Toxicity of nano-scaled Al2O3, SiO2, TiO2 and ZnO to bacteria (Bacillus subtilis, Escherichia coli and Pseudomonas fluorescens) was examined and compared to that of their respective bulk (micro-scaled) counterparts. All nanoparticles (NPs) but TiO2 showed higher toxicity than their bulk counterparts. Toxicity of released metal ions was differentiated from that of the oxide particles. ZnO was the most toxic among the three NPs, causing 100% mortality to the three tested bacteria. TEM images showed attachment of NPs to the bacteria, suggesting that the toxicity was affected by bacterial attachment. The effects of oxide NPs on bacteria cells and bacterial surface biomolecules were studied by FTIR spectroscopy to provide a better understanding of their cytotoxicity. Lipopolysaccharide (LPS) and lipoteichoic acid could bind to oxide NPs through hydrogen bonding and ligand exchange, but the cytotoxicity of NPs seemed largely related to the function-involved or structural changes to proteins and phospholipids. The three NPs decreased the intensity ratio of β-sheets/α-helices, indicating protein structure change, which may affect cell physiological activities. The phosphodiester bond of L-α- Phosphatidyl-ethanolamine (PE) was broken by ZnO NPs, forming phosphate monoesters and resulting in the highly disordered alkyl chain. Such damage to phospholipid molecular structure may lead to membrane rupture and cell leaking, which is consistent with the fact that ZnO is the most toxic of the three NPs. LPS and PE are amphiphilic biomolecules that are major constituents of the outer membrane of Gram-negative bacteria. Their micelles and vesicles were studied as model cell membranes to evaluate NP effects on membrane construction. The adsorption of polysaccharides on Al2O3 and TiO2 NPs dispersed LPS vesicles and micelles. LPS coated Al2O3 NPs, while it caused the aggregation of TiO2 NPs according to atom force microscopy images. Desorption from the two NPs was slow due to the firm hydrogen bonding. For PE, Al2O3 NPs induced large multilamillar vesicles, while ZnO NP converted vesicles to tiny aggregates due to molecular structure breakup. PE stability in solution was disturbed by adding NPs, but its stability was enhanced by increasing pH. The electrostatic force was the determining factor for the vesicle stability.

Trace quantities of pharmaceuticals (including carbamazepine and sertraline) are continuously discharged into the environment, which causes concern among scientists and regulators regarding their potential long-term impacts on aquatic ecosystems. These compounds and their metabolites are continuously interacting with the orgranisms in various life stages, and may differentially influence development of embryo, larvae, juvenile, and adult stages. To fully understand the potential ecological risks of two candidate pharmaceutical chemicals (carbamazepine (CBZ) and sertraline (SERT)) exposure on survival, growth and reproduction of Ceriodaphnia dubia in three sucessive generations under static renewal toxicity test, a multigenerational approach was taken. Results indicate that SERT exposure showed higher sensitivity to chronic exposure to C. dubia growth and reproduction than CBZ exposure. The lowest concentration to affect fecundity and growth was at 50 µg L-1 SERT in the first two generations. These parameters become more sensitive during the third generation where the LOEC was 4.8 µg L-1. The effective concentrations (EC50) for the number of offspring per female, offspring body size, and dry weight were 17.2, 21.2, and 26.2 µg SERT L-1, respectively. Endpoints measured in this study demonstrate that chronic exposure of C. dubia to SERT leads to effects that occur at concentrations an order of magnitude higher than predicted environmental concentrations indicating potential transgenerationals effects. Additionally, a process-based dynamic energy budget (DEB) model is implemented to predict the simulated effects of chronic toxicity of SERT and CBZ to C. dubia individual behavior at laboratory condition. The model‘s output indicates the ecotoxicological mode of action of SERT exposure, which acts on feeding or assimilation with an effect that rapidly saturates at higher concentrations. Offspring size decreases with the toxic effects on feeding, and offspring number is thus less affected than total investment in reproduction. Consequently, CBZ affects direclty in reproduction which are captured by DEBtox model as increased embryonic hazard and reproduction cost as well as growth and maintenance costs. Furthermore, stress factor linearly increased not only with increasing chemical concentrations but also with exposure time. The DEBtox model establishes a cumulative life history consequence of multigenerational exposure to CBZ and SERT. This approach provides a tool to which to understand the effect of chemical to the individual organism and predict the population level effects in ecological risk assessment of the emerging contaminants.

