CC BY
12
Науковий вгсник ЛНУВМБТ 1мен1 С.З. Гжицького Том 10 № 2(37) Частина 4, 2008
Agata Gruzewska1, Barbara Biesiada-Drzazga2, Maria Markowska3 ®
1Department of Agricultural Experimentation, 2Department of Breeding Methods, Poultry Breeding and Small Ruminants,
3Department of Foreign Languages, University of Podlasie, Prusa 14 St., 08-110 Siedlce, Poland
COMPARATIVE STUDIES OF BROILER CHICKEN COMMERCIAL LINES PART 1: AN APPLICATION OF THE CLUSTER ANALYSIS IN RESEARCH ON BROILER CHICKEN COMMERCIAL LINES
Abstract
The aim of the present work has been to simultaneously perform the cluster analysis for twenty-two production characteristics of broiler chickens representing eleven commercial lines.
There were obtained different results of the analysis of clusters for empirical and normalised data. It confirms the need to normalise values of characteristics accompanied by different units while performing the analysis of clusters. The chicken commercial lines compared differed according to the production traits analysed. In the first place only two commercial lines were assigned to two clusters.
Key words: broiler chickens, commercial lines, cluster analysis
Introduction
In the year 2000 sixteen commercial lines were accepted for distribution on the Polish market. However, before issuing a permit to distribute a new commercial line on the area of Poland, the line's assessment must always be carried out. One of the methods of assessing the current usable quality of meat chickens are the so-called tests of random samples carried out on specifically prepared farms called test stations (Gaw$cki,1998). On the basis of such an assessment production potential of chicken breeding and commercial material available in Poland is determined. However, one cannot forget that the production level reached by broiler chickens depends not only on the bird genotype but also a broadly understood impact of environmental factors.
As a result, a need emerges to determine an influence of genetic factors on the production results obtained taking into account both quality and quantity. An existence of such an impact has been confirmed in the studies by Gornowicz and Dziadek (2001).
In the available literature there can be found the works which present production results (Gaw^cki and Gornowicz, 2000; Gaw^cki 2001a and 2001b) and values of chicken meat qualitative characteristics taking into account the origin of the chickens. However, statistical analysis of the results is mainly performed for each characteristics separately. Hence, the aim of the present work has been to simultaneously perform the cluster analysis for twenty-two production characteristics of broiler chickens representing eleven commercial lines.
® Agata Gruzewska, Barbara Biesiada-Drzazga, Maria Markowska, 2008
274
Material and methods
Rearing performance data for broiler chickens from eleven commercial lines were use in the study. The birds were reared in eight poultry houses of the Chotycze Poultry Farm, eastern Poland. All results of rearing (from February 2000 to September 2007) were recorded by farm workers, separately for each bout and poultry house, and random groups of 100 birds were weighed at regular seven day intervals. The same housing conditions were maintained in all poultry houses and bouts. Feed manufacturer's recommendations were applied when feeding all commercial lines.
In order to achieve the assumed aim the cluster analysis for empirical data (tab. 1.1. and 1.2.) was applied using the 1-r Pearson distance as a measure of distance, and a full linkage and normalized data as a method of linking objects by applying the Euclidean distance as a measure of distance, and a full linkage. Calculations were conducted using Statistica PL 6.0 (StatSoft, Inc., 2001) for the twenty two characteristics.
Table 1.1.
Selected production results ( y - mean, V% - coefficient of variability, n - number of observations) for broiler chickens of the commercial lines under comparison.
