SISTEMA DE APOIO NA ESCOLHA DE REPRODUTORES DE BOVINOS DE CORTE AVALIADOS EM PROGRAMAS DE MELHORAMENTO

SUPPORT SISTEM FOR CHOICE OF BEEF SIRES EVALUATED IN GENETIC IMPROVEMENT PROGRAMS

ABSTRACT - The Brazilian beef industry is characterized by extensive exploration, where the low genetic potential is very noticeable and results in incomplete utilization of resources and low productivity. The conduction of selection in beef cattle populations through the use of balanced sires ensures participation in the global market and profitability to production systems. The breeding programs have several sire options in each breed and with many evaluated characteristics. This paper aimed to create a tool, using a logical model, to facilitate and assistance in the sire choice from a sire summary based on the characteristics desired by the user. The model was developed and implemented based on data management through the PostgreSQL (PostgreSQL, 2009) and the PHP programming language that is the basis web development (PHP community, 2009). The incoming data are breed, progeny differences, weaning, yearling, ultrasound, DECAs and accuracies. According to verification tests, the model considers requirements pre-established and launched by user where the list of sires is processed and presented in a restrictive way, facilitating the decision.

---

Introdução

The Brazilian beef industry is characterized for the extensive form of exploration, where stands out the low genetic potential, resulting in an incomplete utilization of resources and low productivity. Factors like market demands and economic globalization require high quality products and competitive costs. Thereby, it is necessary to direct the selection of cattle population inside the production system (FRIES, 1999), using balanced sires, in order to improve the adaptability and ensure participation in the market with economic efficiency. EPDs - Expected Progeny Differences -, are the most used tools in beef cattle genetic improvement. They estimate half the additive genetic value of an individual and are used to compare the merits of the animals in various characteristics with the average values of the analyzed population. (LOBO et al., 2003). The upgrading of the genetic evaluation techniques has been generating increasing number of information, providing EPDs for different characteristics, what has been hampering the decision-making processes of selection (VAL et al; 2008). An alternative to minimize this difficulty is the utilization of the computer science. Black (1993) says that the integration of concepts and knowledge in computational programs, using simulation technics, can improve the use of information directed to the management of the commercial systems of animal production. This allows extraordinary advances in the data analysis and the use of multivariate analytical techniques by researchers (VAL et al., 2008).  Therefore, Tess e Kolstad (2000) say that due to various combinations between the options of management and genetics and the high exigency in monetary rescues and time to evaluate the cattle production systems, the computer simulation is an important tool of evaluation. In this way, the objective of this work was the creation of a tool, through a logic model, of support and facilitation in the moment of the choice of a sire in a summary based in the characteristics wanted by the user.

---

Revisão Bibliográfica

The correct identification of genetically superior animals is determinant for the genetic progress of the population (Everling et al., 2003). Most of the progress tends to happen by the selection of the sires, since they are able to generate a great number of products (BALL & PETERS, 2006). To reach the success in selection of herds it is necessary to have a clean vision of the objects in selection, which can be defined like the combination of genetic values of the characteristics economically important in determined production system (JORGE JR. et al, 2007). Thereby, each producer has a specific program of genetic improvement, according with particular criteria and objects of production. This program should have, according to Lobo (1996), the final objective of translating the genetic value, with expressive improvement of the economic results of the herds. The values, like their levels of importance, attributed to the characteristics that make selection indexes vary with the production systems. Bergmann (2003) says, for example, that for the one who works with the cow/calf operations and produces calves to sell, the weaning weight and the number of calves for cow determine his gross income, and are influenced by the maternal ability, the rates of reproduction and the easy calving. For those who work with the stocking and finishing, the weight gain from weaning to slaughter and characteristics associated to the conformation are more important. The indexes used in the genetic improvement programs are the EPDs - Expected Progeny Differences -, tools that estimate half the additive genetic value of an individual and are used to compare the merits of the animals in various characteristics with the average values of the analyzed population. Predicting, in this way, the ability of genetic transmission of the animal evaluated as progenitor (LOBO et al., 2003). The efficient selection in cattle livestock doesn’t consider just the EPDs of weight gain. Before the aim of excellence in the final product from low production costs, the simultaneous use of characteristics that look for growth, finishing and sexual precocity as selection criteria is necessary (DIBIASI, 2006). In this sense, Fries (2005) says that to acquire a cattle genotype appropriated to a production system of short cycle, is necessary modify genetically the three precocities together, using and taking advantage of the correlation between them. Computer models are useful instruments to the perception of gaps of knowledge and support to the decision, besides learning sources during the development processes. Through them, the growing computerization in modern society suggests great opportunities of data storage and information able to facilitate decisions-making, minimizing the risks to the rural enterprise.

