The use of Artificial Intelligence in predicting Respiratory Syncytial Virus-inhibiting flavonoids / O uso da Inteligência Artificial na predição de flavonoides inibidores do Vírus Sincicial Respiratório

Human Respiratory Syncytial Virus (hRSV) infection results in death and hospitalization of thousands of people worldwide each year. Unfortunately, there are no vaccines or specific treatments for hRSV infections. Screening hundreds or even thousands of promising molecules is a challenge for science. We integrated biological, structural, and physicochemical properties to train and to apply the concept of artificial intelligence (AI) able to predict flavonoids with potential anti-hRSV activity. During the training and simulation steps, the AI produced results with hit rates of more than 83%. The better AIs were able to predict active or inactive flavonoids against hRSV. In the future, in vitro and/or in vivo evaluations of these flavonoids may accelerate trials for new anti-RSV drugs, reduce hospitalizations, deaths, and morbidity caused by this infection worldwide, and be used as input in these networks to determine which parameter is more important for their decision.

A infecção pelo Vírus Sincicial Respiratório Humano (hRSV) resulta na morte e hospitalização de milhares de pessoas em todo o mundo a cada ano. Infelizmente, não existem vacinas ou tratamentos específicos para tais infecções. A testagem de centenas, ou mesmo milhares, de moléculas promissoras é um desafio para a ciência. Neste trabalho, nós integramos propriedades biológicas, estruturais e físico-químicas para treinar e aplicar o conceito de inteligência artificial (IA) capaz de prever flavonoides com potencial atividade anti-hRSV. Durante as etapas de treinamento e simulação, a IA produziu resultados com taxas de acerto superiores a 83%, sendo capaz de prever flavonoides ativos ou inativos contra o hRSV. No futuro, avaliações in vitro e/ou in vivo desses flavonoides poderão acelerar os testes de novas drogas anti-RSV, reduzir hospitalizações, mortes e morbidade causadas por essa infecção. Além disso, a validação futura destes dados poderá determinar qual parâmetro tem maior peso na decisão da inteligência.

The embryo non-invasive pre-implantation diagnosis era: how far are we?

Advancements in assisted reproduction (AR) methodologies have allowed significant improvements in live birth rates of women who otherwise would not be able to conceive. One of the tools that allowed this improvement is the possibility of embryo selection based on genetic status, performed via preimplantation genetic testing (PGT). Even though the widespread use of PGT from TE biopsy helped to decrease the interval from the beginning of the AR intervention to pregnancy, especially in older patients, in AR, there are still many concerns about the application of this invasive methodology in all cycles. Therefore, recently, researchers started to study the use of cell free DNA (cfDNA) released by the blastocyst in its culture medium to perform PGT, in a method called non-invasive PGT (niPGT). The development of a niPGT would bring the diagnostics power of conventional PGT, but with the advantage of being potentially less harmful to the embryo. Its implementation in clinical practice, however, is under heavy discussion since there are many unknowns about the technique, such as the origin of the cfDNA or if this genetic material is a true representative of the actual ploidy status of the embryo. Available data indicates that there is high correspondence between results observed in TE biopsies and the ones observed from cfDNA, but these results are still contradictory and highly debatable. In the present review, the advantages and disadvantages of niPGT are presented and discussed in relation to tradition TE biopsy-based PGT. Furthermore, there are also presented some other possible non-invasive tools that could be applied in the selection of the best embryo, such as quantification of other molecules as quality biomarkers, or the use artificial intelligence (AI) to identify the best embryos based on morphological and/or morphokitetic parameters.(AU)

Artificial neural networks for the management of poultry industry: a simulation based on the broiler production chain / Redes neurais artificiais para o gerenciamento da indústria avícola: uma simulação baseada na cadeia de produção de frangos de corte

The aim of this study was to predict production indicators and to determine their potential economic impact on a poultry integration system using artificial neural networks (ANN) models. Forty zootechnical and production parameters from broiler breeder farms, one hatchery, broiler production flocks, and one slaughterhouse were selected as variables. The ANN models were established for four output variables: "saleable hatching", "weight at the end of week 5," "partial condemnation," and "total condemnation" and were analyzed in relation to the coefficient of multiple determination (R2), correlation coefficient (R), mean error (E), mean squared error (MSE), and root mean square error (RMSE). The production scenarios were simulated and the economic impacts were estimated. The ANN models were suitable for simulating production scenarios after validation. For "saleable hatching", incubator and egg storage period are likely to increase the financial gains. For "weight at the end of the week 5" the lineage (A) is important to increase revenues. However, broiler weight at the end of the first week may not have a significant influence. Flock sex (female) may influence the "partial condemnation" rates, while chick weight at first day may not. For "total condemnation", flock sex and type of chick may not influence condemnation rates, but mortality rates and broiler weight may have a significant impact.