Commercial line Percentage of birds sold to the slaughter house Average body weight (g) Concentrate consumption per 1 kg of body weight gain (kg) EWW
in successive weeks of fattening: over the whole period of fattening
1 2 3 4 5 6 7
F-15 n = 18 y V% 95,4 2,39 136,7 5,98 312,5U 9,54 642,0n 4,53 1048,7 5,93 1475,0 4,99 1936,7 4,96 2112,5 4,96 2047 5,31 1,871 1,32 2,381 0,96
Flex n = 16 y V% 96,6' 2,42 155,3* 6,00 373,3 8,92 706,5 4,60 1110,0 6,02 1610,0 5,00 2025,0 4,82 2303,31 4,82 2253 5,20 1,98 1,35 2,30 1,02
Hybro G n = 26 y V% 90,6" 2,18 149,6 6,02 369,6 9,02 716,11 4,61 1089,6 6,00 1543,6 5,02 1962,1 4,79 2121,3 4,79 2060 4,90 2,18u 1,12 2,22 1,07
Hubbard n = 19 y V% 96,61 2,12 142,7 5,89 338,5 8,95 646,5 4,70 1012,1 5,82 1498,2 5,10 1943,5 4,85 2200,0 4,85 2168 5,50 2,08 1,80 2,03n 1,12
Hybro N n = 28 y V% 95,0 2,05 136,4U 6,03 326,9 7,60 659,4 4,20 1028,3 5,73 1530,4 5,25 1954,1 5,00 2063,9" 5,00 2026 6,01 2,01 1,76 2,21 1,14
Hybro PM n = 23 y V% 95,9 2,30 149,5 5,95 375,01 8,02 700,3 4,18 1032,5 5,75 1442,5 5,18 1947,5 5,02 2103,3 5,02 2053 6,20 2,03 1,62 2,24 1,12
ISA 215 n = 28 y V% 92,7 2,15 140,4 5,87 350,0 8,73 701,7 4,32 1092,51 5,12 1452,0 4,98 1971,7 4,98 2170,0 4,98 2035 5,81 2,05 1,82 2,33 1,03
275
Table 1.1. cont.
Percentage Average body weight (g) Concentrate
of birds sold to the slaughter house in successive weeks of fattening: over the consumption per 1 kg of body weight gain (kg)
Commercial line 1 2 3 4 5 6 7 whole period of fattening EWW
PM 3 y 92,6 141,7 341,7 669,0 986/711 1426/711 1903,3" 2296,7 22971 2,09 2,14
n = 20 V% 2,21 5,62 8,92 4,25 5,18 4,88 4,74 7,80 5,29 1,93 1,08
Ross 308 y 94,4 149,8 363,4 705,0 1035,7 2216,8' 2059,5* 2258,2 2128 1,98 2,47
n = 24 V% 2,13 5,78 8,83 4,33 5,22 4,62 5,03 8,20 6,12 1,42 1,05
Ross 508 y 94,5 148,4 363,4 700,2 1033,2 1562,1 2010,7 2180,9 2068 2,01 2,37
n = 19 V% 2,18 5,80 9,01 4,19 5,34 5,03 5,07 8,35 6,20 1,85 1,08
Ross PM 3 y 93,1 145,3 333,8 645,8 1026,8 1550,0 1910,0 2085,0 2070 2,03 2,14
n = 16 V% 2,03 5,79 9,03 4,80 5,26 4,82 4,76 8,14 6,03 1,92 0,99
Source: Personal calculations
I) the most favourable value of a characteristic (within the group under comparison)
II) the least favourable value of a characteristic (within the group under comparison) EWW calculated according to the following formula:
_liveability (%) • body live weight (kg)_
number of fattening days • feed consumption per 1 kg body weight (kg)
Table 1.2.
