Bart Wanten

• 19/12/2024
  
• 3 p.m.
• Venue: Campus Drie Eiken, S.01
• Supervisor: Annemie Bogaerts
• Department of Chemistry

Abstract

Several established processes face a significant cost to capture and store CO2, required to avoid its emission. Therefore, processes that are electricity driven, coupled to a renewable energy source, while providing a pathway to valorise CO2 and other greenhouse gases, are required.

In this thesis, I focus on the combined conversion of CO2 and CH4 (the so-called dry reforming of CH4) by means of a plasma reactor, in particular a confined atmospheric pressure glow discharge. Specifically, I conducted experiments for a wide range of conditions and additive gases, with the goal to find the optimal reaction conditions with this specific plasma reactor, that leads to the formation of syngas, a mixture of CO and H2. Syngas can be further processed into a variety of liquid fuels, such as methanol and various long-chain hydrocarbons, depending on the relative amount of CO and H2. An excellent performance compared to the state-of-the-art is achieved. Specifically, a CO2 and CH4 conversion of 64 % and 94 % is reached, respectively, at an energy cost of 3.2-4 eV/molecule. The confinement in the plasma is assumed to play a crucial role with regards to these excellent results. Unfortunately, the product output’s value remains limited, as soot formation limits the use of sufficiently high amounts of CH4 with this setup, necessary to achieve a syngas ratio (H2/CO) that is high enough for further processing towards various liquid chemicals.

A potential solution for this issue is the addition of other gases next to CO2 and CH4 that can avoid the formation of soot particles to some extent. When a small fraction of O2 is added (the so-called oxidative CO2 reforming of CH4), the syngas ratio is only slightly improved. The bi-reforming of CH4, where H2O is added instead of O2, effectively helps in counteracting soot formation. Moreover, a syngas ratio of 2 is reached, which is ideal for further methanol synthesis.

Finally, these processes are compared in terms of environmental impact and economic potential. Oxidative CO2 reforming of CH4 is found to be the most promising process, with a score twice as high as for the bi-reforming of CH4 process on almost all environmental parameters considered. Compared with current syngas production technologies, all three plasma processes lead to an improved environmental performance. These results show the potential of a bright future for plasma-based syngas production.

Sofie Gielis

• 10/12/2024
  
• 2 p.m.
• Venue: Stadscampus, S.B003
• Supervisors: Kris Laukens & Pieter Meysman
• Department of Computer Science

Abstract

The human immune system is an ensemble of complex interactions between different cell types. The focus of this dissertation lies on T cells. These form an important part of the adaptive immune system, i.e. the group of cells that learns the identity of foreign particles or aberrant cells over time by building up a memory. Each T cell carries multiple copies of a receptor in their membrane: the T cell receptor or TCR. These receptors can recognize small amino acid sequences that are presented by cells. In the case of cancer cells, these epitopes might be tumor specific and thus provide a way for T cells to specifically kill cancerous cells. Since every protein made by or entering our bodies can be cleaved down into multiple peptides and presented in the form of epitopes, the number of possible epitopes is extremely diverse. As a consequence, our T cell compartment must have a very diverse TCR repertoire. To understand and monitor the immune responses in cancer patients, it is crucial to identify T cells targeting tumor-specific epitopes. Due to the high diversity at both ends of the story, the identification of these T cells is highly complex. The goal of this dissertation is to improve this identification process by providing new computational strategies. More specifically, we introduce prediction models to directly identify epitope-specific TCRs within repertoire data. ​ ​

We will start by explaining how these models were trained and how they can be used to monitor and improve our understanding of TCR repertoires. Hereafter, we demonstrate how the prediction models can reveal new insights in the T cell compartment when monitoring the response of cancer patients during immunotherapy. The main finding of this study is that our prediction framework can reveal patterns in the TCR repertoire that are linked with clinical response. This opens the door for computational predictions in the evaluation and design of immunotherapy. ​ ​

Altogether, we introduce epitope-TCR prediction models to understand the human TCR repertoire, especially during cancer immunotherapy. ​ ​

Anwar Hiralal

• 25/11/2024
  
• 3 p.m.
• Venue: Stadscampus, Klooster van de Grauwzusters - Promotiezaal
• Supervisors: Filip Meysman & Jeanine Geelhoed
• Department of Biology

Abstract

Cable bacteria are filamentous, multicellular microbes that conduct electrical currents across centimeter-scale distances—extending previously known limits of biological electron transport by several orders of magnitude. By transferring electrons from cell to cell, these bacteria can connect spatially separated electron donors and acceptors, establishing a unique redox metabolism with distinct anodic (electron-donating) and cathodic (electron-accepting) zones along the filament. This spatially organized metabolism is unprecedented in biology.

To uncover the genetic basis of these processes, obtaining complete cable bacteria genomes is crucial. However, since their discovery in 2012, genomic data have been limited to incomplete and fragmented assemblies consisting of multiple segments, despite bacterial genomes generally consisting of a single circular chromosome. This challenge stems from cable bacteria being complex environmental organisms that cannot be cultured in isolation, complicating genome sequencing. To address this, we developed a novel approach called "targeted metagenomics”, which allows for genome closure of complex unculturable microorganisms like cable bacteria. Using this method, we successfully generated complete genomes for several cable bacteria species. ​ ​

Having complete genomes allowed us to probe deeper into cable bacteria’s unique metabolism. First, we identified unique genetic adaptations enabling efficient nickel cycling within cells, specifically by transporting large quantities of nickel to the periplasm. This finding aligns with prior studies showing that cable bacteria have a unique nickel cofactor located in electricity-conducting fibers within the periplasm. We also used the genomes to examine the sulphur metabolism genes essential for the anodic half of the redox process, which is crucial to the bacteria’s spatially-separated metabolism. Finally, we identified a “unique core genome”—a set of genes shared across all cable bacteria species yet not found in related organisms. This core set likely holds the key to understanding the mechanisms behind long-distance electron transport, and provides a foundation for future biochemical studies to further unravel how cable bacteria achieve this remarkable capability. ​ ​

Ultimately, this research deepens our understanding of the molecular adaptations enabling cable bacteria to conduct electricity, revealing insights into biological electron transport that could one day lead to biotechnological applications, such as sustainably produced electric wires made of biological material. ​ ​

Ine Arts

• 25/11/2024
  
• 2 p.m.
• Venue: Campus Groenenborger, U.024
• Supervisor: Dirk Lamoen
• Department of Physics

Abstract

Uranium dioxide (UO₂), the primary material in nuclear fuel, is prone to oxidation, which leads to higher oxides like U₃O₈ over time. This oxidation process has significant implications for fuel stability, density, thermal conductivity, and the behavior of fission products—all critical for the safe handling and long-term storage of spent nuclear fuel. In this thesis we explore the oxidation of UO₂ at an atomic scale using advanced computational methods based on density functional theory (DFT). ​ ​

UO₂’s complex magnetic and electronic properties require precise modeling techniques, which are first thoroughly studied in this work in order to accurately describe the material. In early oxidation, oxygen enters as interstitial defects. By focusing on how single oxygen atoms and molecules, like oxygen, water and hydrogen peroxide, interact with the UO₂ surface, we gain insight into the initial stages of the oxidation. ​ ​

As oxidation continues, UO₂ transforms into intermediate compounds where uranium atoms exhibit mixed oxidation states. Interstitial oxygen diffuses into the material, which is enhanced by the presence of grain boundaries —regions where different crystals of the same material meet. Using DFT, the possibility of oxygen accumulation in these UO2 grain boundaries is studied. ​ ​

The study also characterizes U₃O₇, the intermediate phase leading to U₃O₈, using both computational and experimental data. Our findings challenge previous assumptions by identifying a fraction of the uranium atoms in a +VI oxidation state within U₃O₇, which affects its electronic and magnetic behavior. Additionally, a new method for identifying oxidation states of specific atomic sites through DAFS experiments is proposed, laying groundwork for future experimental validation of these computational insights. ​ ​

Overall, this work enhances our understanding of UO₂ oxidation, contributing valuable knowledge for the safe and sustainable management of nuclear fuel.

Yuman Qin

• 22/11/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, V.008
• Supervisor: Shoubhik Das
• Department of Chemistry

Abstract

The advancement of sustainable technologies is imperative for mitigating global environmental issues. Photoredox catalysis leverages light to facilitate chemical reactions, representing an energy-efficient and environmentally benign strategy. Complementary to this, circular chemistry (CC) provides a guiding framework for chemical researchers, emphasizing resource efficiency and waste minimization, which is crucial for driving a sustainable future. By converting plastic waste into valuable resources, advanced catalytic processes like photoredox catalysis can substantially lower environmental impact. Furthermore, integrating CO2 capture technologies with CC principles is vital for combating climate change by reducing atmospheric CO2 levels. Additionally, the direct oxidation of C(sp3)–H bonds using O2 as an oxidant is a significant process in fine chemical industries, underscoring the importance of efficient and sustainable oxidation processes. This thesis investigates the synergistic potential of photoredox catalysis, circular chemistry, plastic waste valorization, CO2 capture, and efficient oxidation processes, illustrating how these methodologies can effectively reduce environmental impact, optimize resource use, and support sustainable development. The thesis is divided into four chapters:

• Chapter 1: An overview and introductory exposition of the fundamental principles and concepts of photoredox catalysis are provided. 

• Chapter 2: We have developed a metal-free photocatalytic system for the valorization of plastic waste, specifically targeting 13 different polystyrene-based plastics that are commonly found in daily life. These plastics were successfully converted into benzoic acid. Subsequently, benzoic acid was transformed into aromatic building blocks such as benzene, toluene, salicylic acid, and biphenyl. This approach demonstrated significant potential in addressing the challenges posed by plastic pollution and the requirement for the production of aromatic compounds. 

• Chapter 3: We have pioneered a methodology for utilizing CO2 from exhaust gases to synthesize γ-lactam to mitigate atmospheric CO2 accumulation. This approach boasted a wide substrate scope and high functional group tolerance. Furthermore, it enabled the construction of an γ-lactam core between diverse bioactive compounds, highlighting substantial potential for drug design applications. 

