Deep Learning Regression Approaches Applied to Estimate Tillering in Tropical Forages Using Mobile Phone Images.

We assessed the performance of Convolutional Neural Network (CNN)-based approaches using mobile phone images to estimate regrowth density in tropical forages. We generated a dataset composed of 1124 labeled images with 2 mobile phones 7 days after the harvest of the forage plants. Six architectures were evaluated, including AlexNet, ResNet (18, 34, and 50 layers), ResNeXt101, and DarkNet. The best regression model showed a mean absolute error of 7.70 and a correlation of 0.89. Our findings suggest that our proposal using deep learning on mobile phone images can successfully be used to estimate regrowth density in forages.

Convolutional Neural Networks to Estimate Dry Matter Yield in a Guineagrass Breeding Program Using UAV Remote Sensing.

Forage dry matter is the main source of nutrients in the diet of ruminant animals. Thus, this trait is evaluated in most forage breeding programs with the objective of increasing the yield. Novel solutions combining unmanned aerial vehicles (UAVs) and computer vision are crucial to increase the efficiency of forage breeding programs, to support high-throughput phenotyping (HTP), aiming to estimate parameters correlated to important traits. The main goal of this study was to propose a convolutional neural network (CNN) approach using UAV-RGB imagery to estimate dry matter yield traits in a guineagrass breeding program. For this, an experiment composed of 330 plots of full-sib families and checks conducted at Embrapa Beef Cattle, Brazil, was used. The image dataset was composed of images obtained with an RGB sensor embedded in a Phantom 4 PRO. The traits leaf dry matter yield (LDMY) and total dry matter yield (TDMY) were obtained by conventional agronomic methodology and considered as the ground-truth data. Different CNN architectures were analyzed, such as AlexNet, ResNeXt50, DarkNet53, and two networks proposed recently for related tasks named MaCNN and LF-CNN. Pretrained AlexNet and ResNeXt50 architectures were also studied. Ten-fold cross-validation was used for training and testing the model. Estimates of DMY traits by each CNN architecture were considered as new HTP traits to compare with real traits. Pearson correlation coefficient r between real and HTP traits ranged from 0.62 to 0.79 for LDMY and from 0.60 to 0.76 for TDMY; root square mean error (RSME) ranged from 286.24 to 366.93 kg·ha-1 for LDMY and from 413.07 to 506.56 kg·ha-1 for TDMY. All the CNNs generated heritable HTP traits, except LF-CNN for LDMY and AlexNet for TDMY. Genetic correlations between real and HTP traits were high but varied according to the CNN architecture. HTP trait from ResNeXt50 pretrained achieved the best results for indirect selection regardless of the dry matter trait. This demonstrates that CNNs with remote sensing data are highly promising for HTP for dry matter yield traits in forage breeding programs.

Deep Learning Applied to Phenotyping of Biomass in Forages with UAV-Based RGB Imagery.

Monitoring biomass of forages in experimental plots and livestock farms is a time-consuming, expensive, and biased task. Thus, non-destructive, accurate, precise, and quick phenotyping strategies for biomass yield are needed. To promote high-throughput phenotyping in forages, we propose and evaluate the use of deep learning-based methods and UAV (Unmanned Aerial Vehicle)-based RGB images to estimate the value of biomass yield by different genotypes of the forage grass species Panicum maximum Jacq. Experiments were conducted in the Brazilian Cerrado with 110 genotypes with three replications, totaling 330 plots. Two regression models based on Convolutional Neural Networks (CNNs) named AlexNet and ResNet18 were evaluated, and compared to VGGNet-adopted in previous work in the same thematic for other grass species. The predictions returned by the models reached a correlation of 0.88 and a mean absolute error of 12.98% using AlexNet considering pre-training and data augmentation. This proposal may contribute to forage biomass estimation in breeding populations and livestock areas, as well as to reduce the labor in the field.

A joint learning approach for genomic prediction in polyploid grasses.

