Biology subjects

Sweeney, E.

Publications and source records attributed to Sweeney, E..

11 recordsLinked to original sources

bioRxiv2026

Drought-Spec-Net: Early Tomato Drought Detection and Potential Yield-Impact Assessment Using Vis NIR Data

Drought stress significantly reduces tomato (Solanum lycopersicum L.) productivity, and early detection is critical to minimize yield losses through timely interventions. In this study, we developed Drought-Spec-Net, a hybrid 1D convolutional neural network that integrates local and global spectral feature extraction to detect early drought stress from visible and near infrared (Vis NIR) spectra data of tomato seedlings. The model was trained on 378 samples using an 80:20 train test split, with 20% of the training set reserved for validation. DroughtSpecNet outperformed the evaluated baseline and state of the art models, achieving 97% accuracy, 95% precision, 98% recall, and an F1 score of 97%. To improve the agronomic interpretation of the model outputs, predicted drought probabilities were converted into a literature-informed potential yield impact indicator using a maximum impact level of 60%. On the test set (76 samples), mapped potential yield-impact values ranged from 0% to 60%, with an average reduction of 12.97%. We also conducted an initial experiment using our greenhouse RGB dataset, collected daily from drought treated and well-watered tomato plants at West Virginia State University (WVSU). From this dataset, 44 images were selected for ilastik-based canopy segmentation, producing plant-level drought severity indices (DSI) with a mean of 0.28, median of 0.14, and range of 0.01 to 0.91. Additionally, we trained and fine-tuned a large language model (LLM) based on PLLaMA7BInstruct, called AgriLLaMA, for automated agronomic report generation from Drought-Spec-Net outputs. The generated reports summarize predicted stress levels, mapped potential yield impacts, and preliminary management considerations. This integrated approach not only improves early drought stress detection but also delivers quantitative and interpretable estimates of potential productivity losses, providing a complete framework connecting physiological stress detection to actionable agricultural outcomes.

plant biology↗

bioRxiv2026

A panel of genotypically and phenotypically characterised WHO priority Gram-negative bacteria to facilitate antimicrobial and diagnostic development

Background Antimicrobial-resistant Gram-negative bacteria pose a major threat to global health. In the WHO Bacterial Priority Pathogen List (BPPL) 2024, several species were classified as critical or high priority for research and development of new antimicrobial therapeutics and diagnostics. However, access to well-characterised and clinically relevant bacterial isolate panels remains an unmet need for development and validation. Methods We constructed five panels of Gram-negative bacterial isolates, cultured from samples originally referred to the UK Health Security Agency Antimicrobial Resistance and Healthcare-Associated Infections Reference Unit for analysis between 2014 and 2025. For each isolate, antimicrobial susceptibilities were determined against a range of clinically relevant antibiotics, including third-generation cephalosporins, carbapenems, colistin and other last-line antibiotics. WGS was performed on all isolates, and previously described resistance determinants were characterised. Results The five panels consist of 145 isolates and represent 81 sequence types across Escherichia coli (30 isolates), Klebsiella pneumoniae (30 isolates), carbapenem-resistant Acinetobacter baumannii (30 isolates), carbapenem-resistant Pseudomonas aeruginosa (30 isolates), and a mixed panel of other healthcare-associated Enterobacterales species (25 isolates). Panels are highly representative of globally relevant strains, including those currently in circulation and on the WHO BPPL 2024, and of resistance mechanisms of public health importance. Conclusion The panels provide a diverse, comprehensive set of multidrug-resistant Gram-negative bacterial isolates, representative of strains currently in circulation globally. Selected isolates carry a breadth of important resistance mechanisms and are highly relevant to the current epidemiological landscape. The isolates and associated metadata are available to industry, academia, and other laboratories for use in developing novel antimicrobial compounds and diagnostic assays.

microbiology↗

bioRxiv2026

LeafyVGG-16: Transfer Learning for Plant Disease Detection with Cyber Risk Analysis