• Chapter 4: We have developed an efficient single-atom heterogeneous photocatalyst for the oxygenation of allylic C-H bonds by utilizing O2 as the oxidant which did not require an additional hydrogen atom transfer (HAT) agent. This reaction system has been meticulously engineered, demonstrating excellent substrate compatibility and functional group tolerance, which underscores its industrial applicability and validates the significance of this approach. ​ ​

Peng Ren

• 21/11/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, T.148
• Supervisor: Shoubhik Das
• Department of Chemistry

Abstract

This thesis explores the development and application of novel heterogeneous photocatalysts for environmentally sustainable production of hydrogen peroxide (H₂O₂) and CO₂ reduction. Utilizing sunlight as an energy source, these photocatalytic processes aim to address the global energy demand and reduce dependence on non-renewable energy sources while mitigating climate change impacts. Conventional methods of H₂O₂ production rely on hazardous chemicals, generating considerable waste. In contrast, photocatalytic H₂O₂ synthesis presents a cleaner alternative, with fewer harmful by-products. Similarly, the rising CO₂ levels, primarily due to fossil fuel combustion, contribute significantly to global warming. Photocatalytic CO₂ reduction, integrating Carbon Capture and Utilization (CCU) strategies, offers a promising solution by capturing CO₂ from exhaust gases and repurposing it for various applications, thus reducing emissions.

The thesis begins with an introduction to semiconductor photocatalysts, with a focus on BiOBr-based, single-atom, and g-C₃N₄-based catalysts, covering their synthesis, photocatalytic applications, and associated reaction mechanisms. Subsequent chapters delve into the specific catalysts developed for H₂O₂ production and CO₂ reduction.

In Chapter 2, a lignin-supported BiOBr photocatalyst (LBOB) is presented, designed for direct H₂O₂ production from seawater. Various characterization techniques (e.g., XRD, SEM, UV-Vis, ssNMR, XPS) were employed to analyze its structural and optical properties. Results indicated that lignin enhanced the photocatalytic performance by providing structural stability, lowering reduction potential, and facilitating electron transfer via functional groups acting as electron sinks.

Chapter 3 introduces a manganese-based single-atom photocatalyst optimized for H₂O₂ production via water oxidation. The catalyst's structural and electronic properties were characterized using UV-Vis, EPR, XANES, EXAFS, HRTEM, XPS, and ssNMR. The Mn centers played a pivotal role in generating hydroxy radicals (•OH), crucial intermediates for H₂O₂ synthesis.

Chapter 4 examines an iron-based photocatalyst for converting CO₂-rich gas streams into valuable products, demonstrating an impressive carbon monoxide (CO) production rate. This catalyst’s efficiency highlights its potential for practical CCU applications, leveraging sunlight for sustainable chemical transformations.

These studies underscore the potential of advanced photocatalysts in achieving efficient, sustainable solutions for H₂O₂ production and CO₂ reduction, aligning with global environmental goals.

Stijn Verschueren

• 14/11/2024
  
• 5.30 p.m.
• Venue: Campus Drie Eiken, O.06
• ​Online PhD defence​
• Supervisors: Herwig Leirs, Hans Bauer, Bogdan Cristescu & Laurie Marker
• Department of Biology

Abstract

Biodiversity is declining globally, with many species facing threats from habitat loss and human encroachment. While conservation strategies often focus on charismatic species, it is key to evaluate whether such focus complies with the broader principles of biodiversity conservation. The cheetah is renowned for its speed and elegance, and stands as a flagship species for grasslands and savannas, particularly in Africa. Despite its global recognition, research on cheetahs has been localized, leaving many populations understudied. In this context, we first explore the role of the cheetah as a flagship species, highlighting the need for a more comprehensive approach to protect both the species and its habitat. Cheetahs inhabit ecologically valuable areas, particularly were human impact is high and protection low. By identifying and mapping conservation priorities, efforts can be targeted more effectively, while contributing to broader biodiversity targets.

Next, we explore methods to improve population monitoring efforts. This is an important aspect for understanding and managing cheetah populations, which are extremely wide-ranging and exist at low population densities, and thereby often remain undetected. Through the application of non-invasive survey techniques such as detection dogs and camera traps, significant improvements were made in detecting cheetahs. These methods offer cost-effective means while providing insights into their spatial ecology and population status. Moreover, the utilization of pattern recognition algorithms for individual identification enhances population estimates by reducing identification efforts. Adopting these innovative approaches can streamline monitoring efforts and contribute to obtaining more accurate data for informed decision-making.

Finally, we present four case studies from different cheetah populations at local and regional scales. The case studies focus on distribution patterns and habitat associations, with implications extending to the broader carnivore guild and the conservation of biodiversity. These studies identify conservation opportunities across human-impacted landscapes, such as the conservation potential of specific landscapes in eastern Namibia and the ability of subordinate carnivores to persist outside of national parks.

The research findings highlight several implications for biodiversity conservation. Firstly, integrating species-centric approaches with broader conservation strategies is crucial for effectively safeguarding flagship species while preserving biodiversity. Secondly, the adoption of innovative survey techniques can improve monitoring efficiency and contribute to new insights into carnivore ecology and conservation. Lastly, the case studies emphasize the need for adaptive management approaches that accounts for local environmental conditions and social needs.

Charlotte Adams

• 13/11/2024
  
• 3 p.m.
• Venue: Stadscampus, D.013
• Supervisors: Wout Bittremieux, Kurt Boonen & Kris Laukens
• Department of  Computer Science

Abstract

Mass spectrometry is a powerful tool for identifying proteins and peptides from biological samples. However, annotating the spectra from a mass spectrometry experiment can be challenging, leaving a significant portion of spectra unexplored. During this PhD thesis, we tackled the challenge of annotating difficult-to-identify spectra in order to extract the hidden knowledge they contain, which could lead to new insights into the underlying biology.

First, we focused on identifying post-translational modifications (PTMs). These are alterations to proteins that can alter their function. Even though there are many different types of PTMs that could be studied, typically only a few are considered when spectra are searched against a reference database. For each PTM that is considered, both the modified and unmodified peptide must be included in the reference database. To prevent the database from becoming too large, which would increase the likelihood of incorrect peptide-spectrum matches (PSMs), researchers are usually limited to investigating only a few PTMs at a time. In contrast, open modification searching enables a peptide with a PTM to match against a reference peptide without the PTM by allowing a mass shift. This approach allows us to investigate any PTM without explicitly defining it and without expanding the reference database.

Another field that suffers from large amounts of unidentified spectra is that of immunopeptidomics. Immunopeptides are peptides on the cell surface that can elicit an immune response. They are frequently studied in the context of infectious diseases or cancer, and their identification is critical for the development of vaccines and immunotherapies. Unfortunately, a lot of spectra in immunopeptidomics research remain unidentified due to intrinsic differences between tryptic peptides and non-tryptic immunopeptides, in addition to a lack of clear digestion rules, resulting in large reference databases and low identification rates. We were able to significantly improve the reliability and number of identifications, by applying PSM rescoring with deep learning-based predictions. Finally we examined privacy concerns related to proteomics data. While privacy implications have been thoroughly explored for human genetics, they have only recently gained attention in proteomic data. To investigate whether it is possible to identify an individual using their raw mass spectrometry data, we propose an approach, inspired by facial recognition, that would look directly at LC-MS profiles.

Joachim Slaets

• 08/11/2024
  
• 2 p.m.
• Venue: Campus Drie Eiken, O.01
• Supervisor: Annemie Bogaerts
• Department of Chemistry

Abstract

Global CO2 concentrations in the atmosphere have reached unprecedented levels, driven primarily by anthropogenic emissions. This alarming rise in greenhouse gases (GHGs) presents a significant challenge to global climate stability, with CO2 being the primary contributor to climate change. Industrial activities are major sources of these emissions, highlighting the urgent need for innovative and sustainable solutions.

Plasma technology emerges as particularly promising, which creates a highly reactive environment through the presence of high-energy electrons. By leveraging such processes, CO2 can be transformed into useful chemicals, contributing to both emissions reduction and resource circularity. One interesting reaction, which can be carried out in a plasma environment, is the dry reforming of methane (DRM), a process that utilizes CO2 and methane (CH4) to produce a mixture of carbon monoxide (CO) and hydrogen (H2). These are valuable intermediates for further chemical synthesis, which can be used to synthesize a variety of chemicals and fuels.

Through chemical kinetics modelling a wide range of conditions is explored to better understand the core chemical kinetics of DRM in warm plasmas. Thereby examining the performance of the process across a wide temperature range and highlighting the limitations of various gas mixtures. The findings demonstrate where plasma-specific kinetics diverges from thermal gas-phase chemistry, offering new insights into the unique behaviour of plasma-driven reactions. Also, the effect of nitrogen (N2) on plasma-based DRM is investigated through computational modelling to support experimental results and demonstrate the role of N2 in the conversion process within a gliding arc plasmatron (GAP) reactor. Revealing a small fraction of N2 can improve the process.

Furthermore, after the plasma has converted the gas molecules, further chemical changes can still occur, influencing the overall efficiency and product distribution. The model demonstrates that quenching the gas temperature does not generally improve performance, except in CO2-rich mixtures where certain reactions are influenced by the cooling process, leading to notable changes in the product distribution. The benefits of combining the hot plasma effluent with unconverted gas are also explored, as the residual heat from the plasma can be reused to drive additional reactions, thereby improving the overall efficiency of the process.

The findings presented improve the understanding of plasma-based DRM technology, forming a foundation for further experimental and theoretical studies, and contribute to the broader goal of reducing GHG emissions and supporting the transition to a sustainable, low-carbon future.

Glen McLachlan

• 05/11/2024
  
• 5 p.m.
• Venue: Stadscampus, Promotiezaal van de Grauwzusters, Lange Sint-Annastraat 7, 2000 Antwerpen
• Supervisors: Herbert Peremans & Bart Partoens
• Department of Physics

Abstract

While the human auditory system is proficient on its own at discerning the direction of incoming sounds, it operates in concert with other sensory modalities to reach accurate spatial awareness. Many studies have investigated the integration of auditory and visual information, but much less attention has been given to the importance of proprioceptive and vestibular information in the localisation process. The vestibular and proprioceptive systems aid in discerning self-motion from source motion and, through this, can stabilise perception and provide additional cues for sound localisation.