---

Materiais e Métodos

The model was developed and implemented based on data management by PostgreSQL (PostgreSQL, 2009) and the PHP programming language, which is the basis of the The increase in profitability of a herd is only possible when the it is effectively managed, being breeding one enhancer of that increase, since it is not possible to target quality and uniform products if animals are not managed under the genetic point of view (LAZZARINI NETO, 2000). The genetic gain possible to obtain with the responsible use of the information generated by breeding programs, leads to a stable system in every step and is always cumulative in the herd. Genetic progress in the desired direction can only be effectively achieved if selection objectives are clearly defined. There are no recipe improvement programs, or unique recommendations of breed, region and livestock system. The demand of characteristics to be selected is specific to the actual production and market requirement. The features should be chosen in order to increase productivity and usually with common goals, to reduce the age at slaughter, improve the performance and carcass finishing of animals, and also to increase the productive and reproductive efficiency of females without increasing its nutritional requirements, and maintenance costs (CARVALHEIRO, 2008). The importance and magnitude of the features vary by each improvement program, as the producer specific needs. The ANC sire summary is a national survey of Bos taurus breeds breeding, controlled by associations, being an annual publication, with the purpose of guiding the breeders in the choice of sires for use through artificial insemination on the herds. All the sires that have progeny evaluated in controlled herds by PROMEBO are listed, as long as they achieve a minimum number of progeny and accuracy of the information generated. This suggests a large number of animals available for the user, which requires time for correct choice. Fries (2005) states that a production system that wants to compete in the global environment should be based on the foundations of eliminating waste, reducing idleness, and optimizing benefits, costs and time in the execution of tasks. In this sense, the model enables practicality in performing the task. Corroborating, Fox (1990) says that the artificial intelligence and expert systems provide technologies that enable a new approach to solving many decision problems. Costa (1991) emphasizes that the advance in the use of simulation systems makes them important tools to support decision-making in complex systems, assisting producers in evaluating strategic decisions in an operation. The study was based on the creation of a simplified model, very useful to the needs of producers. According to the processes of creation and verification, it is noticeable that it greatly facilitates the choice, since the list of sires becomes very restrict to the user demand as the release of information is predetermined according to its needs. For verification of the model it was simulated the situation of a user who needs Braford sires to mate with heifers, sires that print law weight at birth (EPD WB < - 0.5), and who want accelerated weight gain from birth to weaning (EPD GBW > 5) as input data (Figure 2) and the respective model output (Figure 3). In the previously simulated situation, the user considers isolated needs to weaning, however the system allows simultaneous selection of features in the three camps for a particular breed, facilitating further his choice and featuring animals that possess a set of the demanded EPDs. The determining factor in the sire list is the amount of data released, where the more specific and the greater the number of user-supplied information, the stricter the list processed and presented. Besides the selection by EPD that meet the squad of females and production targets, the system offers the user the choice of the best sires to required characteristics within each race, through DECAs. The filter also allows listing of the most reliable sires, through the camps accuracy for weaning, yearling and ultrasound. The increase in profitability of a herd is only possible when the it is effectively managed, being breeding one enhancer of that increase, since it is not possible to target quality and uniform products if animals are not managed under the genetic point of view (LAZZARINI NETO, 2000). The genetic gain possible to obtain with the responsible use of the information generated by breeding programs, leads to a stable system in every step and is always cumulative in the herd. Genetic progress in the desired direction can only be effectively achieved if selection objectives are clearly defined. There are no recipe improvement programs, or unique recommendations of breed, region and livestock system. The demand of characteristics to be selected is specific to the actual production and market requirement. The features should be chosen in order to increase productivity and usually with common goals, to reduce