O objetivo deste trabalho foi predizer os indicadores de produção e determinar o seu potencial impacto econômico em um sistema de integração utilizando as redes neurais artificiais (RNA). Quarenta parâmetros zootécnicos e de produção de granjas de matrizes e de frango de corte, um incubatório e um abatedouro foram selecionados como variáveis. Os modelos de RNA foram estabelecidos para quatro variáveis de saída ("eclosão vendável", "peso ao final da quinta semana", "condenações parciais" e "condenações totais") e foram analisados em relação ao coeficiente de determinação múltipla (R2), coeficiente de correlação (R), erro médio (E), erro quadrático médio (EQM) e raiz do erro quadrático médio (REQM). Os cenários produtivos foram simulados e os impactos foram estimados. Os modelos de RNA gerados foram adequados para simular diferentes cenários produtivos após o treinamento. Para "eclosão vendável", o modelo de incubadora e o período de incubação aumentaram os ganhos financeiros. Para "peso ao final da quinta semana", a linhagem também demonstrou influencia no retorno financeiro, o que não aconteceu com o peso ao final da primeira semana. O sexo do lote possui influência nas taxas de "condenação parcial", ao contrário do peso do frango no primeiro dia. As taxas de mortalidade e o peso do frango apresentaram influência na "condenação total", mas o sexo do lote e o tipo de pinto não tiverem influência.

Artificial Neural Networks to Predict Egg-Production Traits in Commercial Laying Breeder Hens

In recent years, egg production has had an intense growth in Brazil, and Brazilian egg consumption per capita has significantly increased in the last decade. To reduce sanitary and financial risks, decisions regarding the production and health status of the flock must be made based on objective criteria. Our aim was to determine the main "input" variables for the prediction of egg production performance in commercial laying breeder flocks using an ANN model. The software NeuroShellClassifier and NeuroShell Predictor were used to build the ANN. A total of 26 egg-production traits were selected as input variables and eight as output variables. A database of 44,120 Excel cells was generated. For the training and validation of the models, 74.9% and 25.1% of the data were used, respectively. The accuracy of the ANN models was calculated and compared using the analysis of coefficient of multiple determination (R2), mean squared error (MSE), and an assessment of uniform scatter in the residual plots. The models for the outputs "weekly egg production," "weekly incubated egg,", "accumulated commercial egg," and "viability" showed an R2 greater than 0.8. Other models yielded R2 values lower than 0.8. The ANN predicts adequately eight egg-production traits in the breeders of commercial laying hens. The method is an option for data management analysis in the egg industry, providing estimates of the relative contribution of each input variable to the outputs.(AU)

Artificial intelligence applied to enzymatic hydrolysis of lactose: improving the control of industrial processes / Inteligência artificial aplicada a hidrólise enzimática da lactose: melhorando o controle dos processos industriais

Lactose is the main carbohydrate in milk, and its absorption occurs via enzymatic hydrolysis, generating glucose and galactose. Lactose intolerance is the reduction of intestinal hydrolysis capacity due to hypolactasia, which results in the need to consume dairy foods with low levels of this carbohydrate. ß-galactosidase enzymes are used in dairy industries to hydrolyze lactose, thereby allowing intolerant consumers access to dairy products without the negative health implications. Alternative and official analytical methods are used to quantify the carbohydrate content resulting from enzymatic hydrolysis. The objective of this study was to evaluate the enzymatic hydrolysis of two distinct industrial enzymes produced by the microorganisms Bacillus licheniformis and Kluyveromyces lactis using three analytical methods: enzymatic method, cryoscopy, and high performance liquid chromatography (HPLC) using artificial intelligence to improve the control of the industrial processes. After adding the enzymes to skim milk, time kinetics was performed by collecting samples at time 0, every 10 min for 1 h, and every 30 min until the end of 5 h of hydrolysis. In 97% of the cases, a decrease in lactose concentration was observed by HPLC, followed by the deepening of the cryoscopic point. Glucose measurements by absorbance and HPLC quantification were correlated (r = 0.79; p < 0.01) but not concordant (p < 0.01). It was concluded that by means of artificial intelligence, it was possible to indirectly estimate lactose concentration using an algorithm that associates cryoscopy and glucose concentration.(AU)

O principal carboidrato do leite é a lactose e a sua absorção ocorre devido à hidrólise enzimática, gerando glicose e galactose. A intolerância à lactose é a redução da capacidade de hidrólise intestinal devido à hipolactasia, gerando a necessidade do consumo de alimentos lácteos com baixo teor deste carboidrato. As enzimas ß-galactosidase são utilizadas nas indústrias de laticínios para hidrolisar a lactose, proporcionando ao consumidor intolerante a possibilidade de ingerir os produtos lácteos sem prejuízos à saúde. Para quantificar o conteúdo de carboidratos resultante da hidrólise enzimática, são utilizados métodos analíticos alternativos e oficiais. O objetivo deste estudo foi avaliar a hidrólise enzimática de duas enzimas industriais distintas produzidas pelos microrganismos Bacillus licheniformis e Kluyveromyces lactis, por meio de três métodos analíticos: método enzimático, crioscopia e HPLC. A inteligência artificial foi utilizada para melhorar o controle dos processos industriais. Após a adição das enzimas ao leite desnatado, foi realizada a cinética de tempo coletando as amostras no tempo 0, a cada 10 minutos, até completar 1 hora de reação, e a cada 30 minutos até serem atingidas 5 horas de reação de hidrólise. Em 97% dos casos, a diminuição da concentração de lactose por HPLC acompanhou o aprofundamento do ponto crioscópico. As medições de glicose por absorbância e HPLC foram correlacionadas (r = 0,79; p < 0,01), mas não concordantes (p < 0,01). Concluiu-se que, por meio da inteligência artificial, é possível estimar indiretamente a concentração de lactose a partir de um algoritmo que associa a crioscopia e a concentração de glicose.(AU)

