Development of a Recording and Evaluation System for Conformation Traits

Pramod K. Mathur
Canadian Center for Swine Improvement (CCSI),
Central Experimental Farm, Ottawa K1A 0C6, Canada

1. Introduction

Physical soundness of pigs plays an important role in commercial production systems. Animals that fail to maintain minimum levels of physical soundness need to be culled early from the herd resulting in economic losses.

In many herds, feet and leg weakness is the second most frequent reason for culling of sows after reproductive failure (Eliasson-Seling and Lundeheim, 1996; Wood and Rothschild, 2001) which leads to lower lifetime reproductive performance of breeding animals. An evaluation of the conformation of sows is primarily used to predict the risk of involuntary culling that adversely affects sow longevity. Reducing involuntary culling of sows has several advantages such as a reduction in the annual cost of replacement, an increase in the average number of piglets per litter because of the decreasing number of first parity sows, a reduction in the number of non-reproductive days, and increased opportunities for selection on other traits.

Conformation is also important for growth and performance of commercial hogs. Some feet and leg traits are positively correlated with growth rate. In extreme cases, pigs that have problems in standing or walking cannot perform well in modern rearing conditions.

Selection for conformation is usually done separately from other traits, although selection is most effective when based on an index which weighs each trait according to its relative economic value and its relationships with other traits. As a result, selection for conformation is often overemphasized compared to selection for traits with higher economic value or heritability, leading to a reduction in selection intensity for these traits. This leads to a substantial decrease of overall genetic progress for production traits (more than half in many herds) and a corresponding reduction in overall economic benefits to breeders and producers.

The current phenotypic selection for conformation is also far from optimal in terms of improving conformation itself. Some conformation traits have proven economic value while others have little or even negative value. Subjective assessments of conformation often lack adequate adjustments for the effect of various factors, such as the age of the animal or the effect of rearing methods. In addition, selection is often based exclusively on information from the animal itself, even though inclusion of information from relatives into an EBV would make it much more accurate and effective.

Therefore, there are many good reasons to move to a more objective approach of selecting for conformation. These reasons are just as valid for conformation as they are for production traits.

2. Systems used in different countries

2.1 Traits measured

A number of scoring systems have been proposed and used in many countries, especially in Europe (Table 1). These scoring systems are different in terms of the traits recorded, number of categories per trait and differentiation between classes. Following is a summary of the published systems.

Table 1: Systems used for scoring conformation traits in different countries

Recording of a large number of traits allows more detailed description of an animal’s conformation. However, selection based on many traits at once may result in very little progress in any individual trait. In addition, there is a cost associated with the time spent on recording a large number of traits and the process of genetic evaluation. Therefore, it is important to use a simplified approach with a small number of most important traits.

2.2 Number of classes

The scale of recording is also very important. A classification in three categories (three-point scale) has the advantage that the scoring is clear and easy e.g. the fore leg can be classified either as buckled, sickled or normal and pasterns can be classified as high, low or normal. However, this classification does not allow much differentiation (e.g. if the fore leg is slightly buckled compared to severely buckled). The larger problem is the difficulty in the identification of genetic variability. Hence, a classification in three categories is not very suitable for genetic evaluation. On the other hand, if there too many categories, it is difficult to differentiate individuals between adjacent classes and ultimately few classes end up being used. In the study by Van Steenberg (1989) there were 9 main categories with the option to record intermediate categories. Thus there were 17 possible classes. However, intermediate classes were used less than the classes with whole numbers resulting in the use of 9 main categories out of the total of 17 (Fig.1).

Figure 1: Overall distribution of exterior scores into different categories
(source: Van Steenbergen, 1989)

In this scientific study, the distribution in nine classes appears to be fairly normal. However, under field conditions, a system with nine categories may not be used properly due to the difficulty in differentiating between adjacent categories. A five point scale is probably best in practice and also for genetic evaluations.