Poaceae, among the most abundant plant families, includes many economically important polyploid species, such as forage grasses and sugarcane (Saccharum spp.). These species have elevated genomic complexities and limited genetic resources, hindering the application of marker-assisted selection strategies. Currently, the most promising approach for increasing genetic gains in plant breeding is genomic selection. However, due to the polyploidy nature of these polyploid species, more accurate models for incorporating genomic selection into breeding schemes are needed. This study aims to develop a machine learning method by using a joint learning approach to predict complex traits from genotypic data. Biparental populations of sugarcane and two species of forage grasses (Urochloa decumbens, Megathyrsus maximus) were genotyped, and several quantitative traits were measured. High-quality markers were used to predict several traits in different cross-validation scenarios. By combining classification and regression strategies, we developed a predictive system with promising results. Compared with traditional genomic prediction methods, the proposed strategy achieved accuracy improvements exceeding 50%. Our results suggest that the developed methodology could be implemented in breeding programs, helping reduce breeding cycles and increase genetic gains.

A Semi-Automated SNP-Based Approach for Contaminant Identification in Biparental Polyploid Populations of Tropical Forage Grasses.

Artificial hybridization plays a fundamental role in plant breeding programs since it generates new genotypic combinations that can result in desirable phenotypes. Depending on the species and mode of reproduction, controlled crosses may be challenging, and contaminating individuals can be introduced accidentally. In this context, the identification of such contaminants is important to avoid compromising further selection cycles, as well as genetic and genomic studies. The main objective of this work was to propose an automated multivariate methodology for the detection and classification of putative contaminants, including apomictic clones (ACs), self-fertilized individuals, half-siblings (HSs), and full contaminants (FCs), in biparental polyploid progenies of tropical forage grasses. We established a pipeline to identify contaminants in genotyping-by-sequencing (GBS) data encoded as allele dosages of single nucleotide polymorphism (SNP) markers by integrating principal component analysis (PCA), genotypic analysis (GA) measures based on Mendelian segregation, and clustering analysis (CA). The combination of these methods allowed for the correct identification of all contaminants in all simulated progenies and the detection of putative contaminants in three real progenies of tropical forage grasses, providing an easy and promising methodology for the identification of contaminants in biparental progenies of tetraploid and hexaploid species. The proposed pipeline was made available through the polyCID Shiny app and can be easily coupled with traditional genetic approaches, such as linkage map construction, thereby increasing the efficiency of breeding programs.

High-Resolution Linkage Map With Allele Dosage Allows the Identification of Regions Governing Complex Traits and Apospory in Guinea Grass (Megathyrsus maximus).

Forage grasses are mainly used in animal feed to fatten cattle and dairy herds, and guinea grass (Megathyrsus maximus) is considered one of the most productive of the tropical forage crops that reproduce by seeds. Due to the recent process of domestication, this species has several genomic complexities, such as autotetraploidy and aposporous apomixis. Consequently, approaches that relate phenotypic and genotypic data are incipient. In this context, we built a linkage map with allele dosage and generated novel information of the genetic architecture of traits that are important for the breeding of M. maximus. From a full-sib progeny, a linkage map containing 858 single nucleotide polymorphism (SNP) markers with allele dosage information expected for an autotetraploid was obtained. The high genetic variability of the progeny allowed us to map 10 quantitative trait loci (QTLs) related to agronomic traits, such as regrowth capacity and total dry matter, and 36 QTLs related to nutritional quality, which were distributed among all homology groups (HGs). Various overlapping regions associated with the quantitative traits suggested QTL hotspots. In addition, we were able to map one locus that controls apospory (apo-locus) in HG II. A total of 55 different gene families involved in cellular metabolism and plant growth were identified from markers adjacent to the QTLs and APOSPORY locus using the Panicum virgatum genome as a reference in comparisons with the genomes of Arabidopsis thaliana and Oryza sativa. Our results provide a better understanding of the genetic basis of reproduction by apomixis and traits important for breeding programs that considerably influence animal productivity as well as the quality of meat and milk.

Genomic Selection with Allele Dosage in Panicum maximum Jacq.

