Differences Among Breeds, Breed Origins and Gender for Growth, Carcass Composition and Pork Quality

R.O. Ball, J.P. Gibson, C.A. Aker, K. Nadarajah, B.E. Uttaro and A. Fortin

Background

There are many growth and carcass characteristics of pigs that contribute to the overall profitability of the pork production system. The long-term competitiveness of the pork industry will be strengthened by having optimum combinations of growth, carcass characteristics and meat quality to meet diverse market demands. However, many economically important characteristics relating to carcass composition and meat quality are not recorded in routine genetic improvement programs. Thus, informed choice of optimum breeds is not always possible. Some information on breed characteristics is available from previous trials, in other countries. Breed characteristics are not fixed, however, but can differ among different sources from different parts of the world or even among different breeders within a local area. Differences among breeds and herds can also change over time as breeders pursue diverse selection goals with varying degrees of success. It is therefore important to have complete growth and carcass information on the breeds as they exist for the use of our current local industry. The Ontario Pork Carcass Appraisal Project (OPCAP) was set up to document differences among the four principal pig breeds in Ontario. Breeding stock came from 118 different breeding herds representing 57 different seedstock producers. This provided a unique opportunity to estimate the amount of variation among breeding stocks within a breed. The large data set also allowed analyses of breed by sex interactions. This article summarizes the differences among breeds (Yorkshire, Landrace, Duroc, Hampshire) and sex (boar, gilt, barrow) for the main growth and carcass characteristics and selected measures of pork quality. These differences are discussed in relation to the variation observed among different sources of breeding stock.

Methods

Summaries of the trial protocols and carcass dissection are given in the opening article by C.A. Aker and in Appendices 1 and 2. Statistical methods are detailed in Appendix 3.

Results and Discussion

A large number of differences were examined in this experiment. Because of space limitations, only the key breed and sex effects will be discussed here. Additional results are summarized in Tables 1 and 2 and details of breed by sex interactions are given in Appendix 4.

Differences Among the Breeds

Significant differences were observed among breeds for most of the live animal traits. Of particular interest, Landrace pigs had the fewest days to 100 kg and highest average daily gain, while Hampshire pigs required the most days to 100 kg. Yorkshire pigs had the lowest backfat depth, though this was only about 1 mm less than the fattest breeds, Durocs and Landrace. Carcass yield, as estimated from the grading probe, appears in two ways because the grading system changed in the middle of the experiment. Carcass index is shown with two different analyses; one uses weight as a covariate because slaughter weight differed slightly between breeds and sexes. Differences in estimated carcass yield and carcass index among the breeds were small. Estimated carcass yield is not a particularly good measure of carcass leanness, which is better estimated by dissected lean content of the three primal cuts. Breed differences in lean % in three primals were somewhat larger than for carcass yield, with Yorkshires at 54.3% being well ahead of Hampshires at 53.4%, and Durocs having the lowest yield at 52.5%. Rather different breed rankings were observed for lean content of each of the three primals separately (lean content of shoulder, of loin and of ham %) and this was reflected in the distribution of lean as estimated by shoulder lean, loin lean and ham lean (% of three primals). Durocs had the highest proportion of lean appearing in the ham at 38.9%, while Landrace had the lowest at 37.3%. Conversely, Landrace has the highest proportion of lean appearing in the loin at 31.1%, with Durocs and Hampshires having the lowest at about 30%. The differences among sexes and breeds in dissected lean content of the primal cuts are much larger than the differences in the estimated lean yield equations (old and new) and carcass index. This confirms our previous analyses (not shown) that the current grading system severely underestimates differences between animals in lean yield. Chemical fat in the loin, representing a measure of marbling, was highest for the Duroc pigs. This was confirmed by the higher subjective marbling score for Durocs. A number of measurements were made on the belly because although this is a high value cut it is usually neglected in research. Duroc pigs had the leanest bellies (lowest percent fat), which was confirmed by the ratios of linear measurements on the depth and width of the lean streak in the belly. Higher values for the belly ratios indicate more lean. Meat quality was estimated by drip loss and subjective scores for marbling, colour and structure (see also results on taste panel assessments of meat quality by Fortin in these proceedings). Drip loss, an inverse measure of water holding capacity, was highest for the Hampshire pigs. The high drip loss of Hampshire pigs suggests that the Napole gene is probably at high frequency in the Ontario Hampshire population. Tests for this gene may become available in the near future (see article by Rothschild and Ernst in these proceedings). Duroc pigs consistently had the highest values for marbling score, and colour and structure of the loin, while Landrace scored lowest for colour and structure of loin. However, the Landrace had the highest colour score for ham. Higher values for colour and structure indicate darker and firmer appearance of the pork.