The aim of this PhD thesis was to better understand how head movement and position estimation affects sound localisation. To this end, an ideal-observer model, based on Bayesian principles, was developed as a tool to predict dynamic sound localisation in humans and to test how performance is affected by the available information. 

Behavioural experiments were conducted in conjunction with model simulations to determine the acoustic cues and head motions that are relevant to dynamic sound localisation. The results from the psychoacoustic experiments were found to be in general agreement with the model output, though some quantitative differences indicated that dynamic sound localisation may involve processes that can be considered non-ideal. 

These studies offer valuable insights for the field of psychoacoustics and for auditory engineering applications in modern technologies such as hearing aids and virtual or augmented reality.

Fahim Faraji

• 04/11/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, V.008
• Supervisors: Erik Neyts, Mehdi Neek-Amal & Milorad Milosevic
• Department of Physics

Abstract

This thesis initiates an inquiry into the intricate interplay between confinement and material behaviors, addressing cutting-edge topics in nanomechanics and interfacial physics. Each section explores how confinement significantly alters material properties compared to their bulk counterparts, extending the investigation into the nuanced realm of thermodynamic properties at interfaces. ​ ​

We first demonstrate the breakdown of a previously acknowledged universal aspect ratio (height versus diameter) in nanometer-sized bubbles within graphene, laying the groundwork for a detailed examination of adhesion energies. Further, the indentation of graphene nanobubbles reveals failure points reminiscent of viral shells through analysis using the Föpple–von Kármán (FvK) dimensionless number. Additionally, phase transitions of encapsulated noble gases are explored, exhibiting intriguing behaviors under varying temperatures. ​ ​

The formation of anomalous shapes in flat nanobubbles encapsulated by hexagonal boron nitride is also investigated, highlighting the influence of heating rates and hydrogen bonding. The cation-controlled permeation of charged polymers through nano-capillaries is examined, revealing distinct effects of monovalent cations on polymer transmission speed. The ability to manipulate permeation is elucidated based on the differing surface versus bulk preferences of various alkali cations in the presence of an external electric field, offering valuable insights into the interplay between ionic dynamics and nano-confinement effects. ​ ​

The exploration continues with an assessment of the accuracy of the Kelvin equation in nanoscale capillaries, proposing a revision based on disjoining pressure. Finally, critical commentary on the Shuttleworth equation corrects misconceptions and contributes to a comprehensive understanding of interfacial thermodynamics.

Ewoenam Kwaku Tokpo

• 30/10/2024
  
• 4 p.m.
• Venue: Campus Middelheim, G.010
• Supervisor: Toon Calders
• Department of Computer Science

Abstract

As natural language-based technologies continue to develop and play a prominent role in society, increasing attention is being paid to the ethical issues constraining their use, with bias being a prominent concern. There is a growing body of evidence highlighting biases, such as gender bias, within natural language models. Although considerable work has been done to understand and address this issue, significant challenges remain regarding how to detect, measure, and effectively mitigate bias. ​ ​

This thesis addresses two main themes. The first part primarily involves empirical investigations of various existing techniques and approaches for detecting, measuring, and mitigating bias in natural language processing (NLP). The second part focuses on developing solutions to mitigate bias in language-based technologies and human-generated biases. ​ ​

We first investigate existing techniques to measure bias in natural language models. Specifically, we review the literature on fairness metrics for pre-trained language models and empirically evaluate their consistency and compatibility. We investigate how various factors, such as templates, attribute and target seeds, and the choice of embeddings used by existing techniques, affect how bias is quantified. ​ ​

Secondly, we investigate the relationship between bias in pretrained language models and fine-tuned language models for downstream applications. We design a probe to investigate the effects that some of the major intrinsic gender bias mitigation strategies have on downstream text classification tasks. We discover the propensity for some intrinsic bias mitigation techniques to hide bias instead of resolving it and show inconsistencies in how bias measuring techniques measure bias with respect to certain mitigation techniques. We also find that bias inherent in a pretrained model has little material effect on downstream fairness. ​ ​

Thirdly, we develop an automated approach to generating parallel data for training counterfactual text generator models for counterfactual data augmentation (CDA) that limits the need for human intervention. Although CDA has been a widely used mitigation strategy in NLP, existing works have significant issues, which we also highlight in this thesis.

Finally, we propose a text style transfer technique to automatically mitigate bias in textual data. Our text-style transfer model can be trained on non-parallel data. We demonstrate that our approach overcomes the limitations of many existing text style transfer techniques.

​

Daphne Vandemeulebroucke

• 22/10/2024
  
• 4 p.m.
• Venue: Campus Middelheim, G.010
  
• ​Online PhD defence​
• Supervisor: Joke Hadermann
• Department of Physics

Abstract

Climate change is one of the biggest challenges for science in the twenty-first century, and the development of new technologies that lower the carbon footprint has never been more pressing. Many of these alternative energy applications involve inorganic crystalline solids exposed to redox conditions. In this research area, the development of in situ 3D electron diffraction at elevated temperatures in gas and vacuum unlocks great new potential, as it can unravel atomic structure transformations in operational conditions. That way, we can investigate the origins of e.g. high performance and degradation properties. A particularly interesting group of energy materials is based on the perovskite structure, with ABX₃ as chemical formula. This project studied the influence of high temperature reducing atmospheres on the atomic structure of two perovskite-based energy materials : La_xSr_2−xMnO_4−𝛿 and (Ca,Sr)(Fe,Mn)O_3−𝛿.

The first group of materials – i.e. Ruddlesden-Popper La_xSr_2−xMnO_4−𝛿  – have good performance properties as symmetric solid oxide fuel cell electrodes. With 3D electron diffraction, we found new incommensurately modulated structures for La_0.25Sr_1.75MnO_4−𝛿 and La_0.5Sr_1.5MnO_4−𝛿 upon annealing in hydrogen gas, which have never been reported before. Further, we discovered differences in defect structure and ordering that can be linked to previously unexplained differences in electrical conductivity with lanthanum concentration. ​ ​

Next, (Ca,Sr)(Fe,Mn)O_3−𝛿 was studied, which is an oxygen carrier in the CLOU process. This is a carbon capturing combustion technique that inherently separates expelled CO₂ from air. The performance properties of CaMnO_3−𝛿 can be improved by A or B site doping with e.g. strontium or iron. Using in situ 3D electron diffraction, we discovered that this is connected to differences in ordering of oxygen vacancies for the undoped versus doped materials. Further, we could successfully solve and refine the structure of CaMnO_2.75 for the first time ever.

However, at the highest temperatures during in situ experiments in gas atmospheres, a reaction of the samples with the Si₃N₄ heating chips gave rise to a SiO₂ shell around the crystals. This caused experimental limitations and discrepancies between in situ and ex situ annealing results. Therefore, a systematic investigation was performed into the potential of a protective graphene coating.

This research illustrated how a combination of various in situ and ex situ electron microscopy techniques led to new insights in the structure-property relations of redox-based energy materials.

Morgana Sluydts

• 18/10/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.03
• Supervisor: Floris Wuyts
• Department of Physics

Abstract

The goal of this doctoral thesis was to evaluate possible new treatment options for patients suffering from bilateral vestibulopathy (BVP). The treatment options under investigation were vestibular co-stimulation by means of a cochlear implant (CI), electrical vestibular stimulation by means of a cochleovestibular implant (CVI), and subconscious haptic feedback.

The only evidence for possible vestibular co-stimulation was observed during the neural response telemetry (NRT) in the patients who were implanted with the CVI. High current levels were required to evoke such co-stimulation. However, the current levels used during daily life for restoring auditory and vestibular functions, were much lower than those at which possible co-stimulation was observed. Therefore, it is unlikely that such co-stimulation would occur during daily usage of the CVI. An alternative explanation for improved posture and gait in regular CI recipients may be the electrically restored auditory, directional cues in a patient’s surroundings.

Treatment with subconscious haptic feedback showed mixed results. Some patients indeed performed better on objective testing when they received higher levels of subconscious haptic feedback, but the perception of mobility and balance did not improve as well.

The lack of convincing results with subconscious haptic feedback and vestibular co-stimulation supports the need for the development of the CVI. Our results showed that CVI stimulation significantly improved static posture and dynamic gait in severely hearing-impaired patients suffering from chronic BVP, albeit not in all patients. Future research is required to gain insight as to why some patients improved while others did not.

In our prevalence study in a tertiary referral ear-nose-throat center, it was shown that 45% of the patients that were considered eligible for a vestibular implant had substantial, functional hearing. As vestibular implantation carries the risk of iatrogenic hearing loss, structure preserving implantation techniques should be developed for stand-alone VIs, which require atraumatic surgery and preservation of hearing.

Lisa Furu Baardsen

• 11/10/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, Q.002
• ​Online PhD defence
• Supervisor: Erik Matthysen
• Department of Biology

Abstract

Urbanization has profound effects on ecological interactions and the composition of various taxa, including arthropod communities. Recent research has endeavored to examine arthropod community responses along urban-rural gradients and found that some of the most important characteristics driving the changes were body size, dispersal ability, and nesting requirements. Parasites are particularly important in anthropogenic environments, affecting host population dynamics and potentially acting as vectors of diseases. Arthropod communities in nestboxes specific for small birds offer an interesting system to study how urbanization affects trophic interactions in general, and host-parasite interactions in particular. This thesis investigates the complex interactions between urbanization and arthropod communities associated with bird nests, specifically focusing on nests of the great tit (Parus major) and the blue tit (Cyanistes caeruleus). We collected data from Flanders, Belgium, a highly urbanized region in a strict sampling design allowing us to disentangle effects of urbanization at different spatial scales and land-use variables. The research sheds light over various aspects of nest arthropod ecology, such as community build-up and diversity, host-parasite dynamics and parasite dispersal.

