Scientia Agropecuaria

Balancing accuracy, interpretability, and stability in machine-learning models: Live-weight prediction of Andean sheep from morphometric traits

2025-01-28T09:25:23+00:00

The objective of this research was to predict the live weight of Corriedale lambs using morphological measurements and machine learning algorithms. A total of 291 five-month-old lambs from the Corpacancha Production Unit of SAIS PACHACÚTEC SAC were used. These animals represented a homogeneous group in terms of age, sex, and genetics, as they belonged to the Corriedale breed and were offspring of "Category A" ewes. Morphological measurements recorded included Body Length (BL), Withers Height (WH), Thoracic Girth (TG), Rump Width (RW), Abdominal Girth (AG), Cannon Bone Length (CBL), Chest Depth (CD), and Live Weight (LW). The models evaluated were Multiple Linear Regression, Ridge Regression, Decision Trees, Random Forest, and XGBoost. The comparative analysis of the machine learning models identified ModG and Ridge as the most accurate and stable options, standing out for their low Mean Squared Error (MSE = 0.083) and Root Mean Squared Error (RMSE ≈ 0.287 – 0.288). Additionally, they exhibited the highest coefficients of determination (R²= 0.89, R_Adj²= 0.88), indicating excellent predictive capability and data fit. Their low coefficient of variation (CV%) confirms their stability, establishing them as the best choices for applications where precision is paramount, such as predicting critical values in production processes and high-demand scientific studies. While XGBoost proved to be a robust alternative with an MSE of 0.119, an RMSE of 0.345, and a relative error of 2.22%. These findings confirm that prioritizing models that balance accuracy, interpretability, and stability enable faster, data-driven decision-making in Corriedale sheep production. Such an approach optimizes feed allocation, classifies lambs by market weight, and promptly detects growth deviations, thereby improving overall flock profitability.

Applying artificial intelligence in durian fertile lobe detection: Attention-Residual Unet and Test Time Augmentation algorithm

2025-04-21T17:55:51+00:00

The key factor in durian fruit trading is ripeness. Several studies have been conducted on non-destructive durian maturity classification using near-infrared (NIR) spectroscopy. However, most of these studies manually determined the most accurate measurement position, which was the durian's main fertile lobe center. This research aims to automate the stage of detecting this position of the durian by using UNet segmentation method, which leverages differences in rind texture between the center of the main fertile lobe and other areas (lobe grooves and stems), prior to conducting NIR measurements. The rough and non-uniform surface of the durian rind presents a significant challenge for segmentation. However, the large size of the durian spines in the main fertile lobe serves as an identification characteristic for the segmentation model. This study uses the Ri-6 durian in Vietnam as the samples for the experiment. The model was developed using three architectures: Unet, Attention-Unet and Attention-Residual Unet. According to the analysis results on test set, Unet, Attention-Unet and Attention-Residual Unet algorithms achieved %accuracy of 78.22%, 81.34%, 82.89% and %intersection over union of 79.49%, 80.47%, 80.72%, respectively. After that, the model was further enhanced using the test time augmentation algorithm, improving the %accuracy to 85.24%, 85.68%, 86.85% and %IoU to 81.65%, 82.03% and 83.12%. Among the three architectures, the Attention-Residual-Unet model demonstrated the highest efficiency in detecting the center of the durian’s main fertile lobe for non-destructive durian maturity classification. This method can be applied to the development of an automatic durian’s maturity classification machine, which would save time and improve economic efficiency.

Morphological characterization, molecular identification, and phylogenetic analysis of Lasiodiplodia theobromae associated with CCN-51 cacao plants in Ecuador

