Breeding strategies for improving performance of Kuchi chicken ecotype of Tanzania for production under village conditions 

J Lwelamira, G C Kifaro* and P Gwakisa** 

Institute of Rural Development Planning, P.O. Box 138, Dodoma, Tanzania;

jlwelamira@yahoo.com

*Department of Animal Science and Production, Sokoine University of Agriculture, P.O. Box 3004, Morogoro, Tanzania;

**Department of Veterinary Microbiology and Parasitology, Sokoine University of Agriculture, P.O. Box 3019, Morogoro, Tanzania

Abstract 

A study was carried out to evaluate various breeding scenarios geared at improving performance of Kuchi chicken ecotype of Tanzania, a type of native chicken found in drier area of north-west part of the country, for production under extensive (village) management conditions. The purpose of these breeding scenarios was mainly to improve body weight at 16 weeks of age (Bwt16). In some breeding scenarios, primary antibody response against Newcastle disease virus vaccine (Ab) was also considered.

Depending on a breeding scenario, results from the present study indicate that it would take approximately 5 to 10 generations of selection, which corresponds to around 3 to 6 years of selection for improving Bwt16 either singly or together with Ab from their current population mean of 974g and 4.8 (HIlog₂) to the desired population mean of 1300g and 6(HIlog₂), respectively (i.e. improvement by 34 and 25%, respectively). Depending on gain per generation, number of years required to attain the desired mean, expected fitness under village conditions and costs of breeding, some breeding scenarios were recommended.

Key words: body weight, desired gain, extensive management, Newcastle disease, selection

Introduction 

Animal protein intake in developing countries including Tanzania has been very low compared to the recommended level (Pedersen 2002; Nielsen et al 2003). The situation is further aggravated by high incidences of poverty in these countries. Due to their short generation intervals, chickens have high potential to offset this problem of low protein intake, and they can as well act as source of income and hence reducing poverty. Local chickens dominate the chicken industry in developing countries. It has been estimated that more than 70% of rural households in developing countries keep local chickens (Aini 1990; MOAC 1997). These local chickens are mostly kept under free range management system (i.e. extensive management) in villages. According to Melewas (1989) and Roetter et al (2007), more than 85% of people in developing countries reside in rural areas. Since majority of people in developing countries live in rural areas (villages), in which most of them are keeping local chickens, it is obvious that putting an emphasis on local chickens would have an immediate positive impact on animal protein intake and income by most of the people in these countries. Low genetic potential for production traits and frequent outbreak of diseases specifically Newcastle disease has been noted in a significant number of studies to be among of the major factors limiting productivity of the local chickens in the Tropics both under intensive and extensive management systems (Katule and Mgheni 1990; Yongolo 1996; Alexander 2001; Msoffe 2003). Crossbreeding programs with specialized meat type or egg type chickens has been shown by several workers to improve their productivity significantly (Ali et al 2000; Pedersen 2002; Theerachai et al 2003; Segura-Correa et al 2004). However, these crossbreeding programs are threatened by the current global move on conservation of indigenous genetic resources which campaigns against genetic dilution of indigenous genetic resources (Msoffe 2003; Kosgey 2004). This call for a need to look for an alternative approach for genetic improvement of local chickens. In this regard, genetic improvement through selection within local chickens could be a good option.  A study conducted by Lwelamira (2007) both under intensive and extensive management (i.e village conditions) in two Tanzania local chicken ecotypes viz. Kuchi and Tanzania Medium (Medium) have shown existence of significant additive genetic variation in various traits in these ecotypes and hence expecting adequate response to selection. Based on their performances, the study further revealed Kuchi ecotype to be good starting material for improving for meat production and Medium ecotype for egg production. Therefore, this study was carried out to evaluate various selection breeding strategies for improving performance of Kuchi ecotype for management under village (extensive) conditions.