First, this study sought to investigate how urbanization and land-use variables influenced the diversity and species composition of nest-associated arthropods. This endeavour revealed that while urban settings do not significantly diminish overall arthropod community diversity or richness, idiosyncratic responses at the arthropod group level were observed. Then, the dynamics of nest-arthropod communities at three points during the nesting cycle were studied, demonstrating notable shifts in abundances and trophic structure, showing that for many groups the peak in abundances was seen during the nestling stage. Focus was then shifted to the distribution and ecological requirements of parasitic Dermanyssus mites, highlighting species-specific responses to urbanization and habitat features. Finally, the population genetics of the parasitic mite Dermanyssus carpathicus was explored, revealing moderate genetic diversity and significant population structuring potentially influenced by urbanization and host ecology.

Overall, this study shows how arthropod diversity is maintained in urban communities and highlight the intricate relationships between urban landscapes, host species and their parasites and commensals, contributing valuable insights into host-parasite dynamics in human-altered ecosystems.

Shangkun Li

• 10/10/2024
  
• 1.30 p.m.
• Venue: Campus Drie Eiken, R2
• Online PhD defence
• Supervisors: Annemie Bogaerts & Erik Neyts
• Department of Chemistry

Abstract

The direct conversion of CH4 to value-added chemicals has attracted intensive interest from both academic and industrial communities. Plasma technology is a promising approach to activate gas molecules by electricity instead of heat, and it can be operated at mild conditions and allows easy upscaling. The combination of plasma and catalysis often shows synergy. In order to optimise this syngergistic operation, it is important to understand the plasma and plasma-catalyst interactions at a fundamental level. However, it is not straightforward to reveal the entire mechanism because of the intrinsically highly complex reactions of plasma catalysis. This thesis aims to provide a fundamental understanding of plasma-driven direct CH4 conversion into value-added products using various oxidants (i.e., O2, CO2, and H2O), involving plasma gas-phase reactions and plasma-assisted surface reactions on a catalyst by combining experiments and modeling, which could be of great interest for the application of plasma-based gas conversion at a wider scale, boosting the transition towards a more sustainable energy economy.

Coline Le Noir de Carlan

• 02/10/2024
  
• 3 p.m.
• Venue: Stadscampus, Klooster van de Grauwzusters, Promotiezaal
• Supervisors: Erik Verbruggen & Caroline De Tender
• Department of Biology

Abstract

Climate change is a major threat to all ecosystems, but its consequences are particularly severe in high latitude regions, where warming intensity is the highest. Because soil microbes have a central role in terrestrial ecosystem functioning, understanding their response to warmer conditions is essential. This thesis aims to explore the response of soil microbial communities to warming, with a particular focus on plant-microbe interactions.

We first carried out field experiments in two sites with warming treatments. The first one is a Finnish tundra grassland, where air warming induced by Open-top chambers (ca +1.9), in combination with other environmental changes, has been applied for 10 years. There, warming alone did not affect the general community composition of root-associated fungi, although it clearly affected the relative abundance of mycorrhizal fungi. The other is an Icelandic subarctic grassland, where soils were warmed through geothermal activity for 12 (MTW), and over 55 years (LTW), at an intensity ranging from ca +0.5 to +6°C. There, the community composition of both prokaryotes and eukaryotes shifted along the gradient. Besides, in both setups, plant communities have been manipulated by either adding new species (tundra), or excluding plant roots (subArctic grassland). The outputs were similar to untreated conditions in both studies, suggesting a high resistance of soil microbes to plant changes under field conditions. Furthermore, we investigated the effects of warming on soil health and found that, while the microbial diversity remained stable, the microbial connectivity increased along the thermal gradient. The bacterial predicted functional profiles shifted with increasing temperatures, for which genes involved in carbon degradation specifically decreased under LTW.

Last, we carried out a greenhouse experiment, where we grew grasses in soils inoculated with soil microbiomes conditioned by different warming intensities and found that plants grown in soils with microbiomes conditioned by LTW displayed lower belowground biomass. We identified some microbial taxa known to associate with plants that suggested that our observations could be due to changed relative abundance of e.g. putative plant pathogens or arbuscular mycorrhiza.

Overall, our work shows that soil microbial communities are sensitive to soil warming, but not directly by air temperature. However, we found a modest effect of plant communities on the microbial community assembly, while warming-conditioned microbiomes negatively altered plant performance, reminding the high potential of soil microbes on plants, but also that within plant-microbe interactions, one may mediate the response of the other.

Lien Michiels

• 19/09/2024
  
• 4.30 p.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisor: Bart Goethals
• Department of Computer Science

Abstract

In the current digital landscape, recommender systems play a pivotal role in shaping users' online experiences by providing personalized recommendations for relevant products, news articles, media content, and more. Their pervasive use makes the thorough evaluation of these systems of paramount importance. This dissertation addresses two key challenges in the evaluation of recommender systems. Part II of the dissertation focuses on improving methodologies for offline evaluation. Offline evaluation is a prevalent method for assessing recommendation algorithms in both academia and industry.

Despite its widespread use, offline evaluations often suffer from methodological flaws that undermine their validity and real-world impact.

This dissertation makes three key contributions to improving the reliability, internal and ecological validity, replicability, reproducibility, and reusability of offline evaluations.

First, it presents an extensive review of the current state of practice and knowledge in offline evaluation, proposing a comprehensive set of better practices to address the reliability, replicability, and validity of offline evaluations.

Next, it introduces RecPack, an open-source experimentation toolkit designed to facilitate reliable, reproducible, and reusable offline evaluations.

Finally, it presents RecPack Tests, a test suite designed to ensure the correctness of recommendation algorithm implementations, thereby enhancing the reliability of offline evaluations. Part III of the dissertation examines the measurement of filter bubbles and serendipity.

Both concepts have garnered significant attention due to concerns about the potential negative impacts of recommender systems on users of online platforms.

One concern is that personalized content, especially on news and media platforms, may lock users into prior beliefs, contributing to increased polarization in society.

Another concern is that exposure only to content previously expressed interest in may lead to boredom and eliminate surprise, preventing users from experiencing serendipity.

This research makes three contributions to the study of filter bubbles and serendipity. First, it proposes an operational definition of technological filter bubbles, clarifying the ambiguity surrounding the concept.

Second, it introduces a regression model for measuring their presence and strength in news recommendations, providing practitioners with the tools to rigorously study filter bubbles and gather real-world evidence of their (non-)existence.

Finally, it proposes a feature repository for serendipity in recommender systems, offering a framework for evaluating how system design can influence users' experiences of serendipity in online information environments. In summary, the findings and tools developed in this dissertation advance the theoretical understanding of recommender system evaluation while offering practical tools for industry practitioners and researchers.

Ceder Dens

• 18/09/2024
  
• 4 p.m.
• Venue: Stadscampus, C.002, Prinsstraat 13, 2000 Antwerpen
• Supervisors: Kris Laukens & Wout Bittremieux
• Department of Computer Science

Abstract

Understanding the immune system is of utmost importance for advancements in biomedical research and human healthcare. Deep learning has the potential to uncover the hidden secrets of proteins and peptides, leading to significant improvements in diagnostics, vaccine development, and cancer therapies. In this thesis, we developed and applied various deep learning techniques to get insights from peptides that drive the adaptive immune response. 

We investigated pathogen recognition and immune response by looking into T-cell receptor and epitope interactions. We apply interpretable deep learning to interaction prediction models and link this to the three-dimensional structure of the molecules to get insights into the factors determining T-cell–epitope binding affinity. Additionally, our results show the importance of using interpretability techniques to verify machine learning models and avoid that small hard-to-detect problems can accumulate to inaccurate results. 

We address the issue of data bias in machine learning. We show how it can lead to overly optimistic performance that cannot be reproduced on real-world data. We underscore the necessity of rigorous data evaluation and advocate for the use of unbiased benchmarking datasets to ensure generalizability and applicability of prediction models. 

We develop a novel transformer-based machine learning model for interaction prediction of biological sequences. This architecture encodes the interaction between protein or peptide sequences in a biologically meaningful way while also providing valuable visualizations for model interpretability. 

We end by focussing on boosting the identification rate of peptides, such as those recognized by the immune system. We find that the training dataset size can have a substantial impact on the performance of machine learning models used in computational proteomics. This underscores the necessity for high-quality, comprehensive, standardized datasets to train robust machine learning models. To address data scarcity, we explore algorithmic strategies such as self-supervised pretraining and multi-task learning. We find that self-supervised learning can be a very valuable technique and hypothesize that the benefits of multi-task learning could become more apparent when used in combination with more comprehensive datasets for all peptide properties. 

Artificial intelligence shows great potential to revolutionize biomedical research, however, as shown in this thesis, large, unbiased, high-quality datasets are required for it to make an impact. The findings of this research can hopefully make progress towards more accurate, interpretable, and reliable predictive models, which are crucial for future breakthroughs in diagnostics, therapeutic development, and personalized medicine.

Steven Mortier

• 17/09/2024
  
• 4 p.m.
• Venue: Campus Middelheim, G.010
• ​Online PhD defence​
• Supervisors: Tim Verdonck & Steven Latré
• Department of Computer Science

Abstract

In recent years, the rapid increase in available data, accompanied by significant algorithmic advancements, have enhanced our ability to analyze and extract valuable insights from this data. In this thesis, we present novel applications of machine learning (ML) in biological sciences, more particular in neuroscience, ecology, and agricultural research. Throughout the thesis, we pay special attention to the interpretability and explainability 0of our models, ensuring that we understand the reasoning behind their predictions. ​ ​

The thesis is divided into four parts, each presented in a separate chapter. In Chapter 2, we address a binary classification task aimed at detecting attention in electroencephalography (EEG) data. By employing state-of-the-art ML models, we successfully differentiate between target and distractor stimuli using EEG data collected during an audiovisual attention task. Additionally, we examine subject dependence in EEG data by comparing the performance of ML models trained on individual subjects versus multiple subjects. Finally, we apply explainable AI (xAI) techniques to identify the features utilized by the models for their predictions, and find that these features align with the expectations of domain experts. 