Liveability results (y - average, V% - coefficient of variation, n- number of
observations) 'or broiler chickens of the commercial lines under comparison
Average number of fettenin g days Mortality
Commer cial line during transportatio n to the in successive weeks of fattening over the whole period of fattening
slaughter house 1 2 3 4 5 6 7 on average in total
F-15 y 43,3' 19,2 130,3 81,71 88,01 67,31 102,2 150,2 85,0 112,7 676,3
n = 18 V% 4,15 18,81 22,49 28,92 43,37 22,00 35,79 43,30 64,23 18,90 19,01
Flex y 47,5 70,0 122,85 105,3 101,5 93,3 120,3 170,0 202,7 136,5 878,5
n = 16 V% 4,18 25,62 21,3 30,15 45,58 25,30 36,23 45,27 89,38 19,82 26,80
Hybro G y 44,1 40,2 198,2 99,6 109,7 85,4 114,6 990,4U 110,6 260,8U 1668,9U
n = 26 V% 4,12 31,24 20,87 28,23 49,29 26,82 33,58 50,32 88,17 22,80 36,94
Hubbard y 46,6 49,1 171,3 123,6 116,8 96,7 137,6 148,3 102,1 148,3 897,5
n = 19 V% 4,16 32,45 18,23 32,46 46,32 30,14 38,42 44,23 77,26 20,98 42,62
Hybro N y 43,4 73,9 170,0 107,7 116,6 93,3 138,2 234,3 181,3 148,1 980,9
n = 28 V% 4,80 35,71 18,12 30,18 45,24 29,66 34,31 38,95 65,42 19,71 30,43
Hybro PM y 43,7 95,2 140,8 104,3 96,0 89,0 80,81 110,0 62,0 106,6* 667,31
n = 23 V% 4,25 38,82 17,65 29,62 112,86 27,52 22,80 33,26 39,23 15,70 19,55
ISA 215 y 41,7 129,811 284,3 175,3 188,3 128,2 210,5 345,5 110,5 218,3 1387,3
n = 28 V% 4,31 40,93 32,19 32,89 50,15 35,18 40,13 34,34 48,74 25,63 44,90
276
Table 1.2. cont.
Mortality (%)
Average
Commer number of fettening days during transporta tion to the slaughter house in successive weeks of fattening over the whole period of fattening
cial line 1 2 3 4 5 6 7 on averag e in total
PM 3 y 49,0U 17,0' 82,0* 131,7 131,7 127,0 106,7 398,0 229,0 201,0 1360,0
n = 20 V% 4,30 8,96 13,23 36,80 55,90 30,93 36,47 29,77 26,70 22,39 47,84
Ross 308 y 43,8 35,9 159,4 94,3 113,5 88,3 243,6U 143,9 238,3U 162,2 973,0
n = 24 V% 4,82 15,30 28,98 26,41 46,23 19,80 42,99 22,55 20,76 18,76 33,62
Ross 508 y 44,0 26,0 174,6 136,9 163,4 112,4 125,7 220,6 101,3 168,9 1013,4
n = 19 V% 4,21 13,18 27,67 28,95 47,24 21,81 24,65 47,32 11,77 17,63 44,28
Ross PM 3 y 44,5 123,8 311,8" 180,0n 220,8U 138,0n 93,3 97,51 35,51 158,5 1077,8
n = 16 V% 4,93 22,63 35,19 34,62 51,32 23,58 20,86 11,89 18,30 15,70 39,80
Source: Personal calculations
1)1 the most favourable value of a characteristic (within the group under comparison) n) the least favourable value of a characteristic (within the group under comparison)
Results of researches
A tree diagram for the empirical data is presented in figure 1. The most similar chicken commercial lines formed four clusters including the following lines: Hybro PM and F-15; Flex and Hubbard; Hybro N and Ross 508, as well as ISA 215 and PM3. A detailed analysis of distances between the lines in individual clusters revealed that the smallest distance was obtained for the first three pairs of commercial lines. Lines Flex and Hubbard as well as Hybro N and Ross 508 formed the next cluster with a higher probability than for the ISA 215-PM3 pair, the lines forming one cluster of the first order.
It should be noted that values of the characteristics examined belonged to several scales, and ranged from 1.87-2.18 for concentrate consumption to 2063-2303 for body weight at termination of fattening.
Due to such high divergence, values of the latter trait markedly influenced the differences (and similarities) between the examined commercial lines of broiler chickens. In such cases normalisation of values of a characteristic is advised.