Plant disease detection using deep learning is essential for precision agriculture, enabling early and automated crop health monitoring. This study proposes an end-to-end transfer learning pipeline, LeafyVGG-16, for multi-class classification of plant diseases and nutrient deficiencies using a tomato leaf dataset. The framework integrates data preprocessing, augmentation, and a VGG-16 backbone with a two-stage fine-tuning strategy. The proposed model is evaluated against CNN, DenseNet-121, Inception-V3, EfficientNetB0, and ResNet-50, achieving an accuracy of 0.93 with precision, recall, and F1-scores of 0.93, 0.90, and 0.92, respectively. These results demonstrate the effectiveness of transfer learning for fine-grained plant disease recognition. We further evaluate model robustness under adversarial cyber attacks to assess deployment reliability in agricultural systems. Under Fast Gradient Sign Method (FGSM) attacks ({epsilon} = 0.01- 0.05), the model shows an accuracy drop of 1%-7.5%, while Projected Gradient Descent (PGD) attacks ({epsilon} = 0.05, step size = 0.005, 10 iterations) produce similar degradation, highlighting the models vulnerability to adversarial perturbations. These findings highlight potential security and reliability risks in AI-based agricultural decision-making systems. Future work will focus on improving robustness and cyber-resilience and extending this framework to other crops for secure and context-aware deployment in resource-constrained environments.

plant biology↗

bioRxiv2024

Immune checkpoint inhibitor treatment does not impair ovarian or endocrine function in a mouse model of triple negative breast cancer

BackgroundRepresenting 15-20% of all breast cancer cases, triple negative breast cancer (TNBC) is diagnosed more frequently in reproductive-age women and exhibits higher rates of disease metastasis and recurrence when compared with other subtypes. Few targeted treatments exist for TNBC, and many patients experience infertility and endocrine disruption as a result of frontline chemotherapy treatment. While they are a promising option for less toxic therapeutic approaches, little is known about the effects of immune checkpoint inhibitors on reproductive and endocrine function. ResultsOur findings in a syngeneic TNBC mouse model revealed that therapeutically relevant immunotherapies targeting PD-1, LAG-3, and TIM-3 had no effect on the quality and abundance of ovarian follicles, estrus cyclicity, or hormonal homeostasis. Similarly, in a tumor-free mouse model, we found that ovarian architecture, follicle abundance, estrus cyclicity, and ovulatory efficiency remain unchanged by PD-1 blockade. ConclusionsTaken together, our results suggest that immunotherapy may be a promising component of fertility-sparing therapeutic regimens for patients that wish to retain ovarian and endocrine function after cancer treatment.

cancer biology↗

bioRxiv2021

Functional connectome reorganization relates to post-stroke motor recovery and structural disconnection

Motor recovery following ischemic stroke is contingent on the ability of surviving brain networks to compensate for damaged tissue. In rodent models, sensory and motor cortical representations have been shown to remap onto intact tissue around the lesion site, but remapping to more distal sites (e.g. in the contralesional hemisphere) has also been observed. Resting state functional connectivity (FC) analysis has been employed to study compensatory network adaptations in humans, but mechanisms and time course of motor recovery are not well understood. Here, we examine longitudinal FC in 23 first-episode ischemic pontine stroke patients (34-74 years old; 8 female, 15 male) and utilize a graph matching approach to identify patterns of regional functional connectivity reorganization during recovery. We quantified functional reorganization between several intervals ranging from 1 week to 6 months following stroke, and demonstrated that the areas that undergo functional reorganization most frequently are in cerebellar/subcortical networks. Brain regions with more structural connectome disruption due to the stroke also had more functional remapping over time. Finally, we show that the amount of functional reorganization between time points is correlated with the extent of motor recovery observed between those time points in the early to late subacute phases, and, furthermore, individuals with greater baseline motor impairment demonstrate more extensive early subacute functional reorganization (from one to two weeks post-stroke) and this reorganization correlates with better motor recovery at 6 months. Taken together, these results suggest that our graph matching approach can quantify recovery-relevant, whole-brain functional connectivity network reorganization after stroke.

neuroscience↗

bioRxiv2021

Increasing age is independently associated with higher free water in non-active MS brain - A multi-compartment analysis using FAST-T2