Chapter 3 focuses on the relationship between soil temperature, various meteorological variables, and vegetation phenology characteristics. Using ML and xAI, we find that rising soil temperatures result in an earlier onset of the growing season for plants. Additionally, our analysis reveals that the meteorological variables have the most significant impact on the vegetation phenology characteristics, while annual variations are primarily driven by changes in the soil temperature. ​ 

In Chapter 4, we address the issue of missing values in sensor data, with a focus on large-scale wireless sensor networks. We evaluate twelve missing value imputations methods, each using different imputation strategies and originating from diverse backgrounds. To enhance the evaluation process, we define a “masked missings” scenario, offering a more realistic assessment compared to the standard practice of using random missings. Our findings indicate that imputation methods explicitly accounting for spatial correlations between sensors generally perform best, and we use these insights to suggest directions for future research. 

Finally, in Chapter 5, we develop a new global dataset considering the historical application of fertilizers. Specifically, we use ML to predict historical fertilizer application at both the crop and country levels, based on a set of features related to crop classes, along with socioeconomic, agrological, and environmental variables. Additionally, we use xAI to identify the most relevant drivers influencing fertilizer application. ​ ​

Chase Donnelly

• ​16/09/2024
  
• 3 p.m.
• Venue: Campus Middelheim, A.143
• ​Online PhD defence
• Supervisors: Kris Laukens, Gerrit Beemster & Han Asard
• Faculty of Science

Abstract

Anthropogenic effects are causing large shifts in our climate, such as increasingly persistent precipitation regimes, leading to longer periods of alternating drought and flooding stress for plant species. While the effects of drought and flooding have been studied, studies are limited when looking into the effects of persistent precipitation regimes through multiple cycles and over a longer time scale. This thesis looks to address this knowledge gap through the use of multiple omics techniques in combination with previously recorded results from the Regime Shift project, which looked to provide a comprehensive gene to ecosystem approach to study the effect of increasingly persistent precipitation regimes on grassland plant species. The main goal of this thesis was to determine changes that take place at the gene expression level of grassland species under different climate change stressors. In this thesis we investigated the changes that take place at the gene expression level of grassland species under increasingly persistent precipitation regimes. Using the techniques discussed above, we compared the response of four grassland species to increased precipitation regimes along with the effect of soil microbiome. We found that plants grown in soil with a history of long precipitation regime exposure showed a shift in future plant stress responses. Soil legacy of extreme precipitation regime increased hormonal and cell wall regulatory response in plants, which we validated through multiple biochemical tests. Here we developed a model for this shift in plant stress response under increasingly persistent precipitation regimes. Finally, we combine the results of the entire Regime Shift study on precipitation regimes, looking from genes to ecosystem scales. Here we evaluate the main takeaways of the four-year project including biodiversity loss, tipping points, and the importance of plant-microbiome interactions in changing climates.

Jakob Struye

• 12/09/2024
  
• 4 p.m.
• Venue: Stadscampus, S.R.004
• ​Online PhD defence​
• Supervisor: Jeroen Famaey
• Department of Computer Science

Abstract

Extended Reality (XR) is finding increasingly widespread acceptance and adoption. Applications range from simple smartphone apps showing what a new piece of furniture would look like in your living room, to immersive simulations training medical personnel. However, the experience is currently far from optimal. To allow for full mobility and extreme quality, content should be generated off-device and then transmitted wirelessly to the users. Ideally, each video frame should be transmitted within milliseconds at effective data rates of gigabits per second, with minimal loss.

A prime enabler for this type of collaborative XR is Millimeter-Wave (mmWave) wireless communications. The bands that mmWave offers, between 24 and 300GHz, can theoretically meet these requirements, although this necessitates addressing challenges not present in the more commonly used sub-6 bands. Specifically, signal strength degrades very easily, meaning wireless signals must be carefully "beamformed" towards intended receivers. Furthermore, orchestrating channel access and video frame generation is challenging. 

In this dissertation, we first assess the performance of existing mmWave hardware, specifically investigating performance during rotational user motion. We conduct this assessment with both consumer-grade off-the-shelf routers and state-of-the-art experimental hardware. Both are based on the IEEE 802.11 Wi-Fi protocol. In addition to the hardware evaluations, we analyze the IEEE 802.11 specification itself. Through highly realistic performance models, we derive the maximal performance mmWave Wi-Fi could offer in an XR scenario. We validate these models through simulations and, where possible, hardware experiments. 

After evaluating the promising-but-lacking performance of current hardware and protocols, we look towards the future. Specifically, we identify a need for effective beamforming approaches within the XR scenario. Current approaches are reactive and are not expected to adapt rapidly enough to rotational user motion. As such, we design, implement and evaluate CoVRage, the first proactive beamforming approach for highly mobile XR devices. By predicting upcoming motion, it forms beams at the XR device that consistently maintain a high gain towards the transmitter during periods of rapid motion. 

Finally, we combine the different branches of work above. We present an end-to-end system approach for high-fidelity collaborative XR. We perform extensive and highly realistic simulations of the full system, demonstrating that the aforementioned contributions enable consistently high-fidelity collaborative XR, even during rapid and erratic user motion. Overall, we are confident that the work presented in this dissertation brings us one step closer to realizing the goal of hassle-free, truly immersive XR experiences.

Domenico Iuso

• 09/09/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.4
• ​Online PhD defence​
• Supervisors: Jan Sijbers & Jan De Beenhouwer
• Department of Physics

Abstract

Additive Manufacturing (AM), or 3D printing, has revolutionized industrial production by enabling the creation of complex parts and products through the successive layering of materials. This technique offers significant advantages in prototyping and customization, streamlining the transition from digital models to physical objects. However, challenges such as anisotropic properties, surface finish issues, and internal defects necessitate robust quality control measures. 

This thesis investigates X-ray Non-Destructive Testing (NDT) methods, to enhance the reliability and quality of AM components. Key contributions include: 

1) Automatic Simultaneous Multi-Mesh Registration and X-ray system spectral estimation: Introduction of an automated technique for aligning X-ray CT scans with their corresponding CAD models, facilitating precise defect localization and comparison, which can incorporate estimation of the poly-chromatic behaviour of the scanning system. 

2) Compensation for X-ray Scattering: The development of a novel software method to mitigate X-ray scattering effects, thereby improving image clarity and defect detection accuracy, suitable for radiographic and X-ray Computed Tomography (X-CT) setups. 

3) 3D Deep Learning Models: Application of state-of-the-art deep learning methods tailored for 3D defect detection in X-CT images, utilising volumetric data in an efficient, 3D patch-wise approach to identify internal flaws. 

These approaches, addressing critical issues such as X-ray scattering and beam hardening, aim to jointly improve the image quality and ease the 2D/3D comparison of the manufactured object and its digital model. By cross-validating these techniques on real-world data, through a series of experiments, this research makes a step forward towards ensuring the structural integrity and defectiveness of AM samples. The findings contribute to the broader adoption of X-rays inspection setups in AM industries where safety and reliability of samples are critical, such as aerospace, medical devices, and automotive sectors, and where reduction of operational costs is usually desired.

Marjolein Meijdam

• 06/09/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisors: Wendt Müller & Marcel Eens
• Department of Biology

Abstract

Many behavioural and physiological processes fluctuate throughout the day. Such circadian rhythms are predominantly regulated internally by the molecular clock, a negative feedback loop that controls gene expression roughly in a 24 hour schedule. Circadian rhythms are in addition modulated by external factors such as light and temperature (i.e. Zeitgebers) and activity rhythms can be strongly affected by masking factors, which directly affect the timing of behaviour without interacting with the circadian clock. Differences between individuals in the precise duration of the feedback loop or differences in sensitivity to Zeitgebers and masking factors can cause individuals to exhibit differences in activity rhythms, with some individuals tending to be generally active earlier in the day than others. An individual's temporal phenotype is also referred to as its chronotype and individual differences in chronotype are increasingly being reported in free-living species in a variety of taxa. Although the evolutionary history of the development of these differences in chronotype occurred mainly in natural conditions, many habitats today are disturbed by anthropogenic stressors, such as light, noise and chemical pollution, that can affect the timing of activity. In order to understand the possible effects of such stressors on wildlife, it is therefore of utmost importance to first increase our understanding of the functional consequences of differences in chronotypes as well as how and why variation in this trait occurs, as this has remained a mystery to date. Next, it is necessary to investigate how individual activity rhythms are affected by anthropogenic disturbances and what the possible consequences of such disturbances might be. To this end, I studied individual activity rhythms of great tits (Parus major) in a suburban population, by determining their timing of activity onset in the morning and activity offset in the evening.

Altogether, this thesis has provided data that are needed to understand how selection can act on the timing of activities, i.e. chronotypes. It not only highlights the functional consequences of differences between individuals in the timing of behaviour, but it also emphasizes how anthropogenic stressors could impinge on urban birds. This knowledge could be of great importance for a better understanding of the key processes of the impact of urbanization on wildlife and to develop sustainable urban planning frameworks. 

Cindy Lainé

• 03/09/2024
  
• 1 p.m.
• Venue: Campus Drie Eiken, Q.002
• ​Online PhD defence
• Supervisors: Gerrit Beemster & Hamada AbdElgawad
• Department of Biology

Abstract

Maize (Zea mays), essential globally, is highly sensitive to chilling stress (0-15°C). Maize's ability to recover from cold stress plays a pivotal role in the plant's subsequent growth and final yield, especially in regions like Europe, where cold spells are typically transient. This PhD research focuses on the recovery mechanisms of maize leaf growth post-chilling, an area less studied compared to the direct effects of cold stress.

The initial phase involved a meta-analysis spanning 1967 to 2022, comparing growth and biochemical responses in cold-tolerant (Flint) and cold-sensitive (Dent) maize varieties. Key findings revealed that membrane damage via electrolyte leakage effectively indicates cold stress tolerance. Experiments conducted with 4°C treatment for 1-7 days led to more than 50% reduction in shoot dry weight and leaf area, providing a ideal model for studying cold stress.

Subsequently, a new protocol was developed to examine recovery processes in maize leaf growth under controlled and cold stress conditions, using B73 inbred maize plants. This study highlighted differences in recovery rates depending on the developmental stages of the leaves when exposed to cold. Notably, leaves that had emerged displayed more robust recovery than those exposed just prior to emergence. Through kinematic analysis, I discovered that emerged leaves had better recovery due to higher cell division and elongation rates during recovery.