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.(AU)

Artificial neural networks on eggs production data management

Background: Eggs have acquired a greater importance as an inexpensive and high-quality protein. The Brazilian eggindustry has been characterized by a constant production expansion in the last decade, increasing the number of housedanimals and facilitating the spread of many diseases. In order to reduce the sanitary and financial risks, decisions regarding the production and the health status of the flock must be made based on objective criteria. The use of Artificial NeuralNetworks (ANN) is a valuable tool to reduce the subjectivity of the analysis. In this context, the aim of this study was atvalidating the ANNs as viable tool to be employed in the prediction and management of commercial egg production flocks.Materials, Methods & Results: Data from 42 flocks of commercial layer hens from a poultry company were selected. Thedata refer to the period between 2010 and 2018 and it represents a total of 600,000 layers. Six parameters were selectedas “output” data (number of dead birds per week, feed consumption, number of eggs, weekly weight, weekly egg production and flock uniformity) and a total of 13 parameters were selected as “input” data (flock age, flock identification, totalhens in the flock, weekly weight, flock uniformity, lineage, weekly mortality, absolute number of dead birds, eggs/hen,weekly egg production, feed consumption, flock location, creation phase). ANNs were elaborated by software programsNeuroShell Predictor and NeuroShell Classifier. The programs identified input variables for the assembly of the networksseeking the prediction of the variables called outgoing that are subsequently validated. This validation goes through thecomparison between the predictions and the real data present in the database that was the basis for the work. Validation ofeach ANN is expressed by the specific statistical parameters multiple determination (R2) and Mean Squared Error...

Artificial neural networks on eggs production data management

Background: Eggs have acquired a greater importance as an inexpensive and high-quality protein. The Brazilian eggindustry has been characterized by a constant production expansion in the last decade, increasing the number of housedanimals and facilitating the spread of many diseases. In order to reduce the sanitary and financial risks, decisions regarding the production and the health status of the flock must be made based on objective criteria. The use of Artificial NeuralNetworks (ANN) is a valuable tool to reduce the subjectivity of the analysis. In this context, the aim of this study was atvalidating the ANNs as viable tool to be employed in the prediction and management of commercial egg production flocks.Materials, Methods & Results: Data from 42 flocks of commercial layer hens from a poultry company were selected. Thedata refer to the period between 2010 and 2018 and it represents a total of 600,000 layers. Six parameters were selectedas “output” data (number of dead birds per week, feed consumption, number of eggs, weekly weight, weekly egg production and flock uniformity) and a total of 13 parameters were selected as “input” data (flock age, flock identification, totalhens in the flock, weekly weight, flock uniformity, lineage, weekly mortality, absolute number of dead birds, eggs/hen,weekly egg production, feed consumption, flock location, creation phase). ANNs were elaborated by software programsNeuroShell Predictor and NeuroShell Classifier. The programs identified input variables for the assembly of the networksseeking the prediction of the variables called outgoing that are subsequently validated. This validation goes through thecomparison between the predictions and the real data present in the database that was the basis for the work. Validation ofeach ANN is expressed by the specific statistical parameters multiple determination (R2) and Mean Squared Error...(AU)

Milk fraud by the addition of whey using an artificial neural network / Identificação de fraude de leite por adição de soro de leite usando redes neurais artificiais

The adulteration of milk by the addition of whey is a problem that concerns national and international authorities. The objective of this research was to quantify the whey content in adulterated milk samples using artificial neural networks, employing routine analyses of dairy milk samples. The analyses were performed with different concentrations of whey (0, 5, 10, and 20%), and samples were analyzed for fat, non-fat solids, density, protein, lactose, minerals, and freezing point, totaling 164 assays, of which 60% were used for network training, 20% for network validation, and 20% for neural network testing. The Garson method was used to determine the importance of the variables. The neural network technique for the determination of milk fraud by the addition of whey proved to be efficient. Among the variables of highest relevance were fat content and density.(AU)

A adulteração do leite pela adição de soro de leite é um problema que diz respeito às autoridades nacionais e internacionais. O objetivo deste trabalho foi quantificar o teor de soro em amostras de leite adulterado por meio de redes neurais artificiais, usando como variáveis de entrada os resultados de análises rotineiras em amostras de leite. As análises foram realizadas com diferentes concentrações em relação à adição de soro de leite (0, 5, 10 e 20%), e as amostras foram analisadas quanto à gordura, sólidos não gordurosos, densidade, proteína, lactose, minerais e ponto de congelamento, totalizando 164 ensaios, dos quais 60% foram utilizados para treinamento em rede, 20% para validação de rede e 20% para teste de rede neural. O método de Garson foi utilizado para determinar a importância das variáveis. A técnica de redes neurais para a determinação da fraude ao leite por adição de soro provou ser eficiente. Entre as variáveis de maior relevância estavam o teor de gordura e a densidade.(AU)