PurposeTo explore the relationship between the cerebral cortical perivascular space (PVS) and aging in non-active MS subjects by using the multi-echo T2 relaxometry based cerebrospinal fluid fraction (CSFF) map. MethodsMulti-echo spiral T2 data from 111 subjects with non-active multiple sclerosis (MS) were retrospectively investigated by fitting the T2 data into a three-compartment model, the three water compartments including myelin water, intra-extracellular water, and cerebrospinal fluid. Segmentation of T1w image was performed to get the region of interest (ROI) in cerebral cortical regions. The white matter lesion segmentation was conducted using a convolutional neural network (CNN) based segmentation tool. The CSFF in the ROIs were correlated with age by controlling the gender, white matter hyperintensity lesion burden, and MS disease duration. Multiple linear models were created for the analysis of aging effect on the CSFF. ResultsThe ROI analysis shows that the CSFF in the cerebral cortical regions (temporal, occipital, parietal, front, hippo, and mtl) are significantly linear increasing with age (p<0.01). The intra-extracellular water fraction (IEWF) in the ROIs are significantly linear decreasing (p<0.01). ConclusionThe multi-echo T2 based three-compartment model can be used to quantify the CSFF. The linear increase of CSF water contents in the cerebral cortical regions indicates increased perivascular space load in cortex with aging. The quantification of CSFF may provide a way to understand the glymphatic clearance function in aging and neurodegenerations. HighlightsO_LIMR T2 relaxometry is a valid method to quantify the cerebrospinal fluid fraction (CSFF) in cerebral cortical regions C_LIO_LIThe CSFF in the cerebral cortical regions are positively correlated with age by controlling the white matter lesion load in non-active MS subjects. C_LIO_LIQuantification of cerebral CSFF may reflect the perivascular space load in cortex and better interpret the disease progression in neurodegenerative disease, such as MS. C_LI

neuroscience↗

bioRxiv2020

Effect of host-mimicking medium and biofilm growth on the ability of colistin to kill Pseudomonas aeruginosa.

In vivo biofilms cause recalcitrant infections with extensive and unpredictable antibiotic tolerance. Here, we demonstrate increased tolerance of colistin by Pseudomonas aeruginosa when grown in cystic fibrosis-mimicking medium versus standard medium in in vitro biofilm assays, and drastically increased tolerance when grown in an ex vivo CF model versus the in vitro assay. We used colistin conjugated to the fluorescent dye BODIPY to assess the penetration of the antibiotic into ex vivo biofilms and showed that poor penetration partly explains the high doses of drug necessary to kill bacteria in these biofilms. The ability of antibiotics to penetrate the biofilm matrix is key to their clinical success, but hard to measure. Our results demonstrate both the importance of reduced entry into the matrix in in vivo-like biofilm, and the tractability of using a fluorescent tag and benchtop fluorimeter to assess antibiotic entry into biofilms. This method could be a relatively quick, cheap and useful addition to diagnostic and R&D pipelines, allowing the assessment of drug entry into biofilms, in in vivo-like conditions, prior to more detailed tests of biofilm killing.

microbiology↗

bioRxiv2020

Predicting antibiotic-associated virulence of Pseudomonas aeruginosa using an ex-vivo lung biofilm model

BackgroundBacterial biofilms are known to have high antibiotic tolerance which directly affects clearance of bacterial infections in people with cystic fibrosis (CF). Current antibiotic susceptibility testing methods are either based on planktonic cells or do not reflect the complexity of biofilms in vivo. Consequently, inaccurate diagnostics affect treatment choice, preventing bacterial clearance and potentially selecting for antibiotic resistance. This leads to prolonged, ineffective treatment. MethodsIn this study, we use an ex-vivo lung biofilm model to study antibiotic tolerance and virulence of Pseudomonas aeruginosa. Sections of pig bronchiole were dissected, prepared and infected with clinical isolates of P. aeruginosa and incubated in artificial sputum media to form biofilms, as previously described. Then, lung-associated biofilms were challenged with antibiotics, at therapeutically relevant concentrations, before their bacterial load and virulence were quantified and detected, respectively. ResultsThe results demonstrated minimal effect on the bacterial load with therapeutically relevant concentrations of ciprofloxacin and meropenem, with the later causing an increased production of proteases and pyocyanin. A combination of meropenem and tobramycin did not show any additional decrease in bacterial load but demonstrated a slight decrease in total proteases and pyocyanin production. ConclusionsWe demonstrate a realistic model for understanding antibiotic resistance and tolerance in biofilms clinically and for molecules screening in anti-biofilm drug development. P. aeruginosa showed high levels of antibiotic tolerance, with minimal effect on bacterial load and increased proteases production, which could negatively affect lung function. This may potentially contribute to exacerbations and eventual lung failure.