Further analysis on 30 inbred maize lines, subjected to the same cold treatment, identified natural variations in leaf elongation rates and final leaf lengths. A selection of lines for detailed kinematic and biochemical studies revealed that tolerance to chilling was associated with rapid and complete recovery of the leaf elongation rate, marked by heightened cell production and elongation rates, increased soluble sugars, and antioxidant levels.

Finally, I performed a transcriptomic analysis to delve into the molecular mechanisms underpinning recovery. Increased peroxidase activity and jasmonic acid signaling were observed in the transition zone, correlating with slower recovery in non-emerged leaves. While genes regulating cell wall elongation and carbohydrate metabolism were upregulated in the both the meristematic and transition zone, leading to improved recovery in emerged leaves. Overall, this thesis advances the understanding of chilling stress recovery in maize, shedding light on the physiological, cellular, biochemical, and molecular aspects that facilitate resilience in maize leaves against transient cold exposure.

Daphne Lenders

• 03/09/2024
  
• 4 p.m.
• Venue: Stadscampus, S.B.003 - Prinsstraat 13
• Supervisors: Toon Calders & Sylvie de Raedt
• Department of Computer Science

Abstract

Automated decision-making (ADM) systems used in high-stakes areas such as lending or hiring often perpetuate biases present in their underlying data. Consequently, these systems can adversely impact certain population groups, mirroring the sexist or racist practices of our society.

In this thesis, we inspect current approaches to auditing and mitigating such discriminatory biases in ADM systems. We highlight how these approaches typically centre around single definitions of fairness, that aim to express how (un)fair some system is through a single number and try to optimize for fairness accordingly. We explain how these approaches fall short in adequately understanding and resolving discrimination and argue how better approaches should be driven by more nuanced considerations: rather than having one single fairness measure, auditors should focus on which parts of the data a system behaves discriminatory, so that they then can then address this behaviour in a targeted manner.

To that end, our first two chapters focus on new tools and methods for bias detection in ADM systems. The first inspects the potential of interactive auditing toolkits, while the second improves an existing method for measuring individual fairness, allowing auditors to decide for one decision subject at a time whether they received just treatment.

Our third chapter introduces a human-in-the-loop approach to mitigate bias in ADM systems. We design a selective classifier that refrains from making predictions when they are deemed as discriminatory. These rejected instances, along with an explanation for their rejection, can be passed on to human experts who can make better-informed decisions for them.

The fourth chapter shifts focus from new bias mitigation techniques to evaluating their effectiveness. We emphasize how the traditional evaluation scheme, based on single fairness definitions, is not sufficient and instead introduce a benchmarking-dataset to facilitate the evaluation of bias mitigation strategies. This dataset includes a fair and biased version of its decision labels, allowing precise assessment of how well a model can predict the fair labels after being applied on the biased ones. Our fifth and final chapter zooms out from these specific considerations surrounding bias in ADM systems and provides an overview of the research field in general. By highlighting research gaps, we also conclude this thesis with a discussion and its implications for future work.

Renata Turkes

• 03/09/2024
  
• 09.30 a.m.
• Venue: Campus Middelheim, A.143
• Supervisor: Steven Latré
• Department of Computer Science

Abstract

Topological data analysis is a recent and fast growing field that approaches the analysis of (the shape of) data using techniques from algebraic topology. Its main tool, persistent homology, captures information about cycles in the data: its connected components, loops, voids, and so on.

The first part of the thesis investigates the properties of persistent homology. Homology captures cycles in the data, but what additional information is stored in the persistence ("the what")? We show that it can capture the number of holes, but also important geometric notions of curvature and convexity. Stability theorems provide mathematically provable guarantees of desirable properties of persistent homology ("the why"), but how do these theoretical results translate to practice? We show that persistent homology does not always yield noise robust features in a classification task.

The second part explores two applications of persistent homology. Firstly, we demonstrate how it can be used to study the preservation of topology and geometry ("the how"); here we focus on hyperdimensional computing that encodes the input data into a very high dimensional space. Secondly, we extract persistent homology features from EEG data during an audiovisual task in order to detect attention, providing some understanding when it can be useful in neuroscience applications ("the when").

In both parts, we depart from the dominant stream in the literature by highlighting the versatility of filtrations and signatures, the input and output of persistent homology, beyond the canonical choices. In this way, we showcase how persistent homology can be seen as a diverse family of rich descriptors of different aspects of shape.

Jens Renders

• 29/08/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisors: Jan Sijbers & Jan De Beenhouwer
• Department of Physics

Abstract

X-ray computed tomography (CT) is an imaging technique that produces 3D images of objects, with applications in medicine, industry and scientific research. The technique combines multiple X-ray radiographs from different angles to produce this 3D image. Standard CT reconstruction techniques assume that the object does not move during the scan, which is often not valid due to unwanted motion, and sometimes it is the motion itself that is of interest. The subfield of CT that specializes in the imaging of moving objects, called 4DCT, therefore requires specialized acquisition and reconstruction methods. The goal is to reconstruct a full 4D image, which is a sequence of 3D images of frames, depicting the motion during the scan. ​ ​

This thesis focuses on a particular application: the imaging of foam formation, an application that is of interest in industry and material science. It is challenging because foam formation is a fast process, but there are many opportunities for algorithmic improvements in reconstruction techniques, because of the specific characteristics of foams. ​ ​

After an introductory chapter, chapter 2 and 3 present improvements to general 4DCT methods based on image warping. Image warping is a tool to model motion on raster images. It uses an interpolation step, and it is shown that the interpolation method needs to be taken into account when solving for motion and images, to avoid specific artifacts. To this end, efficient implementations of adjoint and differentiated image warping are presented, evaluated and collected in a software package called ImWIP. The package can be applied in other fields than CT, simply by swapping the CT projection operator for a model of another imaging modality. This is demonstrated in multiple applications. ​ 

In chapter 4, another image model than voxel images is considered. Because of the specific structure of foam, the opportunity to use surface meshes instead of raster images arises. Each bubble can be represented by a deformed spherical mesh, of which the shape is determined by a few control points. This leads to the method "BubSub", with many advantages over voxel-based methods: a low number of unknowns, low memory requirements, high noise resistance, and motion does not require interpolation anymore (motion is simply a change in the control points). It is demonstrated that BubSub can successfully be applied to real foam scans, with significant improvements in reconstruction quality.

Miguel Hernando Camelo Botero

• 27/08/2024
  
• 4 p.m.
• Venue: Stadscampus, Klooster van de Grauwzusters, Gebouw S, Lange Sint-Annastraat 7
• ​Online PhD defence
• Supervisor: Steven Latré
• Department of Computer Science

Abstract

As services and networks grow more complex, the need for network management automation rises, encompassing both services and network functions. The primary aim of these systems is to optimize network resources to achieve efficiency while meeting user requirements, leading to the concept of Autonomous Networks (ANs). Advances in Artificial Intelligence (AI) and Machine Learning (ML) enable these networks to manage resources autonomously, learning and adapting to complex environments to perform tasks like self-healing and self-provisioning. 

The radio access domain is evolving into a fully Autonomous Domain (AD), driven by the complexities of wireless networks and the demanding requirements of 5G and future networks. These challenges necessitate advanced management techniques and innovations in spectrum allocation, with Cognitive Radios (CRs) and Dynamic Spectrum Access (DSA) emerging as critical technologies. New AI techniques, particularly Deep Learning (DL), provide the tools to manage these networks intelligently. ​ 

Successful deployment of radio access ADs requires significant innovation in AI/ML algorithms for resource management, addressing challenges in algorithm selection, deployment location, and lifecycle management. Tailored AI solutions must consider sustainability, reliability, scalability, resource awareness, training efficiency, communication overhead, and responsiveness. ​ ​

This dissertation explores these challenges from two perspectives: radio networks and ML. It proposes a novel spectrum-sharing framework for radio networks, using a two-tier architecture for efficient spectrum sharing and protection of incumbent transmissions. A general framework for developing Traffic Classification (TC) algorithms optimized for wireless networks is also presented, along with a DL-based Technology Recognition (TR) algorithm. ​ ​

On the ML side, the dissertation addresses training efficiency and scalability in DL and Reinforcement Learning (RL) algorithms for networking problems. It introduces a Semi-supervised Learning (SSL) based DL algorithm for label-efficient TR and a novel approach to minimize communication overhead in distributed Parallel Reinforcement Learning (PRL) algorithms. These contributions aim to advance the deployment and management of autonomous and intelligent network systems.

Quinten Beirinckx

• 10/07/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, Promotiezaal Q0.02
• ​Online PhD defence
• Supervisors: Jan Sijbers, Arnold J. den Dekker & Marleen Verhoye
• Department of Physics

Abstract

Magnetic resonance imaging (MRI) is one of the most widely used neuroimaging techniques, thanks to its exceptional soft tissue contrast and intrinsic safety for patients. Unfortunately, signal intensities in conventional MRI images are expressed in relative units that depend on scanner hardware and software. While this poses no issue for visual inspection of anatomy, it complicates the quantitative comparison of signal intensities within a scan, between successive scans, and among different subjects. In contrast, quantitative MRI (qMRI) generates quantitative maps of absolute biophysical tissue parameters by voxel-wise fitting a biophysical signal model to a series of contrast-weighted scans. Increasing evidence suggests that qMRI can detect and quantify subtle microscopic tissue damage, aiding in the early detection and accurate diagnosis of various neurodegenerative diseases such as Alzheimer's or multiple sclerosis. The impact of these diseases is growing in our rapidly aging society. However, current qMRI techniques require long scan times due to the need for multiple contrast-weighted scans with high-resolution and high signal-to-noise ratio (SNR). Long scan times increase the risk of motion artifacts, reduce patient throughput, and decrease patient comfort, hindering clinical adoption of qMRI. ​ ​