the age at slaughter, improve the performance and carcass finishing of animals, and also to increase the productive and reproductive efficiency of females without increasing its nutritional requirements, and maintenance costs (CARVALHEIRO, 2008). The importance and magnitude of the features vary by each improvement program, as the producer specific needs. The ANC sire summary is a national survey of Bos taurus breeds breeding, controlled by associations, being an annual publication, with the purpose of guiding the breeders in the choice of sires for use through artificial insemination on the herds. All the sires that have progeny evaluated in controlled herds by PROMEBO are listed, as long as they achieve a minimum number of progeny and accuracy of the information generated. This suggests a large number of animals available for the user, which requires time for correct choice. Fries (2005) states that a production system that wants to compete in the global environment should be based on the foundations of eliminating waste, reducing idleness, and optimizing benefits, costs and time in the execution of tasks. In this sense, the model enables practicality in performing the task. Corroborating, Fox (1990) says that the artificial intelligence and expert systems provide technologies that enable a new approach to solving many decision problems. Costa (1991) emphasizes that the advance in the use of simulation systems makes them important tools to support decision-making in complex systems, assisting producers in evaluating strategic decisions in an operation. The study was based on the creation of a simplified model, very useful to the needs of producers. According to the processes of creation and verification, it is noticeable that it greatly facilitates the choice, since the list of sires becomes very restrict to the user demand as the release of information is predetermined according to its needs. For verification of the model it was simulated the situation of a user who needs Braford sires to mate with heifers, sires that print law weight at birth (EPD WB < - 0.5), and who want accelerated weight gain from birth to weaning (EPD GBW > 5) as input data (Figure 2) and the respective model output (Figure 3). In the previously simulated situation, the user considers isolated needs to weaning, however the system allows simultaneous selection of features in the three camps for a particular breed, facilitating further his choice and featuring animals that possess a set of the demanded EPDs. The determining factor in the sire list is the amount of data released, where the more specific and the greater the number of user-supplied information, the stricter the list processed and presented. Besides the selection by EPD that meet the squad of females and production targets, the system offers the user the choice of the best sires to required characteristics within each race, through DECAs. The filter also allows listing of the most reliable sires, through the camps accuracy for weaning, yearling and ultrasound. development of Web (PHP community, 2009). This is a conceptual model, empiric, static and determinist, which processes information by filters in the database of the improvement programs sire catalogs. The description is done as the conceptual model presented in the Figure 1. According to Figure 1, the input data of the model feed a processor which, by the filtering of information, generates an output list, with the sires that meet the criteria set by the user. The input data are breed, performance registers of the sires, expressed by EPD, values that are estimated to characteristics of economic significance, available in the sire catalogs of the improvement programs, DECAS for each characteristic and accuracy of weaning, yearling and ultrasound. DECA’s are classifications of the individuals based in each EPD and patterned for each breed, being representation of the animals in classes of 10% and allow the quick visualization and classification objective of the EPD of a determined sire in relation with the others. DECA 1 presupposes that sire 1 is between the 10% superior, DECA 2 presupposes that that he is among the 20% superiors and this way successively. The accuracy (Ac) indicates the reliability of determined sire evaluation. The values of accuracy may vary from zero to one (0 to 1.0), more elevated numbers indicate higher level of security in the evaluation of the estimates. The inclusion of new EPDs and its variations for one same animal from one year to other by the programs is constant, in virtue of the improvement progress in some breeds, the variations in the amount of evaluated animals and the connectability of the herds. Therefore, during the implementation of this model, it was used the ones that contained in 2011 PROMEBO (Programa de Melhoramento de Bovinos de Carne) summaries, with average of 20 characteristics from 1200 sires of the breeds Aberdeen Angus, Brangus, Hereford, Braford, Devon, Charolais and Shorthorn, from 29.839 evaluations at weaning  and 8.078 at