Materials and methods 

Comparison of various breeding scenarios/ schemes

Various breeding scenarios were designed and compared through simulation. These scenarios were compared in terms of gain per generation, number of generations required to achieve the desired performances and correlated responses in some breeding scenarios (i.e. those involved selection under intensive management and selected stock expected to perform under extensive management). The emphasis in this study was on improving body weight at 16 weeks of age (Bwt16). In some breeding scenarios, primary antibody response against Newcastle disease virus vaccine (Ab) (i.e. Indicator for immunocompetence to Newcastle disease) was also considered. Egg production and related traits were not considered in this study because no data were available for these traits in this ecotype under extensive management system. For the scenarios which involved single trait selection based on own performance, response per generation were calculated according to Falconer and Mckay (1996) using equation 1. Correlated responses On-farm (extensive management) following single trait selection On-station (Intensive management) were calculated using equation 2 based on Falconer and Mckay (1996). Furthermore, since relative economic weights for the traits considered were not available, those scenarios where more than one trait were considered simultaneously (i.e Bwt16 and Ab), desired gain selection indices were used. The desired gain indices were constructed according to Yamada et al (1975) as applied in subsequent studies (Gill and Verma 1983; Hazary et al 1998; Nishida et al 2001; Noda et al 2002; Kaushik and Khanna 2003). Therefore, gain per generation and number of generations required to achieve the desired performances in multi-trait selection were computed as per Yamada et al (1975).

  (Equation 1)

Where:

R = response to selection per generation
i = intensity of selection,
h² = heritability estimate,
σ_P = phenotypic standard deviation.

Heritabilities and phenotypic standard deviations were obtained from the study by Lwelamira (2007) for this chicken ecotype (Table 1).

     (Equation 2)

Where,

CR₂= Correlated response in environment 2 following selection in environment 1;

r_A= genetic correlation between the same trait measured in two environments (i.e. environment 1 and 2) viz. intensive and extensive management system;

σ_A2 = additive genetic standard deviation for a trait in environment 2;

σ_A1 = additive genetic standard deviation for a trait in environment 1;

R₁= response to selection for a trait in environment 1.

Genetic parameters for the above formula were derived from parameters for this ecotype in Table 1.

Use of index I in selection usually involves calculation of weighting factors b for traits to be used as selection criteria (equation 3).

I = b′ X    (Equation 3)

Where:

I= Selection index;
b =  n x 1 vector of weighting factors;
X = n x 1 vector of source of information, usually phenotypic measurements on candidate for selection or its relatives.

In the present study, information source (X) was individual own performance.

Based on the Yamada index, b in the present study was calculated as:

b = (G′ R) ^–1Q    (Equation 4)

Where:

G = n x m genetic variance-covariance matrix of the traits used as selection criteria and traits in the breeding objectives;
Q = m x 1 vector of intended genetic changes for m traits assigned by breeder;
R = n x n matrix of Wrights coefficient of relationship.

Genetic variance-covariances were derived from genetic parameters estimated for this ecotype in a study by Lwelamira (2007) (Table 1).

Desired genetic changes for various traits (i.e. Q) were calculated as the difference between desired and observed means (Yamada et al 1975; Kaushik and Khanna 2003; Suzuki et al 2005).

Intended performances and hence desired genetic changes were chosen in such a way that they are within the capacity of the population as shown by performance of some individuals in an environment under consideration (Lwelamira 2007). Furthermore, the chosen intended performances were close to the performance of the crosses between local and exotic birds (Ali et al 2000; Theerachai et al 2003; Lwelamira and Katule 2004). In addition, the chosen intended performance for body weight also depended on weight at which a chicken can be marketed (Pedersen 2002; Theerachai et al 2003; Acamovic et al 2005). By considering body weight much further, assessed from maximum weights possible under different ages (i.e. capacity of a population), since possibility of attaining market weight (i.e. at least 1kg) earlier than 16 weeks of age through selection under extensive management conditions was non-existent (Lwelamira 2007). Moreover, as heritability estimates for body weights were relatively higher in later ages compared to earlier ages (Lwelamira 2007). Therefore, body weight at 16 weeks of age (Bwt16) was chosen as the target body weight. Furthermore, based on the capacity of a population (Kuchi), body weight at 16 weeks of age was also chosen as it is possible to target much higher weights in a breeding objective (i.e. above 1kg under extensive management conditions, and hence chickens would be fetching good prices) compared to body weights at 8 and 12 weeks of age, and at earlier age than at 20 weeks of age (Lwelamira 2007).

