The Nelson Mandela African Institution of Science and Technology
Institutional Repository
Research Collection
Select to browse Research collections.
Recent Submissions
Global sensitivity analysis of catheter-associated urinary tract infection models using the eFAST method
(2026-02-26) Nyambo, Devotha; Lyimo, Tumaini; Sauli, Elingarami
Malaria and Dengue are the major vector borne diseases (VBDs) with high prevalence in the tropics, where their existence is significantly linked to the changing climatic conditions which favour the continuous breeding of vectors. In Tanzania, the burden for malaria is more pronounced in rural and periโurban settings than in urban, with prevalence of up to 24% for under 5 children, varying across geographical zones. This paper presents a high-resolution UAV dataset collected from Mtwara and Tabora regions in southern and western zones, respectively, for development of surveillance models for potential mosquito breeding sites. The presented dataset contains 19 classes of mosquito breeding sites that were identified from 471 aerial images as a 80.5% of retained images with useful information from the total of 585 collected in Igunga and Masasi Districts in Tabora and Mtwara regions, respectively. With an experimental approach, the paper presents the use of the imagery dataset on the state of art Gemma 3 N E4B, an instruction-tuned multimodal architecture capable of processing both visual and textual data. By using the 80.5% proportion remained after collaborative image labelling, we trained and validated the instructional model with exceptional convergence characteristics and a stabilization phase yielding a loss of 0.0319. The reduction in loss was 96.7%, from 9.620 to 0.319, within 117 steps.
Graph Convolution Neural Network and Deep Q-Network Optimization-Based Intrusion Detection with Explainability Analysis
(MDPI, 2026-02-24) Mwiga, Kelvin; Dida, Mussa; Maglaras, Leandros; Mohsin, Ahmad; Janicke, Helge; Sarker, Iqbal
As networks expand in size and complexity, coupled with an exponential increase in
intrusions on network and IoT systems, this leads to traditional models failing to cap
ture increasingly intricate correlations among network components accurately. Graph
Convolution Networks (GCNs) have recently acquired prominence for their capacity to
represent nodes, edges, or entire graphs by aggregating information from adjacent nodes.
However, the correlations between nodes and their neighbours, as well as related edges,
differ. Assigning higher weights to nodes and edges with high similarity improves model
accuracy and expressiveness. In this paper, we propose the GCN-DQN model, which
integrates GCN with a multi-head attention mechanism and DQN (Deep Q Network)
to adaptively adjust attention weights optimizing its performance in intrusion detection
tasks. After extensive experiments using the UNSW NB15 and CIC-IDS2017 dataset, the
proposed GCN-DQNoutperformed the baseline model in classification accuracy. We also
applied LIME and SHAPtechniques to provide explainability to our proposed intrusion
detection model.
Global sensitivity analysis of catheter-associated urinary tract infection models using the eFAST method
(2026-03-04) John, Innocent; Stephano, Mussa; Mayengo, Maranya
Catheter-associated urinary tract infection (CAUTI) is one of the most common healthcare
associated infections, posing a significant challenge in clinical settings. This study develops
a mathematical model that incorporates bacterial contamination to investigate the transmission
dynamics of CAUTI. We derive the disease-free and endemic equilibria and compute the basic
reproduction number, ๎พ0, using the next generation matrix method. The modelโs well-posedness
is examined through the existence and uniqueness of solutions, and the long-term behavior
is analyzed to determine the stability of the equilibria. To assess the relative importance of
the model parameters, we conduct a global sensitivity analysis using the extended Fourier
Amplitude Sensitivity Test (eFAST) method. The results identify the catheterization rate (๐),
catheter removal rate (๐ฟ), and transmission coefficients (๐, ๐ฝ1, ๐ฝ2) as the most influential
parameters affecting infection dynamics. These findings highlight key intervention targets for
controlling CAUTI. The model also serves as a foundation for future extensions, including the
incorporation of asymptomatic carriers and environmental sanitation interventions.
Genomic diversity of human adenoviruses in Tanzanian children under five: Insights into F40, F41, B, and rare A18 genotypes
(Elsevier B.V., 2026-01-23) Shayo, Mariana; Kuchaka, Davis; Beti, Melkiory; Kimu, Patrick; Wadugu, Boaz; Jensen, Emilie; Kumburu, Happiness; Kazyoba, Paul; Ali, Mohamed; Consortium, SeqTZ; Clausen, Philip; Muro, Florida; Mmbaga, Blandina; Alifrangis, Michael; Aarestrup, Frank; Sonda, Tolbert
Human adenoviruses (HAdVs) are important pathogens that are associated with a wide array of clinical diseases,
particularly in the pediatric population. Despite numerous reports of HAdV infections in Tanzania, there are
currently no whole genome sequences from this region available in global public databases. This gap presents
challenges to our efforts to understand their dissemination and evolution over time. This study employed
nanopore-based metagenomic sequencing to detect and sequence the whole genomes of HAdV strains in Tan
zanian infants with diarrhea. We present the first whole genome of HAdV-A18 from Africa, representing only the
third worldwide. Additionally, it includes the first complete genomes of HAdV-F40, HAdV-F41, and HAdV-B3
obtained from Tanzania. In addition, this study provides information on the enteric adenovirus lineages circu
lating in Tanzania. These findings provide crucial genomic insights into the diversity of viruses in sub-Saharan
Africa and underscore the importance of genomic surveillance to deepen our understanding of adenovirus
transmission and evolution.
Genomic characterization of a novel human astrovirus MLB1 variant from Tanzania
(Elsevier B.V., 2026-03-03) Shayo, Mariana; Kuchaka, Davis; Beti, Melkiory; Mwing'a, Gerald; Kimu, Patrick; Wadugu, Boaz; Juma, Masoud; Kumburu, Happiness; Kiwelu, Ireen; Ali, Mohamed; Kazyoba, Paul; Clausen, Philip; Muro, Florida; Mmbaga, Blandina; Alifrangis, Michael; Aarestrup, Frank; Sonda, Tolbert
Human astroviruses (HAstV) are emerging viral pathogens responsible for gastroenteritis worldwide, but whole-genome data from Africa, including from Tanzania, remains scarce. A genomic analysis of stools of children with diarrhea was performed using nanopore sequencing.
We successfully assembled HAstV-MLB1 complete genomes from two different stool samples, achieving high sequence coverage of 354.75x and 769.94x, respectively. Analyses included phylogeny, synteny, comparative genomics, and structural characterization of the capsid protein. The two Tanzanian HAstV-MLB1 strains showed high genetic similarity to each other and clustered within the Melbourne-like virus (MLB1) clade and were most closely related (96% similarity) to the Australian reference strain. Both synteny and capsid structure analysis confirmed their conserved nature.
This study presents the first whole-genome data from Tanzania. Our findings align with the previously reported genetic stability of HAstV, regional variability, and potential functional variation of the capsid protein, emphasizing the need for surveillance and functional characterization.