Redes neurais artificiais na estimação de variáveis biométricas de mudas de espécies florestais produzidas em diferentes substratos / Artificial neural networks for estimating the biometric variables of seedlings from forest species produced in different substrates

O objetivo do trabalho foi avaliar o crescimento em diâmetro do coleto e altura, e a produção de matéria seca total de mudas de Myracrodruon urundeuva, Jacaranda brasiliana e Mimosa caesalpiniaefolia. Concomitantemente, desenvolveu-se uma Rede Neural Artificial (RNA) do tipo Multilayer Perceptron que seria capaz de estimar a H e a MST das mudas das espécies estudadas. As mudas foram cultivadas em ambiente protegido com 50% de sombra. Assim, os tratamentos foram considerados com cinco proporções do material orgânico (0, 20, 40, 60 e 80% v/v) na composição do substrato final (solo da área desertificada). Aos 120 dias após a semeadura, as mudas foram coletadas para determinação das variáveis biométricas. A rede MLP foi utilizada empregando-se o algoritmo de treinamento Levenberg-Marquardat. As variáveis utilizadas como entrada da MLP para a estimação da altura e massa seca das mudas foram: diâmetro do coleto, diâmetro mínimo, médio e máximo do coleto, as espécies e fontes de resíduos orgânicos (esterco bovino, esterco caprino e palha de arroz), totalizando dez entradas. Foi utilizada a função de ativação tangente hiperbólica. Como resultados, recomenda-se a proporção 80:20% (esterco bovino e/ou esterco caprino:solo da área degradada) ao substrato de cultivo para o crescimento das mudas das espécies. A adição de doses de esterco bovino e esterco caprino influenciaram o DC do...

The aim of this study was to evaluate the stem growth in diameter and height as well as the production of total dry matter from seedlings of Myracrodruon urundeuva, Jacaranda brasiliana and Mimosa caesalpiniaefolia. Concurrently, an Artificial Neural Network (RNA) of Multilayer Perceptron type that would be able to estimate the H and the MST of the seedlings of the studied species was developed. The seedlings were cultivated in a protected environment with 50% shade. Thus, the treatments were considered with five proportions of the organic material (0, 20, 40, 60 and 80% v/v) in the final substrate composition (desertified area soil). At 120 days after sowing, the seedlings were collected to determine the biometric variables. The MLP network was used with help of the Levenberg-Marquardat training algorithm. The variables used as input of the MLP for height and dry mass estimation of the seedlings were: stem diameter, minimum, medium and maximum diameter of stem; and species and sources of organic residues (cattle manure, goat manure and rice straw), totaling ten entries. The hyperbolic tangent activation function was conducted. As a result, a 80:20% ratio (bovine manure and/or goat manure: soil from the degraded area) is recommended to be used in the growing substrate for seedling growth. The addition of bovine manure and goat manure doses influenced the Jacaranda brasiliana DC...

Redes neurais artificiais na estimação de variáveis biométricas de mudas de espécies florestais produzidas em diferentes substratos / Artificial neural networks for estimating the biometric variables of seedlings from forest species produced in different substrates

O objetivo do trabalho foi avaliar o crescimento em diâmetro do coleto e altura, e a produção de matéria seca total de mudas de Myracrodruon urundeuva, Jacaranda brasiliana e Mimosa caesalpiniaefolia. Concomitantemente, desenvolveu-se uma Rede Neural Artificial (RNA) do tipo Multilayer Perceptron que seria capaz de estimar a H e a MST das mudas das espécies estudadas. As mudas foram cultivadas em ambiente protegido com 50% de sombra. Assim, os tratamentos foram considerados com cinco proporções do material orgânico (0, 20, 40, 60 e 80% v/v) na composição do substrato final (solo da área desertificada). Aos 120 dias após a semeadura, as mudas foram coletadas para determinação das variáveis biométricas. A rede MLP foi utilizada empregando-se o algoritmo de treinamento Levenberg-Marquardat. As variáveis utilizadas como entrada da MLP para a estimação da altura e massa seca das mudas foram: diâmetro do coleto, diâmetro mínimo, médio e máximo do coleto, as espécies e fontes de resíduos orgânicos (esterco bovino, esterco caprino e palha de arroz), totalizando dez entradas. Foi utilizada a função de ativação tangente hiperbólica. Como resultados, recomenda-se a proporção 80:20% (esterco bovino e/ou esterco caprino:solo da área degradada) ao substrato de cultivo para o crescimento das mudas das espécies. A adição de doses de esterco bovino e esterco caprino influenciaram o DC do...(AU)