Genomic selection is an efficient approach to get shorter breeding cycles in recurrent selection programs and greater genetic gains with selection of superior individuals. Despite advances in genotyping techniques, genetic studies for polyploid species have been limited to a rough approximation of studies in diploid species. The major challenge is to distinguish the different types of heterozygotes present in polyploid populations. In this work, we evaluated different genomic prediction models applied to a recurrent selection population of 530 genotypes of Panicum maximum, an autotetraploid forage grass. We also investigated the effect of the allele dosage in the prediction, i.e., considering tetraploid (GS-TD) or diploid (GS-DD) allele dosage. A longitudinal linear mixed model was fitted for each one of the six phenotypic traits, considering different covariance matrices for genetic and residual effects. A total of 41,424 genotyping-by-sequencing markers were obtained using 96-plex and Pst1 restriction enzyme, and quantitative genotype calling was performed. Six predictive models were generalized to tetraploid species and predictive ability was estimated by a replicated fivefold cross-validation process. GS-TD and GS-DD models were performed considering 1,223 informative markers. Overall, GS-TD data yielded higher predictive abilities than with GS-DD data. However, different predictive models had similar predictive ability performance. In this work, we provide bioinformatic and modeling guidelines to consider tetraploid dosage and observed that genomic selection may lead to additional gains in recurrent selection program of P. maximum.

De novo transcriptome assembly for the tropical grass Panicum maximum Jacq.

Guinea grass (Panicum maximum Jacq.) is a tropical African grass often used to feed beef cattle, which is an important economic activity in Brazil. Brazil is the leader in global meat exportation because of its exclusively pasture-raised bovine herds. Guinea grass also has potential uses in bioenergy production due to its elevated biomass generation through the C4 photosynthesis pathway. We generated approximately 13 Gb of data from Illumina sequencing of P. maximum leaves. Four different genotypes were sequenced, and the combined reads were assembled de novo into 38,192 unigenes and annotated; approximately 63% of the unigenes had homology to other proteins in the NCBI non-redundant protein database. Functional classification through COG (Clusters of Orthologous Groups), GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) analyses showed that the unigenes from Guinea grass leaves are involved in a wide range of biological processes and metabolic pathways, including C4 photosynthesis and lignocellulose generation, which are important for cattle grazing and bioenergy production. The most abundant transcripts were involved in carbon fixation, photosynthesis, RNA translation and heavy metal cellular homeostasis. Finally, we identified a number of potential molecular markers, including 5,035 microsatellites (SSRs) and 346,456 single nucleotide polymorphisms (SNPs). To the best of our knowledge, this is the first study to characterize the complete leaf transcriptome of P. maximum using high-throughput sequencing. The biological information provided here will aid in gene expression studies and marker-assisted selection-based breeding research in tropical grasses.

Genotypic value in hybrid progenies of Panicum maximum Jacq / Valor genotípico de progênies híbridas de Panicum maximum Jacq

ABSTRACT: Genetic breeding of forage plants has increasingly contributed to the release of more productive plants. In this regard, evaluating the genotypic value is essential when aiming to rank genotypes based on the mean free of environmental factors. Therefore, this study aimed to predict the genotypic value of agronomic and nutritive value characters of three progenies of Panicum maximum. Hybrids were evaluated in a clonal test in an incomplete-randomized design with three treatments (progenies 1, 2, and 3) and two replications (clones). Six harvests were performed at 25cm from the ground level throughout one year. Progeny 2 provided better results for total and leaf dry mass yield, regrowth, and height, and lower incidence of leaf spot. Progenies 1 and 3 had a better response for qualitative characters such as higher crude protein and digestibility and lower lignin and fiber content. Hybrid progenies of P. maximum have forage characters of interest for breeding, and when using 'Mombaça' grass as parents, the progeny stands out for leaf production and resistance to leaf spot and for 'Tanzania' grass as parent has resulted in better forage quality.

RESUMO: O lançamento de forrageiras resultantes de programas de melhoramento genético tem sido importante fonte de liberação de novas forrageiras mais adaptadas e competitivas. Nessas situações, a avaliação do valor genético é essencial quando se objetiva ranquear os genótipos com base no valor genotípico, isento dos efeitos ambientais. O objetivo com este trabalho foi estimar e avaliar o valor genotípico de características agronômicas e de valor nutritivo de três progênies de P. maximum, resultantes do cruzamento entre duas progenitoras sexuais e as cultivares 'Mombaça' e 'Tanzânia'. O experimento foi implantado em teste clonal no delineamento em blocos incompletos com três tratamentos (progênies 1, 2 e 3) com duas repetições (clones). Os híbridos foram manejados por meio de cortes na altura de 25cm do nível do solo por um ano, realizando seis cortes. A progênie 2 proporcionou melhores resultados para produção de folhas, rebrota, altura de planta e baixa incidência de mancha foliar causada por Bipolaris maydis. As progênies 1 e 3 apresentaram, em média, melhores resultados para características qualitativas como proteína bruta e digestibilidade e menor teor de lignina. As progênies híbridas de P. maximum apresentam características forrageiras de interesse para o melhoramento, sendo que a utilização do capim-mombaça como parental proporciona maior produção de folhas e resistência à mancha foliar, ao passo que o capim-tanzânia como parental proporciona melhoria da qualidade da forragem.