2025-08-08T12:40:07+00:00

Necrotrophic fungi are pathogens that cause tissue death in plants, which negatively impacts their growth and productivity. This study focused on identifying the presence of Lasiodiplodia theobromae in CCN-51 cacao plants in the Simón Bolívar canton in Ecuador. We sampled cacao pods exhibiting necrotic lesions and obtained fungal isolates for morphological and molecular characterization. Techniques, such as culturing on selective media, microscopy, and DNA sequencing were used to confirm the fungal identity. We compared our results with international databases and assessed the genetic variability of the isolates. Morphological characterization placed the fungal isolates within the family Botryosphaeriaceae, and molecular analysis using ITS and EF1-α regions confirmed the species as Lasiodiplodia theobromae, with 100% DNA quality for amplicon analysis and 100% sequence similarity in GenBank. We constructed phylogenetic trees using maximum likelihood methods, which revealed high genetic similarity and recent divergence among sequences despite their varied geographic origins. The fungal isolates specifically confirmed the presence of L. theobromae as the causal agent of necrotic lesions in CCN-51 cacao pods from Simón Bolívar. These findings underscore the importance of studying necrotrophic fungi in cacao plants to inform control strategies, improve crop resistance, and support sustainable production, essential to the global cacao trade.

Antifungal activity of plant extracts against Botrytis cinerea, Lasiodiplodia theobromae, and Fusarium sp.: Effectiveness in controlling Erysiphe necator and phytotoxic effect on wheat seeds

2025-08-08T13:22:32+00:00

The agricultural sector faces risks from damage caused by phytopathogens, and many farmers rely on synthetic fungicides to combat them. However, excessive use of these products pollutes the environment and promotes pathogen resistance. This study evaluated the mycelial growth inhibition of 57 plant extracts against Botrytis cinerea, Lasiodiplodia theobromae, and Fusarium sp., and their efficacy in controlling Erysiphe necator in the field. It also evaluated their phytotoxic effect on wheat seeds and the identification of metabolites present in the extracts. The most effective extracts were those of Ambrosia artemisiifolia, Conyza sumatrensis, Dysphania ambrosioides, Minthostachis mollis, Salvia sp., Pimpinella anisum, and Syzygium aromaticum. The P. anisum extract exhibited the greatest inhibition of B. cinerea growth in tomatoes, while the most effective extracts against E. necator were P. anisum, C. sumatrensis, and S. aromaticum. Furthermore, the A. artemisiifolia extract exhibited phytotoxic effects on wheat seed growth. Flavonoids, tannins, steroids, triterpenoids, alkaloids, leucoanthocyanidins, coumarins, and saponins were identified as the main metabolites in the extracts. These results offer viable alternatives for controlling phytopathogenic fungi using plant extracts, contributing to a more sustainable agriculture that is less dependent on chemicals.

Rhizospheric and phylloplane bacteria from Capsicum annuum: Uncovering candidates for biocontrol of Ralstonia solanacearum

2025-08-18T12:11:55+00:00

Ralstonia solanacearum, the causative agent of bacterial wilt, is a major plant pathogen affecting many economically significant crops, including pepper (Capsicum annuum). This pathogen causes severe yield losses due to the limited effectiveness of current control measures. This study aimed to evaluate potential biocontrol agents for managing Ralstonia solanacearum by isolating and testing microorganisms from the rhizosphere and phylloplane of pepper plants. A total of 32 bacterial isolates were screened, and four strains showed the most pronounced antagonistic activity in vitro, producing inhibition zones ranging from 4.0 to 6.12 cm. The most effective isolates included three rhizospheric strains identified as Bacillus sp., Serratia sp., and Pseudomonas sp., and one phylloplane strain identified as Pseudomonas aeruginosa. These microorganisms effectively suppressed Ralstonia solanacearum growth under laboratory conditions and show strong potential as biocontrol agents for bacterial wilt in pepper crops.

Agronomic, physiological, and phytochemical responses of physalis to pre- and post-emergence herbicides

2025-08-18T12:43:10+00:00

Cape gooseberry (Physalis peruviana L.) is a fruit crop with increasing economic and functional relevance, yet limited research exists on weed management practices for this species. This study aimed to evaluate the agronomic, physiological, and phytochemical responses of P. peruviana to various pre- and post-emergence herbicides under greenhouse and field conditions. Two biotypes were used to assess selectivity and crop tolerance to thirteen post-emergence and two pre-emergence herbicides. Post-emergence trials revealed that chlorimuron, fomesafen, and the mixture atrazine + simazine significantly reduced plant height and caused high phytotoxicity, especially under field conditions. Conversely, quizalofop, clethodim, fluazifop, and clodinafop (ACCase inhibitors) showed excellent selectivity and maintained yield levels comparable to the control. Pre-emergence applications of S-metolachlor exhibited minimal effects on plant growth and effectively reduced weed density, while imazaquin caused a dose-dependent reduction in plant height and yield, particularly in one biotype. Phenolic compound analysis indicated that both herbicide application and weed presence influenced fruit quality. Plants grown under weed-free conditions presented the highest total phenolic content, while high weed pressure or herbicide injury reduced phenolic accumulation, especially in biotype 2. The results demonstrate that while some herbicides pose risks to P. peruviana development, others offer promising weed control options with minimal impact on crop performance and fruit quality. These findings contribute to the development of safe and effective weed management strategies for this emerging crop.

