Browsing by Author "Dallago, Gabriel Machado"

Now showing 1 - 3 of 3

Avaliação de modelos para predizer a produção de leite no primeiro controle leiteiro de primíparas
(UFVJM, 2018) Dallago, Gabriel Machado; Figueiredo, Darcilene Maria de; Andrade, Paulo César de Resende; Santos, Roseli Aparecida dos; Universidade Federal dos Vales do Jequitinhonha e Mucuri (UFVJM); Figueiredo, Darcilene Maria de; Santos, Roseli Aparecida dos; Verardo, Lucas Lima; Glória, Leonardo Siqueira; Andrade, Paulo César de Resende
O período de transição é marcado por profundas alterações morfofisiológicas no corpo do gado leiteiro. A maioria dos distúrbios metabólicos e doenças infecciosas são diagnosticados durante o periparto, comprometendo o sucesso de toda a lactação. Formas de se avaliar o manejo de transição de multíparas têm sido propostas, muito embora o mesmo não tenha sido observado para novilhas. O objetivo do trabalho foi desenvolver um modelo capaz de estimar a produção de leite do primeiro controle de vacas primíparas (P_LEITE) com base em variáveis rotineiramente coletadas nos controles leiteiros, de tal forma que pudesse ser utilizado na avaliação do período de transição de animais nessa categoria. Para tal, foram utilizadas informações do primeiro controle de 3267 primíparas da raça Holandesa, realizados entre 2014 e 2017. As informações foram alocadas em arquivos de treino (n = 2345) e validação (n = 780), posteriormente utilizados na estimativa dos parâmetros e avaliação dos modelos, respectivamente. Variáveis relacionadas com a P_LEITE foram identificadas através da regularized regression por meio da elastic net. Três tipos de modelos foram avaliados: regressão linear múltipla (RLM), random forest (RF) e rede neural artificial (RNA). Os modelos resultantes foram avaliados com base em seis estatísticas de ajuste e 10-fold cross validation. Além disso, foram calculados os coeficientes de correlação de Pearson (r) e de concordância de Lin (ρ) entre os valores observados e preditos pelos modelos, assim como a comparação do valor mediano observado e predito de P_LEITE, adotando 5% como nível de significância. Os três modelos avaliados apresentaram capacidade preditiva para a P_LEITE, com erro absoluto médio abaixo de 4,00 kg no arquivo de validação e 10-fold cross validation. Não foi observada diferença significativa entre os valores medianos preditos e observados. A RNA foi o modelo que apresentou a melhor qualidade de ajuste comparada com a RLM e RF. Trabalhos futuros poderão estudar as possíveis variações existentes entre rebanhos, como forma de melhorar a capacidade preditiva da RNA.
Intake, digestibility, microbial protein production, and nitrogen balance of lambs fed with sorghum silage partially replaced with dehydrated fruit by-products
(Springer, 2019-03) Almeida, Júlio César Silva Almeida; Figueiredo, Darcilene Maria de; Azevedo, Katharine Kelly de; Ribeiro, Enilson Geraldo; Dallago, Gabriel Machado; Universidade Federal dos Vales do Jequitinhonha e Mucuri; Instituto de Zootecnia (SP)
The objective of this study was to evaluate intake, digestibility, microbial protein production, and nitrogen balance of lambs fed with different dehydrated fruit residues replacing sorghum silage. The experiment was conducted using twenty-five mixed-breed male lambs (20.64 ± 4.20 kg) with an average age of 8 months. Treatments consisted of sorghum silage and 75% (DM basis) replacement of sorghum silage for dehydrated pineapple, banana, mango, or passion fruit by-products. The experiment was conducted in a completely randomized design with five treatments and five repetitions. Results were subjected to one-way analysis of variance and means were compared using the Tukey test at 5% of probability. Daily nutrient intake did not differ (P>0.05) between the different treatments evaluated regardless of how it was expressed (kg/day or % body weight [BW]), except for NDFap (% BW) that was higher (P<0.05) in the diet with passion fruit residue than in diets with banana and mango residues. The digestibility of ethereal extract was higher (P<0.05) for diets with sorghum silage, pineapple, and passion fruit compared to the diet with banana residue. The digestibility of neutral detergent fiber corrected for ashes and protein was higher (P<0.05) on animals receiving sorghum silage than diets with mango and passion fruit residues. No difference (P>0.05) was observed for purine derivative excretion, microbial efficiency, nitrogen intake, nitrogen loss (urine and faeces), and nitrogen balance between diets. In conclusion, dehydrated fruit by-products (pineapple, banana, mango, and passion fruit) are good options for partial replacement of sorghum silage (75%) and potentially reduce feeding costs.
The phenotypic relationship between residual intake and gain and other feed efficiency traits in Nellore cattle
(2019-02) Figueiredo, Darcilene Maria de; Mercadante, Maria Eugênia Zerlotti; Pires, Aldrin Vieira; Branco, Renata Helena; Ribeiro, Enilson Geraldo; Dallago, Gabriel Machado; Schorer, Marianne; Rocha, Pedro Ribeiro
This study aimed to compare feed efficiency measures of Nellore beef cattle on different residual intake and gain (RIG) classes. We used data from 610 animals weighing on average 236.33 kg and average of 283 days of age from feedlot performance tests carried out between 2005 and 2012. Animals were grouped based on RIG into three different classes: high RIG (> mean + 0.5 standard deviation (SD), most efficient; n = 193), medium RIG (mean ± 0.5 SD; n = 235), and low RIG (< mean - 0.5 SD, least efficient; n = 182). Residual feed intake (RFI), residual gain (RG), feed conversion ratio (FCR), feed efficiency (FE), relative growth rate (RGR), and Kleiber ratio (KR) of animals in each RIG class were compared by Tukey test at 1% of probability. Phenotypic correlations between variables were evaluated as well. Animals on high RIG class showed lower dry matter intake (P < 0.01) and higher average daily gain (P < 0.01) than low RIG animals. Consequently, high RIG animals had lower FCR (P < 0.01) and higher FE (P < 0.01) than those animals in low RIG class. The most efficient animals based on RIG were also the most efficient animals based on RG and RFI. RIG was negatively correlated to dry matter intake (P < 0.01) and FCR (P < 0.01), and a positive correlation was found between RIG and FE (P < 0.01). Therefore, RIG appears to be a good parameter to select animals with reduced dry matter intake and high productive performance.