Agricultural Sciences, Tjele, Denmark.
- Rodriguez-Zas, S. L. et al. 2003. Bioeconomic evaluation of sow longevity and profitability. J Anim Sci 81: 2915-2922.
- Serenius, T., K. J. Stalder, and M. Puonti. 2006. Impact of dominance effects on sow longevity. J Anim Breed Genet 123: 355-361.
- Stalder, K. J., R. C. Lacy, T. L. Cross, and G. E. Conatser. 2003. Financial impact of average parity of culled females in a breed-to-wean swine operation using replacement gilt net present value analysis. Journal of Swine Health and Production 11: 69-74
- Tarrés, J., J. P. Bidanel, A. Hofer, and V. Ducrocq. 2006. Analysis of longevity and exterior traits on large white sows in Switzerland. J Anim Sci 84: 2914-2924.
- Tsuruta, S., and I. Misztal. 2006. Thrgibbs1f90 for estimation of variance components with threshold-linear models. In: 8th WCGALP, Belo Horizonte, MG, Brasil.