Use of purple corn (Zea mays L.) cob in the formulation of functional teas developed using Flash Profile and CATA methods

2025-08-18T13:44:23+00:00

Purple corn (Zea mays L.) is a superfood native to Peru, highly regarded for its functional properties and commonly used in the preparation of traditional beverages and desserts, such as chicha morada and mazamorra morada. Following processing, the corn cobs remaining as a byproduct retain significant amounts of bioactive compounds with potential for utilization. This study proposes their use as a primary component in the production of teas. To ensure product safety, the moisture content, total ash, and counts of enterobacteria and aflatoxins in the raw materials were first evaluated. Fourteen formulations were developed, varying in the proportions of corn cobs, quince, stevia, cinnamon, and cloves, as well as extraction times (5 and 10 minutes at 100 °C) with hot water. Two rapid sensory evaluation methods using consumer panels were applied sequentially: Flash Profile (FP) and Check-All-That-Apply (CATA), External preference mapping was then conducted, and the most acceptable teas were subjected to instrumental characterization. The FP methodology generated 400 sensory descriptors, classified semantically, from which 12 key descriptors were selected for the CATA test: sweet, stevia flavor, quince flavor, fruity flavor, fruity smell, astringent, bitter, cinnamon smell, reddish color, acid, purple and “Chicha morada” flavor. The confidence ellipses in the FP Multiple Factor Analysis (MFA) space allowed to identify six groups of formulations for the CATA test. This test revealed that the characteristics that improve consumer acceptability are: “Chicha morada” flavor, fruity flavor, sweet and fruity smell. The External Preference Mapping allowed to determine the formulations, with 90% preference among consumers, despite not being the ones with the highest concentrations of total polyphenols, antioxidant activity and monomeric anthocyanins. In conclusion, the sensory methodologies applied in this study help to elucidate the sensory characteristics that influence consumer acceptability, representing valuable tools for the development of new functional products from purple corn.

Precision technologies and their relationship with the management of agricultural inputs in the context of sustainability in vulnerable regions

2025-08-18T13:51:22+00:00

This research analyzes the relationship between the use of precision technologies and the management of agricultural inputs within the framework of sustainability. Structured surveys were conducted with 120 producers, addressing variables such as access to technology, training, technological integration, productivity, and barriers. Multiple linear regression, Spearman correlation, and cluster analysis were applied to identify the factors influencing efficient, environmentally responsible, and socioeconomically viable input management. The findings show that the Productivity and Sustainability dimension has a significant positive effect, while barriers and challenges have a negative influence. The model related to soil health (adjusted R² = 0.304) highlights the potential of precision technologies to enhance yield and reduce environmental impacts. However, structural limitations, such as high costs, insufficient infrastructure, and lack of training, were identified as major obstacles to adoption. The results emphasize the need for public policies, technical training programs, and support strategies to foster more sustainable, resilient, and inclusive agricultural systems aligned with the Sustainable Development Goals.

LiDAR imagery-based relief interpretation for soil organic carbon (SOC) estimation in the Quaternary Sumbing and Tertiary Kulon Progo Volcano Transition Zone, Central Java, Indonesia