2.3 Linear vs. non-linear scoring methods

The scores can be assigned either on the basis of the desirability of the condition or the condition itself. An individual leg trait could be scored on a scale from 1 to 5 in which 5 is the most desirable and 1 is the least desirable. This is a nonlinear system (Schulze et al., 1998). Another, type of scoring system could be based on the severity of the condition on a linear scale e.g. from 1 to 5 where 1 is one extreme and 5 is another extreme while the optimum is often somewhere in the middle, around 3. This is a linear system. For example, according to a linear system, the front view of the legs (leg turning) can be scored on a scale of 1 to 5 where 5 is X–shape with the legs bent extremely outwards while 1 is O-shape with the legs turned extremely inwards. The most desirable score is 3 for straight legs. If a non-linear system is used to record this condition, both X-shape and O-shape would end up in the category 1 being the most undesirable conditions while straight legs would receive a score of 5. Giving the same score to very different, even opposite, conditions is not a desirable option. Therefore, linear systems provide more information and allow recording of the actual variability.

A systematic comparison of the linear and non-linear scoring methods is given by Schulze et al. (1998). In this study, a non-linear scoring system was used for about 20 years to record height, length, muscularity and body type on a scale from 1 to 9 where 9 was the most desirable score. Thereafter, since 1994, an additional linear scoring system was used for six feet and leg traits. The scores ranged from -2 to +2 where 0 was the optimum score. The study recommended the use of a linear scoring system because for each trait much more information about deviations from optimum was obtained, each leg trait represented separate genetic information as indicated by low genetic correlations among them, and each leg trait showed moderate heritability.

Considering genetic improvement as the main goal, it is more appropriate to use a linear system. The EBVs should then be included in a selection index with non-linear economic values depending upon the desirability of the trait.

3. Selection of suitable measures for conformation

In a systematic approach for genetic improvement, the choice of conformation traits should be based on heritability estimates, the economic values of the traits, their correlations with other economically important traits and the accuracy of the measurements.

3.1 Heritability estimates

Genetic parameters for conformation traits are not so common in the literature as those for growth and production traits. However, some of the studies have been very extensive in terms of the number of traits recorded, the number of classes within each trait and the number of records used. Estimates reported in some studies are given by Larochelle (1999). Following is a summary of heritability estimates for important conformation traits from various sources.

Table 2: Heritability estimates for important conformation traits.

Here, the averages for individual traits should be interpreted with caution because the estimates are based on different scoring and production systems. For example, fore leg bones in the study by Grindflek and Sehested (1996) in Norway are scored on a three point scale and include males only, while in the study by Van Steenbergen (1989) in Sweden they are scored on a nine point (or 17 point) scale and include both sexes.

In general, among feet and leg traits, pasterns seem to have the highest heritability. For most other traits, the heritability is around 15%. Locomotion has a lower heritability probably because of the difficulty in recording it accurately.

3.2 Genetic correlations of conformation traits with production traits and longevity

Genetic correlations between conformation and production traits have been evaluated in a long-term study in Holland as a part of a Ph.D. dissertation (Van Steenbergen, 1990). These estimates are probably the most reliable because of the large number of observations and the number of classes used. The estimates for traits included in the current selection index are given below.

Table 3: Genetic correlation between conformation and production traits

These correlations suggest that selection on the index traits can significantly affect conformation and vice versa. However, the magnitude and directions of these correlations should be interpreted in relation to the scales of the measurements. The low genetic correlations may also indicate strong nonlinear relationships. Nonlinear relationships are especially possible because both high scores and low scores are undesirable, for most of these traits. A positive correlation does not necessarily mean an improvement in the production trait with an improvement in conformation.

The sign of the correlation and the population mean for the leg traits together determine if selection for production traits is associated with better conformation or vice versa. This situation makes it rather difficult to draw meaningful conclusions based on these genetic correlations. There is evidence of a negative phenotypic correlation between good conformation and faster growth rate based on the frequencies of the conformation traits in fast and slow growing animals. Grindflek and Sehested (1996) found high frequencies of weak fore leg and hind leg pasterns, uneven and small/narrow hind leg claws, X-shaped fore and hind legs, dipped back and bad locomotion in fast growing animals. This negative phenotypic correlation may result from the fact that very high rates of growth have a negative effect on conformation. One factor may be a disproportionate growth of muscles compared to tendons. Selecting for better conformation may reduce this problem rather than aggravate it.