Diversidade genética molecular entre cultivares e híbridos de Brachiaria spp. e Panicum maximum / Molecular genetic diversity among cultivars and hybrids of Brachiaria spp. and Panicum maximum

A utilização de marcadores moleculares pode servir para direcionar cruzamentos, confirmar novos híbridos e identificar genótipos para fins comerciais. Nesse contexto, o objetivo deste trabalho foi analisar a diversidade genética entre cultivares e híbridos de Brachiaria spp. e Panicum maximum usando marcadores moleculares do tipo RAPD (Polimorfismos de DNA amplificados ao acaso). Foram 22 genótipos analisados com 10 primers, os quais amplificaram 178 fragmentos polimórficos de DNA, que foram usados para estimar a similaridade genética por meio do coeficiente de Jaccard. Os valores de similaridade obtidos variaram de 0,066 a 0,841. A estrutura genética entre todos os genótipos estudados foi estimada pelo método UPGMA (Método de agrupamento com médias aritméticas não ponderadas), revelando três grupos distintos com altos valores de bootstrap (>89 por cento). Os resultados demonstraram que a técnica RAPD oferece uma maneira rápida, relativamente barata e útil para a caracterização da diversidade genética entre as diferentes cultivares e híbridos de Brachiaria ssp. e P. maximum analisados.

The use of molecular markers may serve to direct crossings, confirm new hybrids and identify new genotypes for commercial purposes. In that context, this research aimed to analyze the genetic diversity among cultivars and hybrids of Brachiaria spp. and P. maximum using molecular markers of the type RAPD (Random amplified polymorphic DNA). It was analyzed 22 genotypes with 10 primers, which amplified 178 DNA polymorphic fragments, which were used to estimate the similarity using Jaccard coefficient. The values of similarity ranged from 0.066 to 0.841. The genetic structure among genotypes was estimated by UPGMA (Unweighted pair-group method with arithmetical average) and revealed three distinct groups with high bootstrap values (>89 percent). The results showed that the RAPD is a fast, relatively inexpensive and useful technique for genetic divergence characterization between different cultivars and hybrids of Brachiaria spp. and P. maximum.

Soil flood tolerance of seven genotypes of Panicum maximum Jacq

The soil flood tolerance of seven genotypes of Panicum maximum Jacq. (PM11, PM34, PM40 and PM45, and the commercial cultivars Massai, Mombaça and Tanzânia) was evaluated in plants subjected to two conditions: flooded and not flooded, during a period of 14 days. Flooding significantly decreased the total and above ground biomass of PM40 and PM45. For cultivar Tanzânia, flooding decreased these two variables and also root biomass. The root, total and above ground relative growth rates were significantly reduced by flooding in cultivar Tanzânia, while in PM45 only the above ground relative growth rate was reduced. Cultivar Tanzânia showed significant differences for all variables analyzed, thus was not flood tolerant, as well as PM40 and PM45. It could be concluded that Massai, PM34, Mombaça and PM11 were the most flood tolerant genotypes.

Avaliou-se a tolerância ao alagamento do solo em sete genótipos de Panicum maximum Jacq. As plantas foram submetidas a duas condições: alagado e não alagado, avaliadas por um período de 14 dias. O alagamento reduziu a produção de biomassa seca da parte aérea e total (para PM40, PM45, p < 5 por cento). Para a cv. Tanzânia (p < 1 por cento), além destas variáveis, reduziu a biomassa da raiz. Quanto à taxa de crescimento relativo total sob alagamento em relação à testemunha foi significativa apenas para PM45 (p < 5 por cento), na parte aérea e para a cv. Tanzânia (p < 1 por cento) na taxa de crescimento relativo da raiz, parte aérea e total; a cv. Tanzânia apresentou diferenças significativas em todas as variáveis analisadas, não sendo tolerante ao alagamento, assim como PM40 e PM45; Massai, Mombaça, PM11 e PM34 são tolerantes ao alagamento, sendo que o mais tolerante foi a cv. Massai.