2025-08-18T14:12:00+00:00

Typically, the transition zone of quaternary and tertiary volcanoes is a potential area for agricultural development but is prone to landslides. Landslide occurrences and the use of former landslide zones exhibit a distinct soil organic carbon (SOC) distribution, necessitating analysis to sustain agricultural output. Laboratory SOC measurements on landscape size are not expedient, necessitating the development of an estimating method for representation. This study aims to analyze the relationship between relief and SOC using LiDAR (Light Detection and Ranging) data. LiDAR acquisition was carried out to identify relief units as units of analysis. Soil sample measurements were carried out in the laboratory with the parameters analyzed including pH, Bulk Density, Moisture Content, Organic Carbon, Organic Matter, N-total, CN Ratio and Cation Exchange Capacity. The results showed that SOC and relief had R² = 0.89 in the upper layer (0 – 20 cm) and 0.86 in the lower layer (20 - 40 cm). Relief has a high correlation with soil characteristics at the top and bottom of soil depths. It is because of relatively stable elevation and relatively dynamic land cover that SOC is spread out in a clustered way. This research can be used as a basis for agricultural land management, especially in areas prone to landslides.

Genetic diversity of antagonistic bacterial isolates obtained from Theobroma cacao L. to control Fusarium oxysporum f. sp. cubense race 1

2025-04-02T00:50:25+00:00

The banana (Musa AAA) is affected by Fusarium oxysporum f. sp. cubense, which causes discoloration in the xylem duct, leading to terminal wilting. The use of plant growth promoting rhizobacteria (PGPR) as a biological control produces different antagonistic compounds and inhibits the growth of various phytopathogens. The objective of the study was based on the molecular identification of rhizobacteria that produce phytohormones with biocontrol activity against Foc-R1. The presence of the 225 bp ChiA gene was observed in PGPR. Phylogenetic analysis of 16S rRNA by sequencing and ERIC-PCR showed genetic variability with the formation of four subgroups. Molecular identification by sequencing the 16S rRNA gene defined the genera as Klebsiella, Enterobacter, and Pseudomonas. There is variation in the biosynthesis of the phytohormones AIA, AG, and AS in strains MH-18, W-417, and FZ 9-7 at 72 h. The identification of Foc-R1 by PCR shows an amplicon of 350 bp. Antagonistic assays of bacterial supernatants from strain FZ 9-7 show 71% mycelial inhibition of Foc-R1 and a decrease in spore production of 2.5X10⁶ spores mL^-1. The results provide information on the genetic relationships of PGPRs through the production of secondary metabolites such as proteases, catalases, chitinases, and siderophores, as well as morphological and molecular analysis for the identification of Foc-R1 and its interaction with antagonistic extracts in inhibiting the growth of diseases in bananas and cocoa.

Controlled-release nanofertilizer using chitosan nanoparticles loaded with NPK: Development and impact on the yield and nutritional quality of Solanum tuberosum

2025-03-26T07:53:40+00:00

In recent years, nanotechnology has made significant progress in various fields, including agriculture, where nanofertilizers play a pioneering role in improving crop productivity and reducing environmental pollution. In this study, a controlled-release nanofertilizer was developed using chitosan nanoparticles loaded with nitrogen (N), phosphorus (P), and potassium (K). The chitosan nanoparticles (CS-NPs) were prepared via an ionic gelation method using sodium tripolyphosphate and characterized by scanning transmission electron microscopy and infrared spectroscopy. The results revealed that the nanoparticle size ranged from 17.21 nm to 18.32 nm. The controlled release of N, P, and K was evaluated over 240 hours. The nanofertilizer was then applied foliage to Solanum tuberosum seedlings under greenhouse conditions. The findings indicated that the 0.25% chitosan nanofertilizer formulation resulted in nanoparticles with relatively high nutrient absorption capacity, with average values of 4.65 mg/L nitrogen, 198.55 mg/L phosphorus, and 1345.27 mg/L potassium. However, the most effective nanofertilizer treatment was the 1% chitosan nanoparticle formulation loaded with 5 ppm N, P, or K, resulting in the best nutritional characteristics among all the fertilization treatments and a 37% increase in the mass yield of Solanum tuberosum compared with that of the control. These results suggest that NPK-loaded chitosan nanoparticles could be used as foliar sprays to produce more nutritious and higher-yielding crops.