This work therefore investigates model-based super-resolution reconstruction (SRR) for qMRI, to optimize the trade-off between spatial resolution, SNR, and scan time. SRR enables the reconstruction of a high-resolution scan or parameter map from a series of low-resolution MRI scans. As an additional innovation in this thesis, patient movement is jointly estimated, which is crucial to avoid motion artifacts in the reconstruction. The new SRR framework is extensively tested and compared with existing SRR methods, both in computer simulations and using real brain MRI data. Furthermore, the new framework uses Bayesian statistics to incorporate prior knowledge about noise and brain tissue characteristics into the reconstruction process. This allows for the reconstruction of a 3D high-resolution quantitative tissue parameter map from motion-affected low-resolution MRI scans. Application to real brain MRI shows that the proposed framework significantly improves the accuracy of tissue parameter quantification compared to methods using a separate motion correction step. Finally, the SRR framework is extended to perfusion MRI for improved quantification of cerebral blood flow (CBF), in combination with Arterial Spin Labeling (ASL). This enables direct reconstruction of a high-resolution quantitative CBF map from low-resolution ASL MRI scans, significantly improving SNR and CBF accuracy over traditional methods. ​ ​

Bent Smets

• 04/07/2024
  
• 5 p.m.
• Venue: Stadscampus, Promotiezaal Klooster van de Grauwzusters
• Supervisors: Filip Meysman & Karolien De Wael
• Department of Biology

Abstract

 Long before humans, microbes harnessed the power of electricity by evolving specialized cell structures and electrogenic metabolisms. In the last two decades, several types of electroactive bacteria have been discovered, and so-called “cable bacteria” present a recent and remarkable addition to this list. These long, filamentous bacteria are members of the Desulfobulbaceae family and thrive in marine and freshwater sediments worldwide. Cable bacteria transport electrical currents across centimetre-scale distances to couple the oxidation of free sulphide deep in the sediment column to oxygen reduction within the upper sediment layer. In doing so, cable bacteria substantially impact the local geochemistry to their benefit.

An important outstanding question is how electrical currents are conducted through the cable bacterium filaments. Previously, it has been shown that electrical currents run through a network of parallel, highly conductive protein fibres in the cell envelope. These fibres are connected to one another in the cell junctions by an intriguing cartwheel structure, which is also conductive. This way, the cartwheel adds redundancy to the electric network, making it failsafe. Together, the fibres and cartwheels form the most elaborate electron transport network known in biology. The central objective of this PhD project is to elucidate the molecular structure of this electron transport network.

Chuang Gao

• 03/07/2024
  
• 2 p.m.
• Venue: Campus Groenenborger, U.024
• ​Online PhD defence​
• Supervisor: Timothy Pennycook
• Department of Physics

Abstract

Transmission electron microscopy is an important technique in the exploration of materials’ structures. This is especially true since the development of electron optical aberration correctors greatly facilitated atomic resolution imaging. We are currently experiencing an ongoing revolution in electron microscopy with the widespread adoption of direct electron detectors. Scientists have reported a lot of key scientific findings facilitated by direct electron detectors. One particular research domain is electron ptychography, which holds promise for unraveling the intricate structures of highly beam-sensitive materials like bio samples and achieving super-resolution without the limitation of aperture in the condenser lens system. Nevertheless, challenges persist both in experimental setups and algorithmic processes. Issues such as the comparatively sluggish scanning speed of cameras and contrast reversals of the reconstructed phase for relatively thick specimens, disrupting phase or weak phase approximations, remain noteworthy limitations. This thesis addresses these challenges by the event-driven Timepix3 detector, presenting a viable solution to the speed bottleneck. Moreover, innovative approaches for applying electron ptychography to relatively thick samples, employing a middle focusing strategy, are proposed. This research aims to push the boundaries of electron microscopy, offering solutions to existing limitations and advancing the field towards more efficient and accurate imaging techniques.

Rajeshreddy Ninakanti

• 27/06/2024
  
• 4 p.m.
• Venue: Campus Middelheim, A.143
• Supervisors: Sammy Verbruggen & Sara Bals
• Department of Bioscience Engineering

Abstract

The global challenge posed by increasing levels of greenhouse gases and the associated detrimental impacts of global warming necessitate a strategic shift from traditional fossil fuel-based energy systems to more sustainable, renewable, and circular energy and material solutions. Consequently, the potential of photoactive nanoparticles, particularly those that harness light-driven processes, has captured extensive scientific interest as a viable approach to mitigating energy and environmental challenges on a global scale. Although, the adoption of solar light based solutions in the chemical industry has been very less due to sluggish reaction rates and its cascading effects on its economics.

The primary focus of this dissertation is the study of plasmonic metal nanoparticles and metal oxide nanoparticles, emphasizing their applications in light-driven energy conversion. The distinctive properties of plasmonic materials, especially surface plasmon resonance (SPR), are pivotal in these applications. SPR involves the oscillation of electron clouds at the surface of nanoparticles when resonating with incident electromagnetic radiation, significantly enhancing solar radiation absorption. This feature is crucial for addressing the limitations of semiconductor photocatalysts like TiO2, which typically exhibit restricted absorption of solar irradiation. The objective of this dissertation is to further optimize the plasmonic enhancement mechanisms by strategically tuning the interactions between plasmonic nanoparticles and TiO2. This is achieved through the development of core-shell nanostructures and the self-assembly of supraparticles, designed to enhance plasmonic photocatalytic systems.

The dissertation begins by elucidating the basic concepts and ideations behind the construction of these nanostructures and their roles in enhancing plasmonic photocatalysis, focusing on mechanisms such as near-electric field enhancement, electron transfer, and enhanced photon absorption. To achieve these objectives, modified synthesis techniques were developed to fabricate novel Au@TiO2 core-shell structures with precisely controlled TiO2 shell thickness and self-assembled Au-TiO2 supraparticles with variable sizes. The thesis further delves into the structural characterization of these synthesized nanoparticles, introducing both basic and advanced electron microscopy techniques. For the specific applications of these structures, it was found that Au@TiO2 core-shell nanoparticles with an optimal 4nm TiO2 shell thickness show significant enhancement in the hydrogen evolution reaction. Additionally, the largest Au-TiO2 supraparticles demonstrate superior efficacy in hydrogen peroxide generation.

This work not only deepens the scientific understanding of plasmonic materials but also contributes to the development of renewable energy materials.

Adnan Sabovic

• 27/06/2024
  
• 4 p.m.
• Venue: Campus Middelheim, G.010
• Supervisors: Jeroen Famaey & Eli De Poorter
• Department of Computer Science

Abstract

The Internet of Things is a concept used to connect embedded objects to the Internet, enabling billions of tiny devices to cooperate and communicate with each other while performing different application tasks. Currently, most of these devices rely on batteries as the main power source. These batteries can provide a stable power supply, but even when rechargeable, they are short-lived, lasting at most a few years. They contain toxic chemical components that can be harmful to the environment. Tiny battery-less IoT devices that depend on harvested energy from their environment provide a promising alternative for a more sustainable IoT vision. Using this approach, battery-less devices harvest available energy from different environmental sources (e.g., solar, kinetic, Radio Frequency (RF) energy) and store it in small capacitors. These capacitors act as energy storage, and compared to batteries, they are smaller and lost more than a decade, thanks to their ability to handle a large number of charge cycles. However, they have significantly lower energy density, storing less energy in a given volume, which can be critical for applications that require long-term energy supply and storage. Also, battery-less IoT devices operate in an unpredictable and dynamic energy harvesting environment, which together with the small energy storage, results in intermittent on-off behavior of the device. To effectively use battery-less IoT devices, the applications running on them should be able to properly handle their intermittency. Measuring and understanding the current consumption and execution time of different tasks of the application is one of the most important steps in achieving this. Conventional computing models and static sequential applications cannot handle such behavior, as they cannot guarantee forward progress, relying only on volatile memory, and assuming a stable energy supply during execution. This problem can be solved with task-based models, where each task performs some atomic function, and its output is saved in non-volatile memory after it successfully completes, including energy awareness as a most important factor in helping to avoid power failures and maintain forward progress.

Advancements in processor efficiency, sensors, as well as low-power communication and computing, unlock the possibility for direct deployment for a wide range of traditional and intelligent applications on battery-less devices, enabling them to play an important role as an extreme edge device in an IoT ecosystem.

Jelle Dillen

• 24/06/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, O.01
• Supervisors: Sarah Lebeer & Peter Bron
• Department of Bioscience Engineering

Abstract

Urogenital infections are major contributors to antibiotic prescriptions, but treatment efficacy is compromised by antibiotic resistance. Additionally, antibiotics disrupt beneficial microbiota, increasing the risk of relapses and side effects. This highlights the urgent need for alternatives, such as microbiome-altering therapies. A vaginal microbiome dominated by lactobacilli, such as L. crispatus and L. jensenii, is associated with vaginal and urogenital health and is considered an "optimal" microbiome. A depletion of Lactobacillus species in the vagina has been linked to higher risks of bacterial, fungal, and viral infections in the vagina and other urogenital sites. While lactobacilli are known for their lactic acid production, their other antimicrobial compounds, including bacteriocins, nonribosomal peptides, and polyketides, remain understudied.

To enhance our understanding of their role, we utilized extensive data and samples from the Isala citizen science study, isolating over 900 Lactobacillaceae strains. Genomic analysis of 242 selected isolates revealed 514 biosynthetic gene clusters, mostly novel, including ribosomally-synthesized and post-translationally modified peptides, nonribosomal peptides, and polyketides. This genetic diversity highlights the untapped potential of Lactobacillaceae for specialized metabolite discovery. Further investigation into the diversity of L. crispatus isolates, coupled with their antimicrobial properties against common vaginal pathogens, revealed clear links between phylogeny, biosynthetic potential, activity, and vaginal community composition. One strain exhibited particularly high activity against tested pathogens, which correlated with the presence of a novel lanthipeptide cluster. This cluster's antifungal activity was validated through in silico characterization and heterologous expression in Escherichia coli. Additionally, we aimed to characterize a novel antimicrobial from L. jensenii, conserved across the species, which showed activity against the Gram-negative pathogen Klebsiella pneumoniae, a cause of adverse pregnancy outcomes and vaginal infections.

In summary, this PhD thesis provides an overview of specialized metabolite cluster prevalence and abundance across vaginal Lactobacillaceae, enhancing our understanding of their protective and beneficial roles, particularly focusing on L. crispatus and L. jensenii. By characterizing an antifungal from a specific L. crispatus isolate and a species-conserved antimicrobial from L. jensenii, we have laid the groundwork for developing novel probiotics, live biotherapeutic products, or specialized metabolites.