Differences Among Barrows, Gilts and Boars

The typical and well known differences among the sexes for growth performance were confirmed, with boars having a large advantage in terms of backfat, feed efficiency and lean yield. Feed efficiency for barrows and gilts is equal only because the experimental protocol penned barrows and gilts together. Boars consistently outperformed barrows, but were often similar in performance to gilts. Lean content of the primals was highest for the boars. Particularly striking was the large difference of 6.3 percentage units between boars and barrows in percent lean of the loin. Boars also had dramatically leaner bellies than barrows, with gilts being intermediate. There were fewer and smaller differences between the sexes for meat quality measurements than for performance and carcass measurements. The largest difference in the meat quality measurements was a higher marbling score for barrows. Attention should be drawn to the relative differences between the breeds versus the sexes. In most cases, the differences between the highest and lowest sex for any particular trait was much greater than the difference between the highest and lowest breed.

How Much Variation Is There Between Herds Within A Breed?

Table 2 provides a comparison of the variation that exists between breeding stocks within a breed compared to the differences between breeds. These are unique data because there has not previously been a data set large enough and with a sufficient number of herds to clearly demonstrate these differences in carcass and meat quality traits. The breed range in Table 2 shows the difference between the means (from Table 1) for the top and bottom breeds for each trait. To illustrate the variation among herds as sources of breeding stock, we have presented the estimated difference between the top versus the bottom 10% of herds. We have assumed that there is similar variation within each breed. As an example, the difference between the top and bottom breeds for days to 100 kg was 10.6 days (Hampshire versus Landrace, from Table 1) while the differences between the top 10% and bottom 10% of herds was estimated to be 14 days. This indicates that there is slightly more opportunity to improve days to 100 kg by using pigs from the best herds within a breed than by choosing between breeds. This is indicated by the ratio of herd range to breed range in Table 2. Where this ratio is greater than one, the difference between herds within a breed is greater than the difference between the best and worst breed for that trait. There were some very large differences in this ratio for the performance and carcass yield traits. For example, the ratio for backfat was 3.36, which indicates that even for the fattest breed, there are breeding stocks which are considerably leaner than the average for the leanest breed. Other traits with high herd to breed ratios were estimated yield, carcass index, minimum backfat, minimum loin fat, and lean content of each of the three primals. Analyses of the meat quality traits indicated that there is considerably more variation between breeds than between breeding stocks within a breed for meat quality. Thus, if the goal is lower drip loss, better colour, and firmer structure, it is most important to choose the right breed, with choice of breeding stock within a breed being a lesser concern.

Sex by Breed Interactions

The differences between barrows, gilts and boars were not the same for all breeds. Appendix 4 gives a description of the significant breed by sex interactions. A striking interaction appeared for days to 100 kg, where for the Hampshire breed, boars were similar to the average of barrows and gilts, being only 1.6 days earlier, whereas they were about 5 to 6 days earlier to 100 kg than in other breeds. The interaction for average daily gain showed that gilts grow faster relative to boars and barrows in Hampshires than in other breeds. There were a number of interactions for the traits relating to lean and fat yield of the carcass. Of particular interest is that the interactions for the fat measurements are closely reflected in the interactions for estimated lean. This suggests that the interactions in estimated lean are due to differences in fat distribution rather than lean distribution. The Hampshire breed overall had a higher drip loss in the loin and ham than the other breeds (Table 1). The sex by breed interaction shows that this was mainly due to a much higher drip loss in Hampshire boars, with considerably lower drip loss from barrows and gilts. The interactions for the pork quality traits show that although breed is an important factor, sex differences within a breed can contribute to these differences. The higher marbling in the Duroc breed overall (Table 1) is seen more in barrows and gilts than in boars. Similar breed by sex interaction effects were observed for the colour and structure scores.