UV-C priming enhances antioxidant mechanisms and bioactive compound biosynthesis in broccoli sprouts

2025-09-01T14:47:42+00:00

Effects of six compost products on soil organic carbon and nitrogen levels in alfisol, rice yields, and diazotrophic endophytic populations

2025-09-01T15:01:24+00:00

The decrease in nutrient content in the soil occurs because of an imbalance between nutrient intake and loss. This study aimed to evaluate the effect of applying six types of compost products, with different compost materials and bioactivators, on the soil organic carbon (SOC) and total nitrogen (TN) levels in Alfisol, on rice yields, and diazotrophic endophytic populations. The pot experiment was arranged in a Completely Randomized Design with a single factor consisting of 10 levels and 3 replications. The treatments compared control (R0), 6 types of compost (R1-R6), 50% chemical fertilizer (CF) (R7), 100% CF (R8), and a set combination (R9). Rice with a variety of Mentik Wangi was grown and harvested on 118 DAP. The results showed that the treatment R5 (compost C4 of a mixture of leaf litter, cow dung, peanut plant residue, and rock phosphate with bioactivator RMC) gave the highest increases of SOC and TN by 200% and 228.57%, and the highest plant dry weight and total seed weight by 247.55% and 171.16% compared to the control, respectively. Three treatments of R4 (compost C3), R6 (compost C5), and R9 (50% dosage of compost C3 in combination with mycorrhiza, zeolite, Azolla, and rock phosphate) yielded a similar effect at the second-highest levels in increasing SOC, TN, and rice yields. There was no significant effect of treatments on the population of diazotrophic bacterial endophytes in rice leaves. The present study revealed that compost enrichment with an effective bioactivator contributed significantly higher effects on soil fertility, plant growth, and yield compared to 100% chemical fertilizer (R8).

Nematophagous fungi for integrated management of Meloidogyne (Tylenchida): a review of taxonomic diversity, mechanisms of action and potential as biological control agents

2025-08-08T14:18:01+00:00

Root-knot nematodes (RKNs) are classified under the genus Meloidogyne and are among the most devastating pests affecting strategical agricultural crops. They attack a wide variety of plant species, including vegetables, fruit trees and ornamental plants, causing root deformities and even lead to plant death in severe cases of infestation. These nematodes contribute to substantial crop yield loss and affect the quality of harvested products. Although synthetic nematicides are available for the control of these pest organisms, there is a growing emphasis on exploring sustainable and eco-friendly alternatives, such as nematophagous fungi like the genera Purpureocillium, Arthrobotrys, Dactylellina, Orbilia, and Trichoderma, among others. Here a review of literature on the matter is given, with a focus on the taxonomic classification of the most relevant fungal orders and genera, their diagnostic features, mechanisms of action, and potential as biological control agents (BCAs) against Meloidogyne species. Other relevant aspects addressed in this review include a brief description of the nematode genus Meloidogyne, along with the symptoms it causes in host plants, such as root gall formation, stunted growth, and yellowing of foliage, among others. It also describes integrated pest management (IPM) strategies such as crop rotation, resistant crops, soil solarization, trap crops, as well as currently used chemical control techniques. Biological control alternatives are also presented with particular emphasis on nematophagous fungi. Future research should focus on improving the formulae of biological agents based on nematophagous fungi under field conditions and understanding their ecological roles and interactions in soil microbiomes.

Impact of land use change and climate on the Brazilian Amazon: a review on carbon stocks and greenhouse gas emissions

2025-09-01T15:25:24+00:00

Historically, land-use changes in the Brazilian Amazon, such as the conversion of forests to pastures, have significantly impacted carbon and nitrogen cycles, contributing to greenhouse gas emissions and potentially compromising environmental sustainability. This review explores the effects of these changes on soil carbon and nitrogen stocks, emphasizing the role of sustainable practices and public policies in mitigating environmental impacts. Findings indicate that, although forest-to-pasture conversion may reduce soil organic carbon stocks by up to 11.3%, practices such as agroforestry systems, sustainable pasture management, and crop-livestock-forestry integration (CLFI) have the potential to reverse these effects by promoting carbon sequestration and soil conservation. Public policies such as the Low Carbon Agriculture Plan (ABC Plan) and the Amazon Fund are highlighted as essential pillars for sustainable development in the region. It is concluded that the adoption of sustainable agricultural practices, integrated with robust environmental policies and technological innovation, can transform the Amazon into a global model of balance between economic development and environmental conservation. Future studies should prioritize integrated assessments of carbon stocks, gas emissions, and socioeconomic indicators to support more effective and regionally adapted public policies.