The aim of this study was to evaluate the stem growth in diameter and height as well as the production of total dry matter from seedlings of Myracrodruon urundeuva, Jacaranda brasiliana and Mimosa caesalpiniaefolia. Concurrently, an Artificial Neural Network (RNA) of Multilayer Perceptron type that would be able to estimate the H and the MST of the seedlings of the studied species was developed. The seedlings were cultivated in a protected environment with 50% shade. Thus, the treatments were considered with five proportions of the organic material (0, 20, 40, 60 and 80% v/v) in the final substrate composition (desertified area soil). At 120 days after sowing, the seedlings were collected to determine the biometric variables. The MLP network was used with help of the Levenberg-Marquardat training algorithm. The variables used as input of the MLP for height and dry mass estimation of the seedlings were: stem diameter, minimum, medium and maximum diameter of stem; and species and sources of organic residues (cattle manure, goat manure and rice straw), totaling ten entries. The hyperbolic tangent activation function was conducted. As a result, a 80:20% ratio (bovine manure and/or goat manure: soil from the degraded area) is recommended to be used in the growing substrate for seedling growth. The addition of bovine manure and goat manure doses influenced the Jacaranda brasiliana DC...(AU)

Stability of the hypocotyl length of soybean cultivars using neural networks and traditional methods / Estabilidade do comprimento do hipocótilo de plantas de cultivares de soja por meio de redes neurais e metodologias tradicionais

The length of the hypocotyl has been highlighted as a potential descriptor of the soybean crop. However, there is no information available in the published literature about its behavior over several planting times. The present study aimed to identify soybean cultivars with stability and predictability of hypocotyl length behavior through neural networks and traditional adaptability and stability methodologies. We analyzed 16 soybean cultivars in 6 planting seasons under greenhouse conditions. In each season, a randomized block design with 4 replications was adopted. The experimental unit was composed of 3 plants. The plot mean was used in the analysis. Hypocotyl length data were analyzed by analysis of variance and Tukeys test. Then analyses were carried out using the Traditional Method, Plaisted and Peterson, Wricke, Eberhart and Russell, and Artificial Neural Networks. A significant effect (p<0.01 by the F test) was identified for Cultivars versus Planting Season and Planting Seasons and Cultivars. Cultivars BRS810C, BRSMG760SRR, TMG1175RR, and BMX Tornado RR showed lower averages, high stability, and general adaptability regarding soybean hypocotyl length whereas the cultivar BG4272 presented higher mean, high stability, and general adaptability. Identification of soybean cultivars of predictable and stable behavior as to hypocotyl length contributes to Soybean Improvement as it further our knowledge on the potential descriptor and the possibility of increasing the number of descriptors.(AU)

O comprimento do hipocótilo tem-se destacado como potencial descritor da cultura da soja, no entanto, não se tem informação sobre o seu comportamento ao longo de várias épocas de plantio. Diante disto, objetivou-se identificar cultivares de soja com estabilidade e previsibilidade de comportamento quanto ao comprimento do hipocótilo por meio de redes neurais e metodologias tradicionais de adaptabilidade e estabilidade. Analisou-se 16 cultivares de soja em seis épocas de plantio, em condições de casa de vegetação. Em cada época, adotou-se o delineamento em blocos casualizados com quatro repetições, sendo a unidade experimental composta por três plantas e usou-se a média da parcela na análise. Os dados de comprimento de hipocótilo foram analisados por meio da análise de variância e teste de Tukey e, posteriormente, procedeu-se análises por meio do Método Tradicional, Plaisted e Peterson, Wricke, Eberhart e Russell e Redes Neurais Artificiais. Identificou-se efeito significativo (p<0,01 pelo teste F) para Cultivares x Épocas, Épocas e Cultivares. As cultivares BRS810C, BRSMG760SRR, TMG1175RR e BMX Tornado RR apresentaram menores médias, alta estabilidade e adaptabilidade geral quanto ao comprimento do hipocótilo de soja; enquanto que, a cultivar BG4272 apresentou maior média, alta estabilidade e adaptabilidade geral. A identificação de cultivares de soja de comportamento previsível e estável, quanto ao comprimento do hipocótilo, contribui para o Melhoramento da Soja no tocante ao melhor conhecimento do potencial descritor e à possibilidade de incremento do número de descritores.(AU)

Artificial neural network for prediction of the area under the disease progress curve of tomato late blight

Artificial neural networks (ANN) are computational models inspired by the neural systems of living beings capable of learning from examples and using them to solve problems such as non-linear prediction, and pattern recognition, in addition to several other applications. In this study, ANN were used to predict the value of the area under the disease progress curve (AUDPC) for the tomato late blight pathosystem. The AUDPC is widely used by epidemiologic studies of polycyclic diseases, especially those regarding quantitative resistance of genotypes. However, a series of six evaluations over time is necessary to obtain the final area value for this pathosystem. This study aimed to investigate the utilization of ANN to construct an AUDPC in the tomato late blight pathosystem, using a reduced number of severity evaluations. For this, four independent experiments were performed giving a total of 1836 plants infected with Phytophthora infestans pathogen. They were assessed every three days, comprised six opportunities and AUDPC calculations were performed by the conventional method. After the ANN were created it was possible to predict the AUDPC with correlations of 0.97 and 0.84 when compared to conventional methods, using 50 % and 67 % of the genotype evaluations, respectively. When using the ANN created in an experiment to predict the AUDPC of the other experiments the average correlation was 0.94, with two evaluations, 0.96, with three evaluations, between the predicted values of the ANN and they were observed in six evaluations. We present in this study a new paradigm for the use of AUDPC information in tomato experiments faced with P. infestans. This new proposed paradigm might be adapted to different pathosystems.