Asynchronous meiosis in an interspecific hybrid of Brachiaria ruziziensis and B. brizantha.

Male meiosis is generally synchronous in higher plants. The regulation of the cell cycle is still not well understood, and a powerful tool for gaining an understanding of this regulation is the development of mutations that affect cell-cycle synchrony. We report here asynchronous microsporogenesis in an interspecific hybrid between two important tropical grasses. In young spikelets of the interspecific hybrid 49.10% of anther meiocytes entered meiosis, exhibiting typical phases of the first and second divisions, while the other 50.90% showed distinctive features of early prophase. In older spikelets, anthers containing mature pollen grains also displayed meiocytes still undergoing meiosis. At this time, the latter cells were enclosed by the exine wall. Despite asynchrony, all cells completed meiosis. Old anthers contained only pollen grains that appeared to be in the same stage of development. Pollen fertility was estimated to be 52.76% in dehiscent anthers. An independent genetic control for meiosis synchrony and meiotic stages is suggested.

Estimativas de repetibilidade para caracteres forrageiros em Panicum maximum / Repeatability estimates for forage characters in Panicum maximum

Objetiva-se com este trabalho estimar a repetibilidade para caracteres forrageiros de Panicum, e determinar o número de cortes de avaliação necessários para a seleção de genótipos de Panicum, com confiabilidade. Utilizaram-se os dados de um ensaio conduzido no período de 21/11/2002 a 08/04/2005, no Campo Experimental da Embrapa Gado de Leite, localizado em Valença-RJ, onde foram realizados 15 cortes de avaliação. No ensaio, foram avaliados 23 genótipos de Panicum maximum, em parcelas experimentais, dispostas no delineamento de blocos casualizados, com três repetições. Foram estimados os coeficientes de repetibilidade para as características produção de matéria verde de forragem (PMV); produção de matéria seca de forragem (PMS) e de folhas (PMSF); porcentagem de folhas na PMS ( por centoFOL) e altura da planta (AP), utilizando os métodos da análise de variância, componentes principais e análise estrutural. Para todas as características avaliadas os efeitos de genótipos, cortes e interação genótipos x cortes foram significativos (P<0,01). Quando se considerou o coeficiente de determinação de 85 por cento, o número de avaliações (cortes) necessários para a determinação do valor real dos genótipos foram de 10, 9, 7, 11 e 3, respectivamente, para PMV, PMS, PMSF, por centoFOL e AP. O método dos componentes principais e o da análiseestrutural (baseado na matriz de correlações) foram concordantes para todas as características avaliadas. A realização de 10 cortes de avaliação permite discriminar o valor real dos genótipos de Panicum, com confiabilidade superior a 85 por cento, para a maioria das características avaliadas.

The objective of this work was to estimate the repeatability for forage characters of Panicum and to determinate the necessary number of evaluation cuts to select Panicum genotypes with confidence. Data of a trial with 15 cuts, carried out between 21/11/2002 and 08/04/2005 in the experimental station of Embrapa Gado de Leite located in Valença, RJ, Brazil, were used. In this study. 23 genotypes of " Panicum maximum" were evaluated, in a complete randomized block, with three replications. The coefficient of repeatability for fresh forage production (PMV), total plant dry matter production (PMS) and leaves dry matter production (PMSF) were recorded along with leaves percentage in PMS ( percentFOL) and plant hight (AP), using the variance analysis, main components and structural analysis methods. For all evaluated parameters the effects of genotype, cut and genotype x cut interaction were significant (P<0.01). When considering the determination coefficient as 85 percent, the required number of measures (cuts) to determine the real value of genotypes were 10, 9, 7, 11 and 3, for PMV, PMS, PMSF, percentFOL and AP, respectively. Main components and structural analysis methods (based on the correlation matrix) did agree for all features evaluated. The utilization of data from 10 cuts allows to discriminate, with confidence higher than 85 percent, the real value for most of the evaluated characters of Panicum genotypes.