A tree diagram for normalised values is presented in figure 2. The diagram is different than for the empirical data. The number of clusters including most similar lines was five and the clusters included: Hybro N and Hubbard; Hybro PM and Ross 508; Flex and Ross 308; Hybro G and PM3. It should be stressed that two pairs of most similar lines - Hybro N and Hubbard as well as Hybro PM and Ross 508 form the next two much more similar clusters than the ISA 215-PM3 pair, the lines forming one cluster of the first order.
277
Figure 1. Diagram of a tree for the examined commercial lines of broiler chicken according to the values of production characteristics which were taken into account.
F-15 Hybro PM Flex Hubbard Hybro N Ross 508 Ross PM3 Ross 308 HYbro G ISA 215 PM-3
0,00 0,01 0,02 0,03 0,04 0,05 0,06 Distance of the bond
0,07
0,08
0,09
Figure 2. Diagram of a tree for the examined commercial lines of broiler chickens according to the values of production characteristics which were subjected to normalisation.
F-15 Hubbard Hybro N Hybro PM Ross 508 Flex Ross 308 HYbro G PM-3 ISA 215 Ross PM3
2,0
Distance of the bond
1,0
1,5
2,5
3,0
278
Conclusions
There were obtained different results of the analysis of clusters for empirical and normalised data. It confirms the need to normalise values of characteristics accompanied by different units while performing the analysis of clusters.
The chicken commercial lines compared differed according to the production traits analysed. In the first place only two commercial lines were assigned to two clusters.
The commercial lines forming the clusters originated from different breeding companies.
References
1. Gaw^cki, W., 1998. Niektore hodowlane i poza hodowlane uwarunkowania efektywnosci produkcji kurcz^t brojlerow. Pol. Drob., 10, 9-13.
2. Gaw^cki W. (2001a): Ocena wartosci uzytkowej kurcz^t brojlerow pochodz^cych po roznych stadach rodzicielskich utrzymywanych na terenie Polski - test nr 3/2000. Pol. Drob. (1): 24-26.
3. Gaw^cki W.(2001b): Znaczenie wynikow testow kurcz^t brojlerow wykonywanych w 1998 i 2000 roku dla ksztaltowania zaplecza produkcyjnego w Polsce. Pol. Drob. (4): 12-14.
4. Gaw^cki W. and Gornowicz E, 2000: Wartosc odzywcza i technologiczna mi^sa kurcz^t brojlerow roznego pochodzenia. Zesz. Nauk. Prz. Hod. PTZ Chow i Hodowla Drobiu, 49: 271-272.
5. Gornowicz E. and Dziadek K., 2001: Zmiennosc skladu chemicznego mi^sni piersiowych i udowych w zaleznosci od pochodzenia kurcz^t brojlerow. Zesz. Nauk. Zoot., 28 (2): 89-100.
6. Gaw^cki W., Kielczewski K., Szlinka U., Gaw^cka K.E. (2001): Ocena wartosci uzytkowej kurcz^t brojlerow pochodz^cych z roznych stad rodzicielskich utrzymywanych na terenie Polski (test nr 1/2000). Wyniki oceny uzytkowosci drobiu. (29): 3-77. Instytut Zootechniki.
7. StatSoft, Inc. (2001). STATISTICA (data analysis software system), version 6. www.statsoft.com.
Summary
Rearing performance of chicken broilers from eleven commercial lines was use in the study. They were reared in eight poultry houses of the Chotycze Poultry Farm.
In order to achieve the assumed aim the cluster analysis for empirical data was applied using the 1-r Pearson distance as a measure of distance, and a full linkage and normalized data as a method of linking objects by applying the Euclidean distance as a measure of distance, and a full linkage.
The diagram for normalised values is different than in the case of the empirical. The number of clusters including most similar lines was five and the clusters included: Hybro N and Hubbard; Hybro PM and Ross 508; Flex and Ross 308; Hybro G and PM3.
Cmammn nadiumna do peda^i'i 14.04.2008
279