 Based on Yamada index, expected genetic gains per generation were calculated using equation 5.

   (Equation 5)

Where:

i_I  = intensity of selection based on the index;

σ_I  = standard deviation of the index calculated as shown in equation 6;

∆G = m x 1 vector of genetic gains per generation in m traits;

G = n x m genetic variance-covariance matrix.

   (Equation 6)

Where P is an n x n phenotypic variance-covariance matrix. Again P was derived from parameters in Table 1.

The number of generations q required to attain the pre-defined breeding objectives in multi-trait selection was calculated using equation 7 (Yamada et al 1975; Kaushik and Khanna 2003). All matrix equations were solved using Interactive Matrix Language (IML) procedures of SAS (2000).

    (Equation 7)

Population structure and derivation of selection intensities

In breeding scenarios which involved selection under extensive (village) conditions it was assumed that 100 farmers (households) are involved in the program, and each household keeps 2 to 3 breeding females. Furthermore, it was assumed that, in total in each generation 40 best males and 240 best females are selected, and each cock saves 2-3 households (i.e. a mating ratio of one male to 4 to 9 females). In the study by Lwelamira (2007) for this ecotype it was found that the average number of progeny per dam obtained under field condition to be around 6. Therefore, 240 hens would be expected to produce 1440 progeny in total with 720 birds of each sex. Taking into account the mortality rate/loss of about 33% by the time birds are selected and mated (i.e. average loss of 2 chicks per hen), then the number of birds available for selection would be 240 x 4 = 960 (i.e. 480 chicks of each sex). The chosen percent mortality/loss was based on the obtained percent mortality/loss up to 12 weeks of age of about 30% under extensive management for this ecotype (Lwelamira 2007), and further assuming that once a chick reachies the age of 12 weeks stands a high chance to survive to maturity and breed when regular control of outbreak of major diseases under field conditions specifically Newcastle disease are in place.  Therefore, after accounting for mortalities/loss, selection of 40 best cocks and 240 best females in each generation would result into proportions selected of about 8.3% and 50% for males and female, respectively (i.e. average selection intensity of 1.4). Proportions selected (%) were transformed into selection intensity (i) using a Table by Falconer and Mackay, (1996).

In those breeding scenarios which involved selection under intensive management (On-station) in some stage, taking into account the reproductive performance, survival of the breeding stock under intensive management for this ecotype (Lwelamira 2007) and the recommended cock:hen ratio, proportion selection under this management system was assumed to be  3.7 and 22.2 % for males and females, respectively (i.e. average selection intensity of 1.8). This average selection intensity is within the range of that of around 1.5 to 2 mostly used in commercial chicken breeding (Ameli et al 1991; Su et al 1997).

Results and discussion 

Five breeding scenarios were considered in this study (Table 2).

Due to the presence of genotype by environment interactions observed for this ecotype for Bwt16 in a study by Lwelamira (2007) (i.e. rA= 0.75), some breeding scenarios (1 and 2) were designed to predict response on farm (extensive management) as a result of selection under intensive management. Since apart from genetic correlation between the same trait measured in the two environments, no information were available regarding corresponding phenotypic correlations, together with both genetic and phenotypic correlation between different traits (i.e. trait 1 and trait 2) measured two different environments due to data structure (Lwelamira 2007). Therefore, breeding scenarios which involve computation of correlated responses under extensive management following index selection for Bwt16 and Ab under intensive management was not considered.  Observed means, desired means and hence desired changes are indicated in Table 3.

Results from Table 4 indicate that on-station single trait selection for body weight at 16 weeks of age predicted to yield an on-station response of around 87g per generation would be associated with a correlated response of about 68g per generation under extensive management (scenario 1).

This is close to the response of 65g per generation obtained for single trait selection for body weight at 16 weeks of age carried out in both males and females under on-farm/extensive conditions (scenario 4), but higher that the response of 53g per generation obtained when apart from body weight, antibody response was also considered for selection under on-farm conditions (Scenario 3) (i.e. Index selection for Bwt16 and Ab under extensive management). Selection index (I) for this breeding scenario is shown in equation 8. Furthermore, in a breeding scenario where only improved cocks from the on-station (intensive management i.e. elite nucleus) are taken to the field as breeding males as suggested by Horst (1981), and Mukherjee (1990) (scenario 2), or selection for improving body weight at 16 weeks of age under extensive conditions is done only on males (scenario 5), it was predicted to result into appreciably lower genetic responses per generation (i.e. 34 and 43g, respectively) than those obtained in all other breeding scenarios investigated.    