Implications

The breed differences were generally smaller than many might expect, but considerable variation still exists within breeds. These differences among and within breeds allow producers to fit different breeds and lines to specific production and marketing niches. Several expected breed differences were confirmed. For example, a higher level of marbling in Duroc pigs was clearly shown. For many of the performance and carcass characteristics there was more variation between herds than between breeds. This means that when choosing breeding stock it is often more important to identify the best breeding stock within a breed than to identify the best breed overall. The sex differences strongly emphasize the need to tailor the production and marketing system to take advantage of the sex differences, where these are advantageous, and minimize the differences where these are disadvantageous. The breed by sex interactions are particularly interesting because few data sets have been large enough to demonstrate such differences. These differences were large enough that they could affect choice of breeds for specific production systems. What we cannot determine, however, is how consistent these breed by sex interactions might be across different populations. Collectively, these results show the advantage of conducting a large and comprehensive analysis of carcass and meat quality traits. Although many large projects have been conducted to evaluate growth performance traits, few comprehensive programs have been completed for carcass and meat quality. The data provided by this project provide a sound basis for choice of breeds to match varying production and marketing systems. There is, however, large variation between different genetic stocks within breeds and there is continuous genetic change in different populations (see article by Gibson in these proceedings). This argues strongly for continuous programs to estimate the merits of individual breeding stocks over time, something the OPCAP was not set up to do.