Van Steenbergen (1990) conducted a detailed investigation on the relationship between exterior traits and longevity. Over a two-year period, over 5000 gilts and sows from different strains (Landrace, Yorkshire, Duroc and their crosses) were judged for 20 exterior traits at 24 commercial multiplier herds. Farms were visited monthly by an inspector who judged all gilts of approximately 7-9 months of age and all sows from weaning till 40 days post weaning. Sows culled for leg weakness before the fourth parity were longer and broader at the gilt stage, had less straight rear legs and more tubercles at the rear legs, walked slower and twisted more with hind quarters than animals that had produced at least four litters. In this investigation, estimates of genetic correlations with longevity were not calculated. Research results from a similar study, however, suggest that longevity is strongly influenced by exterior traits (Grindflek and Sehested, 1996). Poor locomotion and straight pasterns lead to decreased longevity. However, contrary to the common belief, they found that weak pasterns have a positive effect on longevity. According to these studies, length, width, hock joints, pasterns and locomotion are important conformation traits with significant effects on longevity.

3.3 Economic values

There is a scarcity of information about the direct economic value of conformation traits in the published literature. In most cases, the economic values are estimated based on relationships between individual conformation traits and longevity. Similar procedures are followed in other livestock species. Therefore, it can be assumed that the traits with the highest economic values are those that have the highest effect on longevity, as described above.

3.4 Accuracy of measurements

Among the important conformation traits that are related to growth and longevity, the traits of fore legs and hind legs can be recorded clearly and easily at the time of probing. The average estimated time for recording fore and hind legs is about 30 seconds/pig (Andersen and Hansen, 1996). However, locomotion is more difficult to measure accurately. It requires a calm animal walking on an even surface for some time. Therefore, it is more time consuming and difficult to record. Locomotion may be replaced by scores for fore and hind legs that can be recorded more easily and accurately.

4. Proposed system of recording

A national system for recording conformation traits in Canada has been developed by a working group including representatives from Regional Centres and the University of Guelph. This system includes scores for fore legs, hind legs and underline.

4.1 Feet and legs

The legs are scored on a scale of 1 to 5 in three ways: from in front or behind the pigs, from the side, and for the condition of the pasterns as given in Table 4. During the testing and evaluation of the proposed system it was felt that some pigs should not be approved for breeding despite satisfactory feet and leg scores. There should be some way of recording the overall opinion of breeders and technicians. Therefore, an additional measure “overall rating” was included to allow breeders and technicians to record their overall opinion about the individual pig and to avoid modifying the feet and leg scores to disqualify the pigs they do not like.

Table 4: Proposed system of scoring feet and leg traits

4.2 Number of functional teats

Scoring of underline is based on the number of functional teats. Functional teats are those from which milk could be obtained at farrowing. However, the number of functional teats needs to be recorded well before farrowing for selection of potential sows for breeding. An ideal time for a clear differentiation between functional and non-functional teats is when the pigs are closer to probing weight or are above 70 kg (Labroue et. al., 2001). At this stage, the teats can be classified into five categories: 1) normal teats, which have a normal size and teat canal, 2) inverted teats which often appear like normal teats with a teat canal, except that the tip is turned inwards with various degrees of inversion, from partial to complete, 3) small teats which may appear functional but are substantially smaller than normal teats, 4) supernumerary or rudimentary teats which are small and blind teats often found in between normal teats, and 5) blind teats, which are teats without a milk canal, and often tend to be small

An interesting observation from the study by Labroue et. al., 2001 was that almost all of the inverted teats become functional and give milk at the time of lactation. Therefore, both normal and inverted teats should be considered as functional teats. The small, supernumerary and blind teats should be considered as non-functional and should be excluded from the counts.

4.3 Time of recording

The feet and leg traits should be recorded on boars and gilts at the time of probing. The number of functional teats can be recorded at any time from weaning to probing and the data can be processed either with pre-listing or probe records. It is best to record the information on all pigs, including pigs that will not be probed, in order to identify sires that have a high percentage of progeny that are culled because of poor conformation.