    (Equation 8)

Considering the average body weight over both sexes of 974g at 16 week of age under extensive management obtained in a study by Lwelamira (2007) for this ecotype, breeding scenarios 1, 2, 3, 4 and 5 would be expected to attain the desired average body weight of 1300g under extensive management (Table 3) after 4.78, 9.57, 6.10, 5.03, and 7.57 generations of selection, respectively (Table 4). By assuming mating, egg laying and incubation and hatching by the hen under extensive conditions takes approximately 2 months (i.e. 1 month for mating and egg laying, and 21 days for incubation and hatching), and age at sexual maturity under this management system is reached at around 8 months of age (Sonaiya 1992; Gunaratne et al 1992; Mwalusanya et al 2002), which would lead to a generation interval of around 10 months for breeding scenario 3, 4 and 5. Further assuming a generation interval of about 8 months (Lwelamira 2007; Su et al 1997) (i.e. 2 months for mating, egg collection, and hatching + 6 months for attaining sexual maturity under intensive system) for breeding scenario 1 and 2, it would require 3.2, 6.4, 5.1, 4.2, and 6.3 years of selection for breeding scenario 1, 2, 3, 4, and 5, respectively for achieving the desired gain.

Breeding scenario 1 and 2 are more convenient (cheap and easy to institute) compared to the other breeding scenarios, however, the problem of these breeding scenarios are reduced fitness under extensive management, and to a large extent in breeding scenario 1 (Horst 1981; Mukherjee 1990 1992; Demeke 2003). Doing selection under farm (extensive) conditions has the advantage of minimizing a problem of fitness under this management system i.e. loss of scavenging habit (scenarios 3, 4, and 5). However, the success of breeding scenarios under extensive conditions requires close supervision on record keeping under farm conditions which could be too involving and expensive. The problem could be more serious in breeding scenario 3 and 4 where both males and females are selected compared to breeding scenario 5 where only males are selected. Despite of being too involving and expensive, results from Table 4 also indicate breeding scenario 3 and 4 to have reasonably high genetic gain per generation. Therefore, in situation where resources are available for carrying out selection program under on-farm conditions, breeding scenario 3 could be recommended as it improves both body weight and humoral immune response to ND which is usually a serious problem under field conditions (Yongolo 1996; Alexander 2001; Ilango et al 2005; Otim 2005) at the expense of one more year of selection relative to breeding scenario 4. In a situation where resources are a problem, breeding scenario 5 or 2 could be advised depending on the magnitude of the problem. However, to achieve the desired gain in body weight, approximately one more year of selection would be required in both breeding scenarios relative to breeding scenario 3. If the problem of resources is not very severe, breeding scenario 5 could be a selection scheme of choice as it is expected to be associated with more fitness under farm conditions compared to breeding scenario 2.

Conclusion 

Approximately 5 to 10 generations of selection, which corresponds to around 3 to 6 years of selection would be required for improving Bwt16 either singly or together with Ab from their current population mean of 974g and 4.8 (HIlog₂) to the desired population mean of 1300g and 6(HIlog₂), respectively. The number of years of selection required to attain the desired gains for the studied ecotype indicates that it won’t take too long to reach the target (desired performance) if selection breeding programs would be initiated.  This suggests that selection breeding programs for improving performance of this ecotype for production under village conditions should be initiated. In addition, for these selection breeding programs to be successful, apart from gain per generation and number of years required to attain the desired mean, the choice of a breeding scenario should also consider expected fitness under village conditions and costs of breeding.

Acknowledgement 

The authors are very grateful for the financial support from Production and Health of Smallholder Livestock (PHSL) project funded by DANIDA which sponsored the senior author in his PhD. Programme.