Table 1. Least square means for breed and sex1,2,3

Breed
Sex
Trait Duroc Hampshire Landrace Yorkshire Barrow Gilt Boar
Age on test (d)
80.5a
81.5a
77.5b
79.0ab
79.9
79.9
79.1
Weight on test (kg)
31.8ab
32.7a
34.5b
31.3b
31.9
31.9
31.6
Weight at slaughter (kg)
107.4a
105.1b
107.1a
105.4b
105.5a
103.8b
109.4c
Days to 100 kg
157.9a
165.5b
154.9c
161.3b
157.9a
164.0b
157.7a
Average daily gain (kg/d)
0.894a
0.811b
0.899a
0.858c
0.888a
0.823b
0.886a
Backfat at 100 kg (mm)
13.89a
13.21ab
13.83a
12.88b
15.22a
13.19b
11.94c
Feed conversion
2.63
2.69
2.65
2.61
2.74a
2.46b
Hot carcass wt (kg)
84.4ab
83.0a
84.9b
84.1a
83.4a
82.3b
86.6c
Cold carcass wt (kg)
75.3
74.3
76.2
75.3
74.7a
73.8a
77.4b
Dressing %
78.5a
78.9a
79.0a
79.5b
79.0b
79.4a
78.5b
Estimated yield (old) (%)
51.0ab
50.9ab
51.0a
51.4b
50.0a
51.5b
51.7b
Estimated yield (new) (%)
60.4
60.7
61.1
60.8
59.5a
61.4b
61.4b
Carcass index (no covariate)
106.3ab
107.7ab
105.9a
107.7b
105.6a
109.5b
105.6a
(slaughter weight covariate)
107.1
107.3
106.5
107.3
105.3a
108.0b
107.9b
Carcass length (cm)
81.5a
81.0a
84.5b
83.2c
81.6a
82.4b
83.6c
Moisture loss (%)
1.28
1.36
1.25
1.29
1.27
1.28
1.32
Skin thickness shoulder (mm)
2.1
2.1
2.0
1.9
2.0a
1.6b
2.4c
Skin thickness back (mm)
2.8a
3.0a
2.5a
2.6a
2.5a
2.6a
3.1b
Skin thickness loin (mm)
3.4a
3.8a
2.8b
3.0b
2.9a
3.1a
3.7b
Skin thickness chop (mm)
3.3a
3.4a
2.5b
3.0a
2.7a
2.8a
3.7b
Max. fat depth, shoulder (mm)
42.9a
38.1b
39.4b
40.2b
42.8a
40.5b
37.1c
Min. backfat (mm)
20.9a
21.1a
20.4a
20.3a
23.9a
20.4b
17.8c
Min. loin fat (mm)
26.7
25.6
26.3
25.9
29.3a
26.6b
22.3c
Shoulder % of side
28.8a
29.1a
28.1b
28.8a
28.2a
28.2a
29.7b
Loin % of side
25.8a
25.7a
26.9b
26.4c
26.3a
26.4a
25.9b
Ham % of side
27.3a
27.0a
25.9b
26.5c
26.6a
26.7b
26.8b
Belly % of side
18.1a
18.2a
19.1b
18.4a
18.9a
18.6a
17.7b
Loin eye area (cm2)
40.1a
45.6b
41.3a
42.9c
39.8a
44.4b
43.2c
Lean content of shoulder (%)
46.9a
47.0ab
47.9b
48.1b
45.9a
47.8b
48.6c
Lean content of loin (%)
50.4a
52.2ab
50.5a
52.1b
47.7a
52.1b
54.0c
Lean content of ham (%)
60.8a
61.8a
61.1a
63.4b
59.4a
62.3b
63.6c
Lean content of 3 primals (%)
52.5a
53.4ab
53.0a
54.3b
50.8a
54.0b
55.1c
Shoulder lean (% of 3 primals)
31.5
31.4
31.6
31.5
31.6a
30.9b
32.0c
Loin lean (% of 3 primals)
29.5a
30.0ac
31.1b
30.3c
29.8a
30.8b
30.1c
Ham lean (% of 3 primals)
38.9a
38.5ac
37.3b
38.2c
38.5a
38.3a
37.8b
Chemical fat loin (% of DM)
18.2a
16.9ab
15.5b
16.2ab
18.8a
16.4b
14.7b
Chemical nitrogen loin (% of DM)5
11.9ab
11.6a
12.2b
12.2b
11.7a
12.0b
12.2b
Chemical fat belly (% of DM)
60.6a
62.4ab
68.3b
63.4a
70.6a
63.1b
57.4c
Chemical nitrogen belly (% of DM)5
5.2a
5.1ab
4.4b
5.2a
4.0a
5.0b
6.1c
Belly ratio 14
0.923ac
0.930abc
0.900b
0.924c
0.925
0.912
0.920
Belly ratio 24
0.104a
0.100ab
0.065b
0.086c
0.077a
0.099b
0.090ab
Belly ratio 34
0.319a
0.282ab
0.276b
0.303a
0.277a
0.291ab
0.317b
Loin drip loss (%)
10.2a
14.2b
12.1c
11.7c
11.8a
12.6b
11.7a
Ham drip loss (%)
10.5a
14.1b
8.0c
9.5d
10.9a
11.4a
9.2b
Marbling score, loin
2.92a
1.78b
1.78b
1.71b
2.32a
1.97b
1.84c
Ag. Canada colour, loin
2.85a
2.63abc
2.41b
2.62c
2.60ab
2.70a
2.60b
Jap. colour, loin
3.52a
3.40ac
3.11b
3.35c
3.32
3.34
3.37
Ag. Canada structure, loin
2.94a
2.41bc
2.39b
2.57c
2.62
2.59
2.53
Ag. Canada colour, ham
3.10a
2.52b
3.60c
3.47d
3.08a
3.18ab
3.25b
Ag. Canada structure, ham
2.83a
2.25b
2.86a
2.78a
2.64
2.66
2.75
1From analyses including breed, sex and PSS status and all two-way interactions. Details of analyses given in Appendix 3.
2 Breeds or sexes with shared superscripts or without superscripts do not differ significantly (P > 0.05).
3 Analyses for the first eight traits (i.e. age on test to hot carcass weight) and carcass index (no covariate) did not include hot carcass weight as a covariate. Analysis of carcass weight (slaughter weight covariate) included slaughter weight as a covariate.
4 See Appendix 2 for definitions. 5 Multiply by 6.25 to get chemical protein %.

