A carregar...
Publicação
Modelo de previsão da persistência de células somáticas elevadas (>200 000 células/ml) durante o período seco a partir de contagens da lactação anterior em vacas Holstein-Frísia
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.26 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Atualmente, a produção leiteira enfrenta diversos constrangimentos. A mastite assume especial relevância. Apresenta prejuízos diretos na produção e um impacto negativo na qualidade do leite. Esta pode originar alterações físicas, químicas, bacteriológicas e organoléticas no leite. O presente estudo teve como objetivo desenvolver um modelo preditivo para calcular a probabilidade de cura ou persistência de mastites durante o período seco em vacas leiteiras de raça Holstein-Frísia. Para apoiar a toma de decisões informadas no maneio individual dos animais. Para tal, utilizaram-se dados referentes a 366 lactações completas. Estas apresentavam valores superiores a 200 000 células/ml na última contagem de células somáticas da lactação. A definição deste ponto de corte baseou-se no seu reconhecimento generalizado como limiar para a identificação de mastite subclínica em vacas leiteiras. Construiu-se uma tabela contendo todos os dados obtidos da exploração. A partir dos quais se criaram variáveis potencialmente relevantes para o desenvolvimento do modelo preditivo. Incluindo a variável “Grupo”, que permitiu classificar cada animal em dois grupos, cura ou persistência, com base no valor da primeira contagem de células somáticas após o parto. Também se procedeu ao cálculo dos Linear Scores, que são a transformação logarítmica dos valores das contagens de células somáticas. De seguida, utilizou-se o SAS (SAS 9.4, SAS Institute Inc., Cary, NC), um programa de análise estatística. Para determinar as variáveis que teriam maior capacidade discriminativa entre os dois grupos e realizar a construção, treino e teste do modelo preditivo. As variáveis selecionadas foram: o número de contrastes com um linear score superior a 4 ao longo de toda a lactação e nos últimos 5 meses da mesma; a produção leiteira em Kg ao último contraste; a duração da lactação; e o número de lactações por vaca. Contudo, estas só capturaram parcialmente os fatores determinantes da cura ou persistência da mastite, sendo capazes de explicar 12% da variabilidade dos dados. O modelo preditivo obteve uma exatidão de 62,30% e um erro estimado de 33,62%. Este revelou ser insuficientemente robusto para aplicação prática isolada. No entanto, poderá servir de ponto de partida para o desenvolvimento de ferramentas que permitam realizar esta previsão
Currently, dairy production faces several constraints. Mastitis holds particular significance. This is not only due to direct economic losses in production, but also due to its detrimental effect on milk quality. It can induce physical, chemical, bacteriological, and organoleptic alterations. The present study aimed to develop a predictive model to estimate the probability of cure or persistence of mastitis during the dry period in Holstein-Friesian dairy cows, thereby supporting informed decision-making in individual animal management. For this purpose, data from 366 complete lactations were used, in which the last somatic cell count exceeded 200 000 cells/mL. The definition of this threshold was based on its widely recognized status as the cutoff point for identifying subclinical mastitis in dairy cows. A comprehensive dataset was constructed containing all collected information. From this, potential predictive variables for model development were derived. One of these variables, “Group,” enabled classification of each cow into either cure or persistence, based on the first somatic cell count value after calving. In addition, Somatic Cell Scores were calculated as the logarithmic transformation of somatic cell count values. Data analysis was performed using SAS statistical software (SAS 9.4, SAS Institute Inc., Cary, NC, USA). This allowed identification of variables with the greatest discriminative capability between the two groups and construction, training, and testing of the predictive model. The selected variables were the number of test days with a somatic cell score above 4 throughout the entire lactation and within its last five months, milk yield (kg) at the last test day, lactation length, and the number of lactations per cow. However, these variables captured only part of the factors influencing the cure or persistence of mastitis, explaining approximately 12% of the data variability. The predictive model achieved an accuracy of 62,30% and an estimated error rate of 33,62%, indicating that it is not yet sufficiently robust for independent practical application. Nevertheless, it may serve as a starting point for the development of more effective tools to perform this prediction
Currently, dairy production faces several constraints. Mastitis holds particular significance. This is not only due to direct economic losses in production, but also due to its detrimental effect on milk quality. It can induce physical, chemical, bacteriological, and organoleptic alterations. The present study aimed to develop a predictive model to estimate the probability of cure or persistence of mastitis during the dry period in Holstein-Friesian dairy cows, thereby supporting informed decision-making in individual animal management. For this purpose, data from 366 complete lactations were used, in which the last somatic cell count exceeded 200 000 cells/mL. The definition of this threshold was based on its widely recognized status as the cutoff point for identifying subclinical mastitis in dairy cows. A comprehensive dataset was constructed containing all collected information. From this, potential predictive variables for model development were derived. One of these variables, “Group,” enabled classification of each cow into either cure or persistence, based on the first somatic cell count value after calving. In addition, Somatic Cell Scores were calculated as the logarithmic transformation of somatic cell count values. Data analysis was performed using SAS statistical software (SAS 9.4, SAS Institute Inc., Cary, NC, USA). This allowed identification of variables with the greatest discriminative capability between the two groups and construction, training, and testing of the predictive model. The selected variables were the number of test days with a somatic cell score above 4 throughout the entire lactation and within its last five months, milk yield (kg) at the last test day, lactation length, and the number of lactations per cow. However, these variables captured only part of the factors influencing the cure or persistence of mastitis, explaining approximately 12% of the data variability. The predictive model achieved an accuracy of 62,30% and an estimated error rate of 33,62%, indicating that it is not yet sufficiently robust for independent practical application. Nevertheless, it may serve as a starting point for the development of more effective tools to perform this prediction
Descrição
Dissertação de Mestrado Integrado em Medicina Veterinária, área científica de Clínica
Palavras-chave
Contagem de células somáticas Holstein-Frísia Mastite Período seco Vaca Cow Dry period Holstein-Friesian Mastitis Somatic cell count