microbiology↗

bioRxiv2019

Building a better biofilm - formation of in vivo-like biofilm structures by Pseudomonas aeruginosa in a porcine model of cystic fibrosis lung infection

Pseudomonas aeruginosa biofilm infections in the cystic fibrosis (CF) lung are highly resistant to current antimicrobial treatments and are associated with increased mortality rates. The existing models for such infections are not able to reliably mimic the clinical biofilms observed. We aimed to further optimise an ex vivo pig lung (EVPL) model for P. aeruginosa CF lung infection that can be used to increase understanding of chronic CF biofilm infection. The EVPL model will facilitate discovery of novel infection prevention methods and treatments, and enhanced exploration of biofilm architecture. We investigated purine metabolism and biofilm formation in the model using transposon insertion mutants in P. aeruginosa PA14 for key genes: purD, gacA and pelA. Our results demonstrate that EVPL recapitulates a key aspect of in vivo P. aeruginosa infection metabolism, and that the pathogen forms a biofilm with a clinically realistic structure not seen in other in vitro studies. Two pathways known to be required for in vivo biofilm infection - the Gac regulatory pathway and production of the Pel exopolysaccharide - are essential to the formation of this mature, structured biofilm on EVPL tissue. We propose the high-throughput EVPL model as a validated biofilm platform to bridge the gap between in vitro and CF lung infection.

microbiology↗

bioRxiv2019

An ex vivo cystic fibrosis model recapitulates key clinical aspects of chronic Staphylococcus aureus infection

Staphylococcus aureus is one of the most prevalent organisms isolated from the airways of people with cystic fibrosis (CF), predominantly early in life. Yet its role in the pathology of lung disease is poorly understood. Clinical studies are limited in scope by age and health of participants and in vitro studies are not always able to accurately recapitulate chronic disease characteristics such as the development of small colony variants. Further, animal models also do not fully represent features of clinical disease: in particular, mice are not readily colonized by S. aureus and when infection is established it leads to the formation of abscesses, a phenomenon almost never observed in the human CF lung. Here, we present details of the development of an existing ex vivo pig lung model of CF infection to investigate the growth of S. aureus. We show that S. aureus is able to establish infection and demonstrates clinically significant characteristics including small colony variant phenotype, increased antibiotic tolerance and preferential localisation in mucus. Tissue invasion and the formation of abscesses were not observed, in line with clinical data.

microbiology↗

bioRxiv2018

Objective Evaluation of Multiple Sclerosis Lesion Segmentation using a Data Management and Processing Infrastructure

We present a study of multiple sclerosis segmentation algorithms conducted at the international MICCAI 2016 challenge. This challenge was operated using a new open-science computing infrastructure. This allowed for the automatic and independent evaluation of a large range of algorithms in a fair and completely automatic manner. This computing infrastructure was used to evaluate thirteen methods of MS lesions segmentation, exploring a broad range of state-of-the-art algorithms, against a high-quality database of 53 MS cases coming from four centers following a common definition of the acquisition protocol. Each case was annotated manually by an unprecedented number of seven different experts. Results of the challenge highlighted that automatic algorithms, including the recent machine learning methods (random forests, deep learning, ...), are still trailing human expertise on both detection and delineation criteria. In addition, we demonstrate that computing a statistically robust consensus of the algorithms performs closer to human expertise on one score (segmentation) although still trailing on detection scores.

neuroscience↗