Table 2. Variation between herds relative to variation between breeds.

Trait
Breed Range1
Top vs Bottom 10% of Herds2
Ratio of Herd Range to Breed Range3
Age on test (d)
4.0*
15.3***
3.8
Weight on test (kg)
3.2**
1.1ns
0.35
Weight at slaughter (kg)
2.3***
2.2ns
0.94
Days to 100 kg (kg)
10.6***
14.0***
1.32
Average daily gain (kg/d)
0.088***
0.087**
0.99
Backfat at 100 kg (mm)
1.01**
3.40***
3.36
Feed conversion
0.08ns
0.24***
3.06
Hot carcass wt (kg)
1.9*
1.5ns
0.79
Cold carcass wt (kg)
1.9ns
1.6ns
0.83
Dressing %
1.0***
1.2ns
1.20
Estimated yield (old) (%)
0.9ns
1.9***
2.07
Estimated yield (new) (%)
0.7ns
2.6**
3.71
Carcass index (old)
1.8**
3.3ns
3.65
Carcass length (cm)
3.5***
2.2***
0.62
Moisture loss (kg)
0.11ns
0.05ns
0.47
Skin thickness, loin (mm)
1.0ns
0.7ns
0.73
Max. fat depth, shoulder (mm)
4.8***
4.1**
0.85
Min. backfat (mm)
0.8ns
4.2***
5.19
Min. loin fat (mm)
2.1ns
5.7***
2.70
Shoulder % of side
1.0***
0.9*
0.93
Loin % of side
1.2***
1.0**
0.80
Ham % of side
1.4***
0.9**
0.67
Belly % of side
1.0***
1.2***
1.22
Loin eye area (cm2)
4.5***
6.9***
1.43
Lean content of shoulder (%)
1.2**
3.3***
2.73
Lean content of loin (%)
1.8***
5.6***
3.10
Lean content of ham (%)
2.6***
3.8***
1.44
Lean content of 3 primals (%)
1.8***
4.1**
2.26
Shoulder lean (% of 3 primals)
0.2ns
1.1*
5.74
Loin lean (% of 3 primals)
1.6***
1.6***
1.00
Ham lean (% of 3 primals)
1.6***
1.3***
0.83
Chemical fat, loin (% of DM)
2.7*
3.4ns
1.24
Chemical nitrogen, loin (% of DM)5
0.6*
0.4ns
0.61
Chemical fat, belly (% of DM)
7.7**
8.4*
1.09
Chemical nitrogen, belly (% of DM)5
0.8**
1.2**
1.50
Belly ratio 14
0.030**
0.030ns
1.0
Belly ratio 24
0.041***
0.000ns
0
Belly ratio 34
0.043**
0.038ns
0.88
Loin drip loss (%)
4.0***
2.1*
0.53
Ham drip loss (%)
6.1***
2.6*
0.42
Marbling score, loin
1.21***
0.54**
0.45
Ag. Canada colour, loin
0.44***
0.24ns
0.54
Jap. colour, loin
0.41***
0.25ns
0.61
Ag. Canada structure, loin
0.55***
0.13ns
0.26
Ag. Canada colour, ham
1.08***
0.42*
0.39
Ag. Canada structure, ham
0.61***
0.35*
0.59
1 Difference between highest and lowest performing breed for each trait.
2 Predicted difference between top 10% and bottom 10% of breeders, assuming between herd variation is the same for all breeds.
ns,*, **,*** Indicates statistical significance of breed range and between herd variation; ns = not statistically significant, * = P < 0.05, ** = P < 0.01 or *** = P < 0.001.
3 A ratio greater than 1.0 indicates that the predicted difference between the top and bottom 10% of breeders for a given breed is greater than the difference between the best and worst breed for that trait.
4 See Appendix 2 for definitions. 5 Multiply by 6.25 to get chemical protein %.