Automation in accession classification of Brazilian Capsicum germplasm through artificial neural networks

Germplasm classification by species requires specific knowledge on/of the culture of interest. Therefore, efforts aimed at automation of this process are necessary for the efficient management of collections. Automation of germplasm classification through artificial neural networks may be a viable and less laborious strategy. The aims of this study were to verify the classification potential of Capsicum accessions regarding/ the species based on morphological descriptors and artificial neural networks, and to establish the most important descriptors and the best network architecture for this purpose. Five hundred and sixty-four plants from 47 Brazilian Capsicum accessions were evaluated. Neural networks of multilayer perceptron type were used in order to automate the species identification through 17 morphological descriptors. Six network architectures were evaluated, and the number of neurons in the hidden layer ranged from 1 to 6. The relative importance of morphological descriptors in the classification process was established by Garson's method. Corolla color, corolla spot color, calyx annular constriction, fruit shape at pedicel attachment, and fruit color at mature stage were the most important descriptors. The network architecture with 6 neurons in the hidden layer is the most appropriate in this study. The possibility of classifying Capsicum plants regarding/ the species through artificial neural networks with 100 % accuracy was verified.

Automation in accession classification of Brazilian Capsicum germplasm through artificial neural networks

Germplasm classification by species requires specific knowledge on/of the culture of interest. Therefore, efforts aimed at automation of this process are necessary for the efficient management of collections. Automation of germplasm classification through artificial neural networks may be a viable and less laborious strategy. The aims of this study were to verify the classification potential of Capsicum accessions regarding/ the species based on morphological descriptors and artificial neural networks, and to establish the most important descriptors and the best network architecture for this purpose. Five hundred and sixty-four plants from 47 Brazilian Capsicum accessions were evaluated. Neural networks of multilayer perceptron type were used in order to automate the species identification through 17 morphological descriptors. Six network architectures were evaluated, and the number of neurons in the hidden layer ranged from 1 to 6. The relative importance of morphological descriptors in the classification process was established by Garson's method. Corolla color, corolla spot color, calyx annular constriction, fruit shape at pedicel attachment, and fruit color at mature stage were the most important descriptors. The network architecture with 6 neurons in the hidden layer is the most appropriate in this study. The possibility of classifying Capsicum plants regarding/ the species through artificial neural networks with 100 % accuracy was verified.(AU)

Artificial neural network for prediction of the area under the disease progress curve of tomato late blight

Artificial neural networks (ANN) are computational models inspired by the neural systems of living beings capable of learning from examples and using them to solve problems such as non-linear prediction, and pattern recognition, in addition to several other applications. In this study, ANN were used to predict the value of the area under the disease progress curve (AUDPC) for the tomato late blight pathosystem. The AUDPC is widely used by epidemiologic studies of polycyclic diseases, especially those regarding quantitative resistance of genotypes. However, a series of six evaluations over time is necessary to obtain the final area value for this pathosystem. This study aimed to investigate the utilization of ANN to construct an AUDPC in the tomato late blight pathosystem, using a reduced number of severity evaluations. For this, four independent experiments were performed giving a total of 1836 plants infected with Phytophthora infestans pathogen. They were assessed every three days, comprised six opportunities and AUDPC calculations were performed by the conventional method. After the ANN were created it was possible to predict the AUDPC with correlations of 0.97 and 0.84 when compared to conventional methods, using 50 % and 67 % of the genotype evaluations, respectively. When using the ANN created in an experiment to predict the AUDPC of the other experiments the average correlation was 0.94, with two evaluations, 0.96, with three evaluations, between the predicted values of the ANN and they were observed in six evaluations. We present in this study a new paradigm for the use of AUDPC information in tomato experiments faced with P. infestans. This new proposed paradigm might be adapted to different pathosystems.(AU)

PREDIÇÃO DA ÁREA DO MÚSCULO Longissimus dorsi A PARTIR DE IMAGENS ULTRASSONOGRÁFICAS EM OVINOS UTILIZANDO VISÃO COMPUTACIONAL