4.4 National standards

Recording can be done either by breeders or by technicians. In either case, it is very important to ensure the system is accurate and credible. This requires an objective accreditation system just like for recording backfat and lean depth. The proposed scoring system has been tested at the national and regional standard sessions in Eastern Canada, Ontario and Western Canada. At these sessions, each technician scored at least 25 pigs. The same group of pigs is scored again some time later. The scores are then analyzed for consistency and accuracy. Through this process, standard procedures are being developed for training and accreditation of technicians in all regions.

5. Genetic evaluation and selection

Some herds have started collecting data using the proposed system. Data from one herd were analyzed to begin development of a genetic evaluation system. The data included records from 104 Landarce and 74 Yorkshire pigs that were the progeny of 18 sires and 27 dams. The analysis included distribution of scores and estimation of differences between breeds, sexes, sires and dams using simple linear models excluding genetic effects.

The amount of data is very small for a conclusive analysis. However, there were some interesting observations. There were no apparent abnormalities in the distribution of the scores. It is likely that some herds may not have the full range of scores for each trait and the variances within some groups can be small (Table 5). In this data set, there were no significant differences between sexes or between breeds. However, there were significant (P < 0.01) differences between sire families suggesting a reasonable level of genetic determination for the traits. Genetic evaluations should therefore be useful to identify sires responsible for very high or very low scores in their progeny. There were significant differences between effects of dams (P < 0.01) as well. These could be due to genetic differences among the dams as well as to the common environmental conditions for the litter. These preliminary results suggest that it is important to investigate these effects in further detail.

Once sufficient data from the proposed scoring system are available, more steps will be taken to develop genetic evaluations for major conformation traits. The process will include standardization of conformation scores based on variances within contemporary groups and single trait or multiple trait evaluations of conformation scores. The resulting genetic evaluations will be incorporated in selection indices to facilitate joint selection for production, reproduction, feed efficiency and conformation. Genetic trends for conformation will be calculated and extension information will be made available to the industry.

Breeders and producers on the Canadian Swine Improvement Program are therefore looking forward to more effective ways to select for conformation traits and incorporate them in their selection program. They will soon be able to produce pigs that are not only genetically superior for performance traits but also for physical soundness. A major benefit will come from commercial sows with higher longevity.

6. References

Andersen, S. and Hensen, U. G. (1996) Selection for conformation and longevity in Danish breeding systems. Proceedings of NJF-Seminar no. 265, Denmark. 27-28 March 1996: 72-76
Eliasson-Selling.L. and Lundeheim, N. (1996) Longevity of sows in Swedish sow pools. Proceedings of NJF-Seminar no. 265, Denmark. 27-28 March 1996
Grindflek, E. and Sehested, E. (1996) Conformation and longevity in Norwegian pigs. Proceedings of NJF-Seminar no. 265, Denmark. 27-28 March 1996: 77-83
Labroue, F., A. Caugant, B. Ligonesche and D. Gaudré (2001) Evolution of “abnormal” teats in gilts during growth and first lactation. Journées Rech. Porcine en France,
33: 143-150
Larochelle, M. (1999) Selection for conformation traits, Review of literature. Centre de developpement du porc du Quebec inc. (CDPQ), 2795 boul. Laurier, Bureau 340
Sainte-Foy, Quebec G1V 4M7, Canada.
Lundeheim, N. (1996) Conformation scoring in the Swedish pig progeny testing scheme. Proceedings of NJF-Seminar no. 265, Denmark. 27-28 March 1996: 70-71
Schulze, V., R. Rohe, H. Looft and E. Kalm (1998) Genetic improvement of exterior and leg soundness in swine. Züchungskunde 70 (1) 43-60.
Van Steenbergen, E.J. (1989) Description and evaluation of a linear scoring system for exterior traits in pigs. Livest. Prod. Sci. 23:163-181.
Van Steenbergen, E.J. (1990) Relevance of exterior appraisal in pigs breeding. Doctoral thesis, Department of Animal Breeding, Wageningen Agricultural University, The Netherlands.
Webb, A. J., W. S. Russell and D.I. Sales (1983) Genetics of leg weakness in performance tested boars. Anim. Prod. 36:117-130.
Wood, C.M. and M.F. Rothschild (2001) Feet and leg soundness: In Swine Industry Handbook, PIH-101, Purdue University Cooperative Extension Service, West Lafayette, Indiana, US.