yearling (post weaning) in 215 properties. By this way, the update of the data that fed the system should be done annually as the publication of the summaries. In order to facilitate understanding of the variables, a list with the definition of acronyms used in the text, in the conceptual model and in the system was created (Table 1). The characteristics used in the weaning filter are those adjusted to 205 day. Birth weight (BW) is used to identify the sires that can cause problems at the calving and to calculate the weight gain until weaning. The birth and weaning weight are not determined just by the individual performance, but also by the maternal environment, mostly represented by the milk yield and maternal ability (MEYER, 1992). Thus, maternal ability depends on the mother's dedication to raise the calf and her capacity to produce milk, indicating the ability of the sire to transmit genes to their daughters with a direct effect for growth and genes related to maternal ability, which will affect in the gain from birth to weaning (GBW) of his grandsons. The selection to GBW allows the transition of genes with direct effect upon the growth velocity from birth to weaning (205 days) of the progeny. This EPD highlights the most promising animals in performance and feed conversion. Conformation characteristics (Conformation, Precocity, Muscling and Size) are formed by a set of EPD’s to select balance and productivity of the herd. The animals are evaluated normally at weaning (205 days) and at post weaning (365 or 550 days) through visual scores, in which higher values suggest the presence a gene remarkable for the characteristic. Through these scores, we seek to evaluate the potential production of carcasses adequate to the meat production. Based on the observation of precocity, the capacity of finishing with smallest weight and age is evaluated. By the score of muscling, the muscular development of the animal is evaluated. The EPD C (carcass) estimates the amount of meat in the carcass. The evaluation is done by imagining the slaughtered animal. This characteristic is influenced by size (especially the length) and by muscularity degree. The EPD P (precocity) indicates the ability of the animal to reach a minimum degree of size in the carcass, with live weight not elevated, which is fixed by the market. The EPD M (muscling) evaluates the development of muscle mass as a whole, observed at points like the as the forearm, shoulder, loin, rump, and especially in the back. The S (size) is an indication of height and length of the animal. Characteristics of yearling are EPDs evaluated after weaning and adjusted to 550 days. The weight gain from waning to yearling (GWY) predicts the growth rate from weaning (205 days) to yearling (550 days). Weight gain from birth to weaning (GBY) indicates the growth rate from birth to yearling or (365 or 550 days). This information is obtained by summing the EPDs weight gain from birth to weaning and weight gain from weaning to yearling. The DEP SC (scrotal circumference)  is expressed in centimeters (cm) and is used to identify the sires that effectively can improve the herd, which highlights the correlation of SC characteristics with sexual precocity of females, sperm production, and also by been another body measurement of the growth velocity. Ultrasound indexes are those predicted by evaluating the quality of the carcass with the ultrasound technique. According to Yokoo et al. (2009), the methodology allows the assessment of carcass characteristics by a noninvasive procedure, leaving no residues in meat. It is considered a low-cost technology and it is easy to use when compared to evaluations carried out directly on the carcass after slaughter. Among these characteristics, REA (Rib eye area)  is based on measurements obtained in the anatomical site of the intercostal space between the 12th and 13th ribs measuring the total area of longissimus dorsi muscle and it is presented in square centimeters (cm²). A positive EPD indicates animals who can produce offspring with a higher yield percentage of retail cuts, since there is a correlation between these and the REA. The EPD SFT (subcutaneous fat thinckness) is estimated by the same methods of the REA, but it is given in millimeters and estimates finishing. The genetic indexes (weaning, carcass and final) are single values ​​calculated from weights of EPDs aggregated and express the full merit of the animal. The model works through filters to values ​​of EPDs, DECAs and accuracies, from the launch of minimum and/or maximum, as criteria to form a list of sires, of the breed chosen as input to the model, that attend the characteristics pre-determined by the user, been this list, the output of the model. The model allows the user to choose, for all the features that will be taken into consideration, options less than (<), greater than (>), equals (=), greater than or equal (≥) and smaller and equal (≤) of the data elected by him in each variable, allowing all restrictions according to his need.