A ovinocultura de corte é um setor que se encontra em constante expansão. Este setor também está ganhando espaço no agronegócio brasileiro e na região Nordeste. Entretanto, a produtividade é baixa. Portanto, este setor exige alternativas tecnológicas dentro do melhoramento genético dos rebanhos, como fenotipagem de precisão, através das coletas de informações fenotípicas com precisão para permitir melhorar o rendimento de carcaça dos animais, a fim de atender às demandas e a avaliação de carcaça com resultados mais precisos, a qual tem sido realizada a partir da análise de imagens ultrassonográficas por profissionais especializados. Sendo assim, o objetivo com pesquisa foi o desenvolvimento de uma abordagem com o uso de inteligência artificial para avaliação de carcaça ovina de forma acurada por meio do reconhecimento de imagens ultrassonográficas do músculo Longissimus dorsi. A metodologia proposta para este trabalho foi dividida em 4 etapas. Na etapa de coleta de dados, 121 imagens ultrassonográficas de fêmeas ovinas foram coletadas com auxílio de aparelho de ultrassom, Durante a segmentação das regiões de interesse, utilizou-se apenas um método de segmentação automatizado com base em algoritmo de redes neurais (U-Net). Para a avaliação de segmentações automáticas, utilizou-se as métricas de coeficiente de dados Dice e a métrica de intersecção sobre união (IoU). Na extração de recursos, objetivou-se encontrar características importantes para a previsão de AOL (Área de Olho de Lombo). Na última etapa, foi realizada uma análise de regressão, sendo a variável independente os valores dos atributos obtidos com os descritores utilizados e a variável dependente a AOL previsto para a imagem de ultrassom do animal, aonde utilizou-se as métricas de quadrado médio do resíduo (QMR) e erro médio absoluto (EMA). Dois algoritmos de regressão foram utilizados, AdaBoost Regressor(ABR)e Random Forest Regressor (RFR). Realizou-se uma análise de variância e o teste t de student para comparar as médias de AOL observada e predita. Os valores obtidos pela métrica Dice foi de 0,94 e a da IoU foi de 0,89 o que demonstra alta similaridade entre o real e o previsto. Os valores de QMR, EMA e R² para o ABR é de, respectivamente, 2,61, 1,22 e 0,51 e para o RFR é de 2,15, 1,12 e 0,61, o que demonstra uma correlação positiva entre os valores preditos e os valores reais. Observou que não houve diferenças significativas entre as médias de AOL observada e predita. Portanto, a medição automatizada da AOL a partir de imagens ultrassonográficas é promissora e possibilitará maior eficiência na realização desta medida em grandes quantidades de imagens com alta precisão, pois dispensa a intervenção humana na delimitação da área do músculo Longissimus dorsi em ovinos de corte de carcaça.

Beef sheep is an expanding sector. This sector is also gaining space in Brazilian agribusiness and in the Northeast region. However, productivity is low. Therefore, this sector requires technological alternatives within the genetic improvement of herds, such as precision phenotyping, through the collection of phenotypic information needed to improve the carcass yield of the animals, to meet the demands and the carcass evaluation with more accurate results. , one that has been carried out from the analysis of ultrasound images by specialized professionals. Therefore, the objective with research was the development of an approach with the use of artificial intelligence to evaluate sheep carcass accurately through the recognition of ultrasound images of the Longissimus dorsi muscle. The methodology proposed for this work was divided into 4 stages. In the data collection stage, 121 ultrasound images of sheep from collections were collected with the aid of an ultrasound device. During the selection of the regions of interest, using only an automated supply method based on a neural network algorithm (U-Net) For the evaluation of automatic segmentations, the Dice data coefficient metrics and the intersection over union (IoU) metrics were used. In the extraction of resources, the objective was to find important characteristics for a forecast of REA (Rib Eye Area). In the last step, a regression analysis was performed, with the independent variable being the values of the resources provided with the descriptors used and the REA-dependent variable predicted for the animal's ultrasound image, where it was used as mean square error (MSE) and mean absolute error (MAE). Two regression algorithms were used, AdaBoost Regressor (ABR) and Random Forest Regressor (RFR). An analysis of variance and the student's t test were performed to compare how REA averages were compared and predicted. The values obtained by the metric Dice was 0.94 and that of the IoU was 0.89, which demonstrates a high similarity between the real and the predicted. The values of MSE, MAE and R² for ABR are 2.61, 1.22 and 0.51, respectively, and for RFR it is 2.15, 1.12 and 0.61, which demonstrates a correlation between predicted and actual values. He observed that there were no significant differences between the observed and predicted REA averages. Therefore, an automated measurement of REA from ultrasound images is promising and will enable greater efficiency in carrying out this measurement in large quantities of images with high precision, since it does not require human intervention in the delimitation of the Longissimus dorsi muscle area in carcass sheep. .

Diagnose do estado nutricional de nitrogênio, potássio e cálcio em Urochloa brizantha cv. Marandu utilizando-se técnicas de machine learning

As concentrações de nutrientes no tecido vegetal têm estreita relação com a produção das plantas forrageiras, que manifestam desordens nutricionais através de um padrão simétrico. Em Urochloa brizantha cv. Marandu diagnosticar deficiências através de análise de imagens, baseadas na manifestação de sintomas visuais, e determinar a relação do estado nutricional com as características produtivas resultará em informações práticas sobre aspectos relativos a nutrição e produção da forrageiras bem como estratégias de manejo inovadoras para uso no campo. Esta pesquisa foi desenvolvida com o objetivo de avaliar a influência dos macronutrientes nitrogênio (N), potássio (K) e cálcio (Ca) sobre a composição química, crescimento e determinação da eficiência da análise de imagens no diagnóstico do status nutricional para N, K e Ca. A Urochloa brizantha cv. Marandu foi cultivada em casa de vegetação sob cultivo hidropônico. Os tratamentos foram as seguintes concentrações de cada nutriente na solução nutritiva: 6%, 20%, 100% e 200%. As avaliações foram realizadas durante 3 ciclos de crescimento. Ao final de cada ciclo foram determinados os seguintes parâmetros da cultura: altura, índice de vegetação (NDVI), índice SPAD, massa seca da parte aérea e número e tipo de perfilhos (aéreos ou basais). Em cada ciclo, as folhas diagnósticas (folha +1 e folha +2) foram escaneadas, avaliadas através de imagens para os macronutrientes estudados e quimicamente para macro e micronutrientes bem como clorofila através de SPAD. Foram calculados os níveis críticos dos nutrientes. Ao final da pesquisa foram realizadas as avaliações de desenvolvimento de raízes e determinados os teores de nutrientes nas mesmas. Conclue-se que a disponibilidade de N, K e Ca na solução nutritiva afeta a absorção de nutrientes no capim marandu, sendo possível quantificar um padrão de remoção de nutrientes. É possível detectar o estado nutricional de N, Ca e K em capim-marandu usando técnicas de classificação de aprendizado de máquina a partir de imagens RGB, com diferença de desempenho entre as metodologias e métodos utilizados.