The rapid growth of user-generated content on social media platforms in the form of text caused moderating toxic language manually to become an increasingly challenging task. Consequently, researchers have turned to artificial intelligence (AI) and machine learning (ML) models to detect and classify toxic comments automatically. However, these models often exhibit unintended bias against comments containing sensitive terms related to de- mographic groups, such as race and gender, which leads to unfair classifications of samples. In addition, most existing research on this topic focuses on fully supervised learning frame- works. Therefore, there is a growing need to explore fairness in semi-supervised toxicity detection due to the difficulty of annotating large amounts of data. In this thesis, we aim to address this gap by developing a fair generative-based semi-supervised framework for mitigating social bias in toxicity text classification. This framework consists of two parts, first, we trained a semi-supervised generative-based text classification model on the bench- mark toxicity datasets. Then, in the second step, we mitigated social bias in the trained classifier in step 1 using adversarial debiasing, to improve fairness. In this work, we use two different semi-supervised generative-based text classification models, NDAGAN and GANBERT (the difference between them is that the former adds negative data augmenta- tion to address some of the problems in GANBERT), to propose two fair semi-supervised models called FairNDAGAN and FairGANBERT. Finally, we compare the performance of the proposed fair semi-supervised models in terms of accuracy and fairness (equalized odds difference) against baselines to clarify the challenges of social fairness in semi-supervised toxicity text classification for the first time. Based on the experimental results, the key contributions of this research are: first, we propose a novel fair semi-supervised generative-based framework for fair toxicity text classification for the first time. Second, we show that we can achieve fairness in semi- supervised toxicity text classification without considerable loss of accuracy. Third, we demonstrate that achieving fairness at the coarse-grained level improves fairness at the fine-grained level but does not always guarantee it. Fourth, we justify the impact of the labeled and unlabeled data in terms of fairness and accuracy in the studied semi- supervised framework. Finally, we demonstrate the susceptibility of the supervised and semi-supervised models against data imbalance in terms of accuracy and fairness.

Introduction: Acute chemotherapy toxicity is common and can have negative effects for the patient and health economy and hospitalisation can be necessitated. Aims: To identify the incidence of toxicity and admission, and predictors of toxicity occurrence, severity, hospitalisation and length of stay. Method: Data was obtained from a proactive telephone assessment of acute toxicity 24 hours after administration of a first cycle of chemotherapy to patients in a large UK NHS teaching hospital. Results: 1539 patients were studied and the overall incidence of toxicity was 35.6% (530 patients). Disease site and number of chemotherapy agents given were shown to predict toxicity, with breast and upper gastrointestinal cancers having a higher likelihood of toxicity. Disease was predictive of toxicity grade, with urology, gynaecology and lung cancer patients experiencing higher grades of toxicity than other tumour sites. The rate of hospital admission due to toxicity was 13.1% (203 patients) and median length of stay 3 days (1-28). The risk of admission had some risk factors in common with toxicity. Disease and the number of drugs in the regimen affected the risk of admission, with gynaecology, head and neck and lung cancer patients and patients who received 3 drugs having a higher likelihood of admission. Predictors in the subgroups of breast, colorectal and lung cancer patients did not differ greatly from the whole population and the number of drugs was shown to be a predictor of nausea, vomiting and fatigue when explored as secondary outcomes. Conclusion: The research partly addressed the main aim and highlighted areas where further research is required. Keywords

The main objectives of this study were to investigate the applicability of chitosan as an effective wood preservative against subterranean termites, conduct metagenomic analysis of the bacterial hindgut community of Reticulitermes flavipes exposed to chitosan-treated wood, and perform chitosanase activity assay of metagenomics suggested bacterial species potentially responsible for chitosan breakdown. Chitosan showed termiticidal effects on subterranean termites at varying retention levels. Termite mortality increased when exposed to samples treated with higher chitosan concentration solutions. Approximately 40 - 100% of chitosan retained in treated-wood was leached depending on the initial retention. Post-leaching results indicate chitosan is not suitable for protection against both subterranean termites in outdoor conditions, but should be effective in non-leaching/indoor applications. For metagenomic analysis of the bacterial hindgut community of Reticulitermes flavipes, two methods were used for sequence data interpretation. The Illumina BaseSpace program identified twenty-six bacteria phyla with significant differences in abundance between the chitosan-treated and control groups. The second method, mothur, identified fifteen bacterial phyla also with significant differences in abundance between both treatment groups. Similar bacterial taxa were uniquely assigned to samples from termites fed on chitosan-treated wood using both methods. These results suggest a treatment driven effect on the hindgut bacteria diversity. While majority of the bacterial taxa were common to both methods, inconsistencies detected using the BaseSpace program suggests that the Greengenes database in its present state is not reliable for 16S rRNA gene sequence analysis. As for chitosanase activity of bacterial species with significance abundance from chitosan-treated wood exposed termites, three bacteria species, Lactococcus raffinolactis, Lactococcus lactis, and Dysgonomonas gadei, were examined. After culturing on chitosan media plates and broth, no conclusive activity could be detected from all three species. Further studies need to be conducted to understand the mechanism of chitosan toxicity to termites and insects in general and to prevent chitosan leaching from treated wood. A comparative metatransciptomic study needs to be implemented to supplement the metagenomic study performed herein, so as to elucidate the exact bacteria species involved in chitosan breakdown and the enzymes produced. Also, other bacterial species suggested by the metagenomic data to possess chitosanase activity should be investigated.