---

Resultados e Discussão

The increase in profitability of a herd is only possible when the it is effectively managed, being breeding one enhancer of that increase, since it is not possible to target quality and uniform products if animals are not managed under the genetic point of view (LAZZARINI NETO, 2000). The genetic gain possible to obtain with the responsible use of the information generated by breeding programs, leads to a stable system in every step and is always cumulative in the herd. Genetic progress in the desired direction can only be effectively achieved if selection objectives are clearly defined. There are no recipe improvement programs, or unique recommendations of breed, region and livestock system. The demand of characteristics to be selected is specific to the actual production and market requirement. The features should be chosen in order to increase productivity and usually with common goals, to reduce the age at slaughter, improve the performance and carcass finishing of animals, and also to increase the productive and reproductive efficiency of females without increasing its nutritional requirements, and maintenance costs (CARVALHEIRO, 2008). The importance and magnitude of the features vary by each improvement program, as the producer specific needs. The ANC sire summary is a national survey of Bos taurus breeds breeding, controlled by associations, being an annual publication, with the purpose of guiding the breeders in the choice of sires for use through artificial insemination on the herds. All the sires that have progeny evaluated in controlled herds by PROMEBO are listed, as long as they achieve a minimum number of progeny and accuracy of the information generated. This suggests a large number of animals available for the user, which requires time for correct choice. Fries (2005) states that a production system that wants to compete in the global environment should be based on the foundations of eliminating waste, reducing idleness, and optimizing benefits, costs and time in the execution of tasks. In this sense, the model enables practicality in performing the task. Corroborating, Fox (1990) says that the artificial intelligence and expert systems provide technologies that enable a new approach to solving many decision problems. Costa (1991) emphasizes that the advance in the use of simulation systems makes them important tools to support decision-making in complex systems, assisting producers in evaluating strategic decisions in an operation. The study was based on the creation of a simplified model, very useful to the needs of producers. According to the processes of creation and verification, it is noticeable that it greatly facilitates the choice, since the list of sires becomes very restrict to the user demand as the release of information is predetermined according to its needs. For verification of the model it was simulated the situation of a user who needs Braford sires to mate with heifers, sires that print law weight at birth (EPD WB < - 0.5), and who want accelerated weight gain from birth to weaning (EPD GBW > 5) as input data (Figure 2) and the respective model output (Figure 3). In the previously simulated situation, the user considers isolated needs to weaning, however the system allows simultaneous selection of features in the three camps for a particular breed, facilitating further his choice and featuring animals that possess a set of the demanded EPDs. The determining factor in the sire list is the amount of data released, where the more specific and the greater the number of user-supplied information, the stricter the list processed and presented. Besides the selection by EPD that meet the squad of females and production targets, the system offers the user the choice of the best sires to required characteristics within each race, through DECAs. The filter also allows listing of the most reliable sires, through the camps accuracy for weaning, yearling and ultrasound.

---

Conclusões

Simulation models, besides representing a growing space in lines of research and industry, arise like alternatives of management and important tools to help in the decision making in several areas of the complex scenarios of production in a beef operation. According to the stages of creation and verification, the model developed to the sire choice presents itself like a simple and practical tool, satisfying the pre-established needs and lanced by the user, where the information processed in a list of sires is presented in a restrict way, serving to support the decision.