References 

Acamovic T, Sinurat A, Natarajan A, Anitha K, Chandrasekaran D, Shindey D, Sparks N, Oduguwa O, Mupeta B and Kitaly A 2005 Poultry. In: Livestock and Wealth Creation. Improving the husbandry of animals kept by resource-poor people in developing countries. (Edited by Owen E, Kitaly A, Jayasuriya N and Smith T). Nottingham University press, UK, pp 304- 324

Aini I 1990 Indigenous chicken production in South-east Asia. World’s Poultry Science Journal 46: 51-57

Alexander D J 2001 Newcastle disease. British Poultry Science 42 (1): 5-22

Ali K O, Katule A M and Syrstad O 2000 Genotype x environmental interaction in growing chickens: comparison of four genetic groups on two rearing systems under tropical conditions. Acta agriculturæ Scandinavica 50: 65-71

Ameli H, Flock D K and Glodek P 1991 Cumulative inbreeding in commercial White Leghorn lines under long-term reciprocal recurrent selection. British Poultry Science 32: 439- 449.

Demeke S 2003 Growth performance and survival of Local and White Leghorn chickens under scavenging and intensive systems of management in Ethiopia. Livestock Research for Rural Development. 15 (11): http://www.lrrd.org/lrrd15/11/deme1511.htm  Site visited on 26/3/2005.

Falconer D S and Mckay T F C 1996 Introduction to quantitative genetics. 4^th edition. Longman Group, Essex, UK. 463pp.

Gill H S and Verma S K 1983 Construction of selection index in poultry ignoring relative economic values. Indian Journal of Animal Sciences 53 (10): 1110-1112

Gunaratne S, Chandrasir A, Hemalatha W and Roberts J 1992 The productivity and nutrition of village chickens in Sri-lanka. In: Newcastle disease in village chickens (edited by Spradbrow PB). ACIAR proceeding. Canberra, Australian Centre for International Agricultural Research 39: 144- 148

Hazary R C, Johri D C, Kataria M C, Sharma D and Singh D P 1998 Evaluation of efficiency of multisource weight-free selection index for desired gains in egg type chicken. Indian Journal of Animal Sciences 68(7): 662-666

Horst P 1981 Constraints on the genetic improvement of non-ruminants in the tropics. Animal Research and Development 14: 120-135

Illango J, Olaho-Mukani W, Mukiibi-Muka G, Abila P P and Etoori A 2005 Immunogenecity of a locally produced Newcastle disease I-2 thermostable vaccine in chickens in Uganda. Tropical Animal Heath and Production 37: 25-31

Katule A M and Mgheni M 1990 Performance of crosses between exotic and local Tanzania chickens. In: Proceedings of the 4^th World Congress on Genetics Applied to Livestock Production. Volume XVI. 23- 27 July 1990, Edinburgh, pp 62- 64

Kaushik R and Khanna A S 2003 Efficiency of different selection indices for desired gain in reproduction and production traits in Hariana Cattle. Asian-Australasian Journal of Animal Sciences 16(6):789-793

Kosgey I S 2004 Breeding objectives and breeding strategies for small ruminants in the tropics. PhD thesis. Wageningen University, Wageningen, The Netherlands. 272pp. http://library.wur.nl/wda/dissertations/dis3546.pdf

Lwelamira 2007 Prospects for improving performance of two Tanzanian chicken ecotypes through selection. PhD thesis. Sokoine University of Agriculture, Morogoro, Tanzania. 190pp.

Lwelamira J and Katule A M 2004 Genetic determination of immune responses to Newcastle disease virus vaccine in chickens. Bulletin of Animal Health and Production in Africa 52: 186 – 197

Melewas J N 1989 The contribution of poultry industry to the national economy. In: Proceeding of the 7^th Scientific Conference of the Tanzania Veterinary Association, Volume 7 (Edited by Msolla P and Kazwala R R). 3-5 December 1989, Arusha, Tanzania, pp 13- 36

MOAC 1997 Ministry of Agriculture and Cooperatives, Tanzania. National Sample Census of Agriculture 1994/95- Report Volume II.

Msoffe P M M 2003 Diversity among local chicken ecotypes in Tanzania. PhD thesis. Sokoine University of Agriculture, Morogoro, Tanzania, 223pp.