No description available.

No description available.

No description available.

<p> Quantum dots are semiconductor nanocrystals with unique optical properties that give them the potential to be excellent probes for bio-imaging applications. However, before quantum dots can be employed for such applications, their toxicity and cellular interactions need to be thoroughly assessed. The Caenorhabditis elegans (C. elegans) embryo was chosen as a test environment to study both the toxicity and dynamics of carboxyl terminated CdSe/ZnS quantum dots. Using confocal imaging, it was found that the C. elegans embryo is not morphologically affected by the introduction of quantum dots up to a concentration of about 1 OOnM. However, the embryo was observed to respond to the nanomaterial by packaging it into aggregates during development in a dose and time dependant manner. Image analysis and fluorescence correlation spectroscopy revealed that this packaging process happens from the nm scale to the J.Ull scale and that it reduces quantum dot mobility over development. This work shows that the dynamics of the quantum dots are highly influenced by the cellular environment in the embryo, as they appear to aggregate and possibly also interact with cellular structures and organelles in the embryo. </p> / Thesis / Master of Applied Science (MASc)

The toxicity of the metals: Cd, Cu, Pb, Ni, and Zn to the freshwater aquatic larvae of the midge fly Chironomus riparius was investigated using 24 h waterborne exposures. Even at the most sensitive life stage, first instar, the chironomids were extremely metal tolerant, with LC50 values for all metals being orders of magnitudes above both the CCME Canadian water quality guidelines for the protection of aquatic life, and the USEP A Water Quality Criteria. This high tolerance of C. riparius to metal toxicity, combined with an exceptional ability to accumulate and tolerate high internal metal burdens makes the chironomid an ideal organism to use in studies on factors affecting the trophic transfer of Cd. Zebrafish were fed with Cd-contaminated chironomids for 7 days, followed by 3 days of gut clearance with clean chironomids. Chironomids loaded with Cd by exposure to Cd-contaminated sediments exhibited a significantly higher trophic transfer efficiency (TTE) than did zebrafish 1 fed chironomids contaminated with Cd by waterborne exposure, although in both cases the TTE's were low (<2%). The majority ofCd transferred to zebrafish was stored in the gut and carcass, irrespective of ingestion of a natural diet (chironomids loaded with Cd), or a manufactured pellet diet of identical Cd concentrations. On a tissue concentration basis, the highest tissue accumulations in zebrafish were (in decreasing order): kidney> gut> liver> gill> carcass; this accumulation pattern w;ls also independent of concentration of Cd in the diet or of prey exposure route. Subcellular fractionation of chironomids found most of the Cd in the metal rich granule fraction followed by the organelle fraction. It also revealed that sediment exposed chironomids had significantly more Cd in the metallothionein-like protein fraction, and significantly less Cd in the cellular debris fraction than water-borne exposed chironomids, although these fractions accounted for only a small percent (~7%) of the total accumulated Cd. Despite this difference in prey subcellular fractionation, the subcellular storage of Cd in zebrafish fed on sediment-exposed chironomids and zebrafish fed on water-borne exposed chironomids was the same, with the highest accumulations in the organelles, enzymes, and metal rich granules fractions. Main areas of subcellular storage in zebrafish fed on a manufactured pellet food were identical to those ofzebrafish fed on chironomids. However, zebrafish fed on chironomids had significantly more Cd in the metallothionein-like protein fraction (5-10%) than did zebrafish fed on pellets :o%). Overall, TTE's were independent of concentration, but were dependent on route of prey exposure. Tissue-specific accumulations and tissue-specific distributions in zebrafish were independent of both concentration and route of prey exposure, and the gut consistently accounted for the highest proportion of overall body burden, and had the highest accumulation of all tissues; verifying its importance in preventing the internalization of Cd. Subcellular distributions were also independent of concentration, although zebrafish fed a biological food (chironomids) had more Cd in the metallothionein-like protein fraction than zebrafish fed a manufactured food. This has biological implications for Cd detoxification. / Thesis / Candidate in Philosophy