---

Gráficos e Tabelas

---

Referências

BALL, P.J.H., PETERS, A.R. Reprodução em bovinos. São Paulo: Roca, 2006. BERGMANN, J.A.G. Workshop Seleção em Bovinos de Corte. UFERSA, 2003. Disponível em: www.ufersa.edu.br, acesso em 15 de junho de 2011. BLACK, J.L. et al. Rol of computer simulation in the applications of knowledge to animal industries. Australian Animal Journal of Agricultural Research, v.44, p.541-555, 1993. CARVALHEIRO, R., CAVALCANTI, J.R. O biotipo dos animais do seu rebanho está de acordo com seu objetivo de seleção? In:VII Simpósio Brasileiro de Melhoramento Animal. SP, 2008. COSTA, G.Z. et al. Estimativas de parâmetros genéticos e fenotípicos de escores visuais e de ganho médio de peso do nascimento a desmama de bovinos formadores da raça brangus. Arq Veterinária, Jaboticabal, SP ,v.24, n.3, 172-176, 2008. COSTA, M.A.B. Uma abordagem sobre inteligência artificial e simulação, com uma aplicação na pecuária de corte nacional. Produção, Rio de Janeiro, VoI. 2, Nº1 – p.51-59, 1991. DIBIASI, N.F. Estudo do crescimento, avaliação visual, medidas por ultra-sonografia e precocidade sexual em touros jovens pertencentes a vinte e uma raças com aptidão para corte. Fac. Cien. Agr. e Vet., Campus de Jaboticabal, SP, 2006. DIBIASI, N.F. et al. Proposta de metodologia para a conversão do escore de avaliação visual de conformação em escore de estrutura, em bovinos de corte. Arq Veterinaria, Jaboticabal, SP, v.26, n.2, 082-087, 2010. EVERLING, D.M. et al. Estimativas de herdabilidade e correlação genética para características de crescimento na fase de pré-desmama e medidas de perímetro escrotal ao sobreano em bovinos Angus-Nelore. Rev. Bras. Zootec., 30(6S):2002-2008, 2001. FONTOURA JUNIOR, J.A.S. et al. Utilização de modelos de simulação em sistemas de produção de bovinos de corte. Vet. e Zootec., v.14, n.1, jun., p. 19-30, 2007. FRIES, L.A. Genética de gado de corte orientada para lucratividade. In: LOBATO, J.F.P.; BARCELLOS, J.O.J.; KESSLER, A.M. Produção de Bovinos de Corte. Porto Alegre: EDIPUCRS, 1999. p 193-233. FRIES, L.A. Avanços do uso dos recursos genéticos e biotécnicas reprodutivas com vistas ao melhoramento do gado de corte. In:I SIMBOI- Simpósio sobre desafios e novas tecnologias na bovinocultura de corte. UPIS: Departamento de Zootecnia. Brasília, 2005. FOX, M.S. AI and Expert Systems Myths, Legends and Facts Al Magazine, p.8-20, Fev 1990. JORGE JR., J. et al. Objetivos de seleção e valores econômicos em sistemas de produção de gado de corte no Brasil. R. Bras. Zootec., v.36, n.5, p.1549-1558, 2007. LAZZARINI NETO, S. Reprodução e melhoramento genético. 2 ed. Viçosa: Aprenda fácil, 2000. LOBATO, J.F.P. Considerações efetivas sobre seleção, produção e manejo para maior produtividade dos rebanhos de cria. In: LOBATO, J.F.P.; BARCELLOS, J.O.J.; KESSLER, A.M. Produção de Bovinos de Corte. Porto Alegre: EDIPUCRS, p 235-285, 1999. LOBO, R. B. Programa de Melhoramento Genético da Raça Nelore. USP. Ribeirão Preto, 1996. LOBO, R.B. et al. Avaliação genética de animais jovens touros e matrizes. Ribeirão Preto. GEMAC/FMRP/USP, 2003. 86p. MARCONDES, C. R. et al. Análise de alguns critérios de seleção para características de crescimento na raça Nelore. Arq. Bras. Med. Vet. Zootec. vol.52 n.1 Belo Horizonte, 2000 MAYER, K. Variance components due to direct and maternal effects for growth traits of Australian beef cattle, Livest. Prod. Sci., 31(3-4):179-202. 1992 PostgreSQL Software Livre – Software de desenvolvimento. disponível em: http://www.postgresql.org/, acesso em 31 de julho de 2009. PHP community - Software livre de linguagem de programação. Disponível em: http://www.php.net/, acesso em 31 de julho de 2009. SILVEIRA, V.C.P. Pampa corte – um modelo de simulação para o crescimento e engorda de gado de corte. Ciência Rural, v. 32, n.3, 2002. TESS, M.W.; KOLSTAD, B.W. Simulation of cow-calf production system in a range environment: II. Model development. Journal of Animal Science, 78:1159-1169. 2000. TREVISAN, N.B. et al. Desempenho de bovinos simulado pelo modelo Pampa Corte e obtido por experimentação. Ciência Rural, v.39, n.1, 2009 VAL, J.E. et.al. Alternativas para seleção de touros da raça Nelore considerando características múltiplas de importância econômica. Arq. Bras. Med. Vet. Zootec., v.60, n.3, p.705-712, 2008. YOKOO, M.J.I., et al. Correlações genéticas entre escores visuais e características de carcaça medidas por ultrassom em bovinos de corte. Pesq. agropec. bras., Brasília, v.44, n.2, p.197-202, fev. 2009.