Mukherjee T K 1990 Breeding selection and biotechnological developments for improvement of poultry in tropics- Current progress and future perspectives.In: Proceedings of the 4th World Congress on Genetics Applied to Livestock Production. Volume XIV. 23-27 july 1990, Edinburg, pp 337- 348

Mukherjee T K 1992 Usefulness of indigenous breed and imported stocks of poultry production in hot climates. Proceedings XIX World’s Poultry Congress, Symposia. 2:31-37

Mwalusanya N A, Katule A M, Mutayoba S K, Mtambo M M A, Olsen J E and Minga, U M 2002 Productivity of local chickens under village management conditions. Tropical Animal Health and Production 34:405- 416

Nielsen H, Roos N and Thilsted S H 2003 The impact of semi-scavenging poultry production on the consumption of animal source food by women and girls in Bangladesh. Journal of Nutrition 133: 4027-4030 http://jn.nutrition.org/cgi/reprint/133/11/4027S

Nishida A, Ogawa T, Kikuchi Y, Wakoh K, Suzuki K, Shibata T, Kadowaki H, Shinohara H and Ohtomo Y 2001 A hopeful prospect for genetic improvement of chronic disease resistance in swine. Asian-Australasian Journal of Animal Sciences 14:106-110

Noda K, Kino K, Miyakawa H, Banba H and Umezawa Y 2002 Persistency of laying strain building by index selection including oviposition time as selection trait in laying hens. Journal of Poultry Science 39: 140- 148

Otim M O 2005 Newcastle disease in village poultry: Molecular and phylogenetic studies of the virus and disease epidemiology. PhD thesis. The Royal Veterinary and agricultural University (RVAU), Copenhagen. Denmark. 140pp.

Pedersen C V 2002 Productivity of semi- scavenging chickens in Zimbabwe. PhD thesis. The Royal Veterinary and Agricultural University (RVAU), Copenhagen, Denmark. 133pp.

Roetter R P, Van Keulen H, Kuiper M, Verhagen J and Van Laar H H 2007 Rural livelihood: Interplay between farm activities, non- farm activities and the resource base. In: Science for Agriculture and Rural Development in Low- Income Countries. Springer, The Netherlands. Pp 77- 95

SAS (Statistical Analysis System) 2000. SAS/STAT Users' Guide, Release 6.12 Edition, SAS Institute Inc, Cary, North Carolina. USA.

Segura-Correa J C, Sarmiento-Franco L, Magaňa-Monforte J G and Santos-Ricalde R 2004 Productive performance of Creole chickens and their crosses raised under semi-intensive management conditions in Yucatan, Mexico. British Poultry Science 45 (3): 342-345

Sonaiya E B 1992 Development strategy for improving sustainable smallholder rural poultry production. Proceedings of the 19th World’s poultry congress. Amsterdam, The Netherlands, pp 684- 687

Su G, Sørensen P and Sorensen D 1997 Inferences about variance components and selection response for body weight in chickens. Genetics Selection Evolution 29: 413- 425

Suzuki K, Kadowaki H, Shibata T, Uchida H and Nishida A 2005 Selection for daily gain, loin-eye area, backfat thickness and intramuscular fat based on desired gains over seven generations of Duroc pigs. Livestock Production Science 97: 193- 202

Theerachai H, Ezzat T and Michael Z 2003 Options for native Chicken (Gallus domesticus) production in Northeastern Thailand. In: Proceedings of the conference on International Agricultural Research for Development. 8-10 October 2003, Göttingen, Germany. http://www.tropentag.de/2003/abstracts/full/166.pdf  site visited on 12/5/2005

Yamada Y, Yokouchi K and Nishida A 1975 Selection index when genetic gains of individual traits are of primary concern. Japanese Journal of Genetics 50 (1): 33-41 http://www.journalarchive.jst.go.jp/jnlpdf.php?cdjournal=ggs1921&cdvol=50&noissue=1&startpage=33&lang=en&from=jnlabstract

Yongolo M G S 1996 Epidemiology of Newcastle disease in village chickens in Tanzania. MVM Thesis, Sokoine University of Agriculture, Tanzania. 234pp.

Received 26 March 2008; Accepted 18 April 2008; Published 6 November 2008

Go to top