Nutrient concentrations in plant tissue are closely related to the production of forage plants, which manifest nutritional disorders through a symmetrical pattern. In Urochloa brizantha cv. Marandu, diagnosing deficiencies through image analysis, based on the manifestation of visual symptoms, and determining the relationship of nutritional status with productive characteristics will result in practical information on aspects related to forage nutrition and production as well as innovative management strategies for use in the field. This research was developed with the objective of evaluating the influence of the macronutrients nitrogen (N), potassium (K) and calcium (Ca) on the chemical composition, growth and determination of the efficiency of image analysis in the diagnosis of nutritional status for N, K and Ca. Urochloa brizantha cv. Marandu was cultivated in a greenhouse under hydroponic cultivation. The treatments were the following concentrations of each nutrient in the nutrient solution: 6%, 20%, 100% and 200%. Evaluations were carried out during 3 growth cycles. At the end of each cycle, the following crop parameters were determined: height, vegetation index (NDVI), SPAD index, shoot dry mass and number and type of tillers (aerial or basal). In each cycle, the diagnostic sheets (leaf +1 and leaf +2) were scanned, evaluated through images for the studied macronutrients and chemically for macro and micronutrients as well as chlorophyll through SPAD. Critical nutrient levels were calculated. At the end of the research, root development evaluations were carried out and nutrient content was determined. It is concluded that the availability of N, K and Ca in the nutrient solution affects nutrient uptake in marandu grass, making it possible to quantify a nutrient removal pattern. It is possible to detect the nutritional status of N, Ca and K in marandu grass using machine learning classification techniques from RGB images, with a difference in performance between the methodologies and methods used.

AVALIAÇÃO DO USO DE ACELERÔMETRO NA OBSERVAÇÃO DO COMPORTAMENTO INGESTIVO DE BOVINOS EM PASTEJO

O objetivo desse trabalho foi avaliar o potencial de substituição da tradicional metodologia de avaliação do comportamento ingestivo de bovinos em pastejo, a observação visual, pelo uso de um registrador eletrônico provido de acelerômetro. Para isso foram realizadas filmagens, monitoramento visual e utilização do registrador na identificação de padrões de aceleração para as atividades de pastejo e ruminação de bovinos, o que possibilitou a comparação das imagens com gráficos gerados pelo equipamento, resultando na identificação do potencial de interpretação gráfica superando a capacidade de observação visual. Na sequência, foi avaliado um algoritmo para automatização das leituras dos gráficos, qualificando e quantificando o tempo de atividades de pastejo, ruminação e ócio. Para concluir o estudo, foram realizadas comparações das metodologias (observação visual, interpretação dos gráficos e versões do algoritmo) através de uma análise de regressão linear simples, buscando o nível de correlação entre as metodologias e o potencial de substituição. Foi observado que as correlações entre as metodologias passaram de 90%, além da precisão dos dados coletados pelo registrador, o que remete a capacidade de monitorar o comportamento ingestivo de bovinos em pastejo com o uso do equipamento provido de acelerômetro. Novos testes de desenvolvimento do algoritmo podem tornar essa ferramenta ainda mais precisa e capaz de computar dados mais afinados do comportamento ingestivo como número de refeições, taxa de bocado, taxa de ruminação, entre outros.

The aim of this work was to evaluate the potential for substitution of traditional methodology for assessing the ingestive behavior of cattle grazing, visual observation, by the use of an electronic recorder provided with an accelerometer. For this purpose, filming, visual monitoring and use of recorder were carried out to identify acceleration patterns for cattle grazing and rumination activities, which made it possible to compare images with graphics generated by the equipment, resulting in the identification of potential for graphic interpretation surpassing the ability for visual observation. Then, an algorithm was evaluated to automate the readings of graphs, qualifying and quantifying time of grazing, rumination and other activities. To conclude the study, comparisons of methodologies (visual observation, graphs interpretation and algorithm versions) were carried out through simple linear regression analysis, seeking the level of correlation between methodologies and the potential for substitution. It was observed that correlations between methodologies were more than 90%, and accuracy of data collected by recorder able to monitor ingestive behavior of cattle grazing using equipment provided with an accelerometer. New tests to develop the algorithm can make this tool even more accurate and able to compute finer data on ingestive behavior such as number of meals, bit rate, rumination rate, among others.