Browsing by Author "Acur, Amos"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Genetic diversity of aflatoxin-producing Aspergillus flavus isolated from groundnuts in selected agroecological zones of Uganda
    (BMC microbiology, 2020) Acur, Amos; Arias, Renée S.; Odongo, Steven; Tuhaise, Samuel; Ssekandi, Joseph; Muhanguzi, Dennis; Adriko, John; Buah, Stephen; Kiggundu, Andrew
    Background Aspergillus is the main fungal genus causing pre- and post-harvest contamination of groundnuts. Aspergillus flavus belongs to section Flavi, a group consisting of both the aflatoxigenic species (A. flavus, A. parasiticus and A. nomius) and non-aflatoxigenic species (A. oryzae, A. sojae and A. tamarii). Aflatoxins are food-borne toxic secondary metabolites produced by Aspergillus species, causing hepatic carcinoma and stunting in children and are the most toxic carcinogenic mycotoxins ever identified. Despite the well-known public health problems associated with aflatoxicosis in Uganda, information about the genetic diversity of the main aflatoxin causing fungus, Aspergillus flavus in this country is still limited. Results A cross-sectional survey was therefore carried out in three main groundnut-growing agro-ecological zones (AEZs) of Uganda; West Nile farming system, Lake Kyoga basin mixed farming system and Lake Victoria basin farming system. This was to assess the genetic diversity of A. flavus and to establish the contamination rates of groundnuts with Aspergillus species at pre- and post-harvest stages. Out of the 213 A. flavus isolates identified in this study, 96 representative isolates were fingerprinted using 16 insertion/deletion microsatellite markers. Data from fingerprinting were analyzed through Neighbor Joining while polymorphism was determined using Arlequin v 3.5. The pre- and post-harvest contamination rates were; 2.5% and 50.0% (West Nile farming system), 55.0% and 35.0% (Lake Kyoga basin mixed farming system) and 32.5% and 32.5% (Lake Victoria basin farming system) respectively. The Chi-square test showed no significant differences between pre- and post-harvest contamination rates among AEZs (p = 0.199). Only 67 out of 96 isolates produced suitable allele scores for genotypic analysis. Analysis of genetic diversity showed higher variation within populations than among populations. Two major clusters (aflatoxigenic and non-aflatoxigenic isolates) were identified as colonizing groundnuts at pre- and postharvest stages. Conclusions These findings provide a first insight on the existence of non-aflatoxigenic strains of A. flavus in Uganda. These strains are potential candidates for developing local Aspergillus biocontrol agent.
  • Thumbnail Image
    Item
    Leveraging edge computing and deep learning for the real-time identification of bean plant pathologies
    (Elsevier B.V, 2024-12) Katumba, Andrew;; Okello, Wayne Steven;; Murindanyi, Sudi ;; Nakatumba-Nabende, Joyce;; Bomera, Moses;; Mugalu, Ben Wycliff;; Acur, Amos
    Beans are essential crops globally, standing out as one of the most consumed and nourishing legumes, thereby playing a significant role in human nutrition and food security. Their cultivation faces several challenges, such as pests, diseases, unpredictable weather patterns, and soil erosion. Of these challenges, diseases are recognized as a key challenge, resulting in a decline in both yield quality and quantity, and inflicting substantial financial losses on farmers. This work proposes a deep learning-based approach for precise in-field identification of diseases in bean plants. We evaluate image classification and object detection models using state-of-the-art Convolutional Neural Network (CNN) architectures to identify Angular Leaf Spot (ALS) and bean rust diseases, key bean diseases in Uganda and the region in general, from smartphone images of bean leaves collected in various districts of Uganda. The dataset employed to train these models is the Makerere University beans image dataset, comprising 15,335 images categorized into three (3) classes (ALS, bean rust, and healthy). To improve in-field performance, the dataset was expanded to include an additional class (unknown class) consisting of a diverse collection of 2,800 images to account for images unrelated to the three (3) predefined classes. Adversarial training was further employed to enhance model robustness in identifying the target classes. In addition, two (2) Out-of-Distribution (ODD) detection techniques, i.e., confidence thresholding and training with an auxiliary class (unknown class), were utilized to handle inputs unrelated to bean leaves. Our results show that our custom CNN, BeanWatchNet, achieved 90% accuracy when tested on unseen data for the classification of the three (3) target classes, i.e., ALS, bean rust and healthy. EfficientNet v2 B0 and BeanWatchNet demonstrated superior performance for the four-class (with unknown class) image classification task, achieving 91% and 90% accuracy, respectively, when evaluated on the test dataset. YOLO v8 exhibited superior performance for the object detection models, attaining mAP@50 of 87.6. The custom CNN model and YOLO v8 model were quantized and deployed across two (2) edge platforms: a smartphone (through a mobile application) and a Raspberry Pi 4B to facilitate in-field disease detection. The benchmarking code and models are publicly available on GitHub.1 •Deep learning approach for real-time identification of bean rust and Angular Leaf Spot (ALS) diseases in bean leaves.•Image classification models trained with adversarial examples and ODD detection mechanisms to enhance model robustness.•Object detection models trained using CNN architectures to localize disease-affected areas.•Quantization and edge deployment of deep learning models on a smartphone and a Raspberry Pi 4B to ease in-field disease detection.
  • Thumbnail Image
    Item
    Promoting insect farming and household consumption through agricultural training and nutrition education in Africa: A study protocol for a multisite cluster-randomized controlled trial
    (Public Library of Science, 2023-07-19) Alemu, Mohammed Hussen; Halloran, Afton; Olsen, Søren Bøye; Anankware, Jacob Paarechuga; Nyeko, Philip; Ayieko, Monica; Nyakeri, Evans; Kinyuru, John; Konyole, Silvenus; Niassy, Saliou; Egonyu, James Peter; Malinga, Geoffrey Maxwell; Ng'ang'a, Jeremiah; Ng'ong'a, Charles Adino; Okeyo, Nicky; Debrah, Shadrack Kwaku; Kiiru, Samuel; Acur, Amos; Roos, Nanna
    Edible insects are a sustainable source of high-quality animal protein. Insect farming is gaining interest globally, particularly in low-income countries, where it may provide substantial nutritional and economic benefits. To enhance insect farming practices in Africa, new farming systems are being developed. However, knowledge on how to best promote uptake of these systems is lacking. This study aims to fill this gap by investigating the effectiveness of educational interventions in promoting insect farming for household consumption in Africa. The study is designed as a multi-site randomized controlled trial to evaluate the impacts of agricultural training alone or in combination with nutrition education on the adoption of insect farming in Ghana, Kenya and Uganda. In each of the three countries, ninety-nine villages are randomly assigned to one of three arms: two intervention arms and a control arm with no interventions. Focusing on production (P), the first intervention arm covers agricultural training on insect farming combined with provision of insect production starter kits. Focusing on both production and consumption (PC), the second intervention arm involves the same intervention components as treatment P plus additional nutrition education. The impacts of the interventions are measured by comparing baseline and endline data collected one year apart. Primary outcomes are adoption of insect farming and consumption of the farmed insects. Understanding the drivers and impacts of novel agricultural practices is crucial for transitioning to sustainable food systems. The current project is the first to investigate how educational interventions promote insect farming for household consumption in low-income countries. The results will contribute evidence-based knowledge to support sustainable development through insect farming in Africa.