Simon Reynaert

• 19/06/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.07
• ​Online PhD defence
• Supervisors: Ivan Nijs, Hans De Boeck, Erik Verbruggen & Tommy D'Hose
• Department of Biology

Abstract

Global warming is changing the intra-annual variability of weather patterns around the world. Recent observations indicate that summer weather persistence is increasing in Europe, resulting in both longer dry and wet spells compared to historic averages. However, the consequences of these newly emerging climate regimes for terrestrial ecosystems have barely been studied. This thesis investigates how increasing weather persistence influences mesic grassland ecosystem functioning in a series of manipulation experiments with model grassland systems under different levels of management intensity. In addition, we explore several solutions for the observed consequences in terms of plant and soil remediation.

In unfertilized, species-rich systems, we found evidence that precipitation regimes with longer dry and wet spells reduce aboveground productivity and native plant diversity in the short-term (one to two years) (Chapters II, III & IV). These changes related to differences in species-specific responses to drought and more pronounced fluctuations in nutrient availability over time as well as shifting species interactions under drought. In particular, the length of the longest drought period throughout the growing season was the strongest determinant of short-term diversity and productivity declines. Moreover, the disproportionate loss of legumes and other forbs compared to graminoids likely negatively affected forage quality. Despite these adverse effects, ecosystems also showed acclimation to more extreme regimes through community reorganization and altered biochemistry, buffering productivity under recurrent and contrasting extreme soil conditions. ​ ​

In fertilized monoculture systems, increased weather persistence reduced yield yet slightly improved forage digestibility (Chapters V & VI). Moreover, soils from permanent grasslands with elevated organic carbon (OC) showed declines in soil water availability. However, not all species and soil types were affected equally. Growing Festulolium, Festuca arundinacea or Dactylis glomerata led to significantly smaller yield losses compared to Lolium perenne cultivars. Regarding soil remediation, high OC soils with a permanent grassland history limited negative impacts on plant quality and yield in sandy-loam soils. Moreover, mixing basalt powder with the top soil layer improved yield in unfertilized monoculture systems under prolonged dry and wet spells. ​ 

To conclude, these findings indicate that increasing weather persistence will likely negatively affect grassland ecosystem service provisioning (i.e., plant diversity, productivity and soil water availability) by altering soil water and nutrient dynamics. However, choosing more drought resistant species and management practices that improve soil carbon sequestration shows potential to limit adverse effects on plant performance, ultimately making natural and managed grassland systems more climate resilient. ​ ​

Mats Van Gestel

• 18/06/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, O.05
• Supervisors: Erik Matthysen (UA), Dieter Heylen (UA) & Kris Verheyen (UGent)
• Department of Biology

Abstract

The sheep tick (Ixodes ricinus) vectors, among others, Borrelia burgdorferi s.l., the causative agent of Lyme disease. Differences in its density and pathogen prevalence between forests have been attributed to climatological values, vegetation characteristics and the abundance of hosts. However, variation within forests has only been studied sporadically, whereas its understanding would greatly benefit forest management. This knowledge gap is addressed through a multi-year sampling campaign in recreational forests; Tick density, vegetation characteristics and local host abundance were measured and captured ticks were tested for pathogens. This was conducted near recreational infrastructure (benches and trails), as sharp changes in vegetation characteristics, microhabitat and the habitat use of tick hosts are expected here. This in turn was expected to affect tick density and pathogen prevalence. Additionally, the distinction was made between trails in the forest core and those at forest edges. In addition to the highest tick density being recorded in the forest interior, with lower values near infrastructure, an edge effect was observed with lower values at trails in forest edges. These differences within forest stands are attributed to differing vegetation characteristics and the habitat use of hosts. Density was higher in stands dominated by deciduous vegetation with dense canopy, compared to open, coniferous vegetation. The same was true for stands with a more developed shrub layer. Small, consistent differences in pathogen prevalence were also observed and attributed to varying drivers for different pathogens. These conclusions support the occurrence of amplification and inhibition mechanisms on a within-forest stand scale. I herein highlight that the effect of tick borne disease risk drivers may differ based on the considered spatial scale.

Additionally, experiments on the effect of herb layer vegetation on nymphal survival in summer and larval detachment patterns from wood mice were conducted, to estimate whether these processes may affect tick distribution. The experiment on survival implied that canopy and shrub layer vegetation determine microclimate buffering, and therefore also tick survival, to a larger degree than herb layer vegetation. The experiment on larval detachment offers a first indication that environmental factors can alter the detachment probability of larvae from wood mice, independent from temporal effects. The thesis is concluded by the translation of the fundamental findings to a forest management context. The contact rate between ticks and humans is outlined as a persistent knowledge gap with regards to the quantification of tick-borne disease risk. ​ ​

Nithin Bharadwaj Kummamuru

• 10/06/2024
  
• 3.30 p.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisors: Patrice Perreault & Sammy Verbruggen
• Department of Bioscience engineering

Abstract

In a world increasingly reliant on alternative energy sources, the quest for efficient and secure storage solutions is paramount. This thesis explores the exciting potential of a familiar material – water - to act as a vault for next-generation energy sources like hydrogen (H2) and methane (CH4). Nature offers a solution in the form of clathrate hydrates, fascinating cage-like structures formed from water molecules that can trap these gas molecules within their framework. This research investigates on improving the formation kinetics and gas storage capabilities of clathrate hydrates utilizing porous materials and the interstitial space between non-porous materials to augment the contact between gas and water thereby catalysing the growth of hydrates and unlocking their full potential as efficient and secure energy storage reservoirs. A key outcome of this research is the formulation of an empirical correlation, offering predictive insights into CH4 hydrate phase equilibrium conditions. Innovative approaches utilizing thermally conductive beads have yielded substantial enhancements in CH4 uptake. Furthermore, the identification of optimal water content within porous materials showcases a pathway to maximize CH4 storage capacity and hydrate growth kinetics. In the domain of hydrogen storage, attention is also directed towards unstirred systems, where the integration of functionalized porous materials has demonstrated a significant improvement in the rate of hydrate formation and the overall H2 storage capacity. A noteworthy achievement of this research lies in the successful storage of H2 within confined CH4 hydrates through a gas exchange process and the preliminary results show the potential for a safer and more sustainable method for H2 storage at mild thermodynamic conditions, offering promising prospects for future energy systems.

Robin Lasters

• 06/06/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.01
• ​Online PhD defence
• Supervisors: Lieven Bervoets & Marcel Eens
• Department of Biology

Abstract

In the past decade, homegrown food consumption has surged in rural, urban, and industrial areas. However, organic pollutants in private gardens, including per- and polyfluoroalkylated substances (PFAS), pose health risks by entering the food chain through bioaccumulation. Very little is known about the driving factors of PFAS accumulation in homegrown food. Therefore, this thesis project aimed to assess PFAS accumulation in various homegrown food categories and related human exposure risks, exploring factors affecting PFAS bioavailability. The results showed that multiple PFAS are omnipresent in homegrown food and can accumulate to concentrations that frequently exceed available health guidelines, even under modest consumption scenarios, especially with regard to egg intake. Within the crop category, higher accumulation was noticed in annual crops in comparison to perennial crops, potentially linked with differences in terms of life-history strategies between these two plant taxa. Large spatial and temporal differences in soil PFAS profile and concentrations were found within private gardens, suggesting that site-specific characteristics and functional usage play a major role in shaping local PFAS contamination. Predictive models could be constructed for some major PFAS in eggs, which show promising potential for applicability in risk assessment by policy makers. Moreover, mitigation and remediation measures could be formulated that should be readily usable for private gardeners to ultimately lower PFAS exposure via homegrown food.

PFAS pollution in gardens within ± 4 km from the fluorochemical plant in Antwerp could be strongly linked with both historical and recent fluorochemical emissions. On the other hand, diffusive mechanisms (e.g. atmospheric transport) and site-specific soil management may be mainly affecting levels at gardens further away from point sources. The accumulation in chicken eggs was generally higher closer to the major fluorochemical plant, although soil characteristics (e.g. organic matter, clay content and pH) could strongly affect this pattern. Conversely, the PFAS accumulation in the crops was not affected by the distance from the plant site and soil characteristics played only a minor role in governing crop accumulation. Long-term declining concentrations in soil and eggs could be observed for some PFAS, although this trend stagnated over recent years. Short-term increases of short-chain and long-chain PFAS concentrations could be observed, mainly in the soil from the chicken enclosure. These findings underpin that homegrown food cannot be neglected as a relevant human exposure source to PFAS and show the urgent necessity for further regulation steps and monitoring efforts.

Sanah Majid

• 03/06/2024
  
• 4.30 p.m.
• Venue: Campus Groenenborger, T129
• Supervisors: Ronny Blust & Karen Smeets
• Department of Biology

Abstract

The presence of metals in the environment has often been associated with various adverse effects to living systems, including humans. Given their ubiquitous nature, metals are commonly accepted to occur as complex mixtures and have therefore always been of concern. Despite the considerable attention given to this issue, there remains a lack of sufficient knowledge of the mechanisms underlying toxicological outcomes specific to each metal and/or each mixture. In addition, knowledge of the toxicity of metal mixtures in taxonomically distinct species is inadequately explored. Addressing these key questions is important for gaining a better understanding of the magnitude and mechanisms of toxicity in biological systems. In this context, our study aims to explore how exposures in a single and binary metal pollution scenario differentially affect aquatic vertebrate and invertebrate organisms across various biological levels. The study focuses on copper (Cu) and cadmium (Cd) as metal toxicants in three animal models: the zebrafish (Danio rerio), the water flea (Daphnia magna) and the planarian flatworm (Schmidtea mediterranea). Our study contributes to a better understanding of the interactive effects of metal toxicity across different biological levels and emphasises the importance of an in-depth approach in assessing the risks associated with metal pollution. The findings presented in this work show a stronger action of Cu and Cd as a mixture in different aquatic organisms under controlled conditions. In particular, our study reveals important developmental, behavioural, and molecular changes with some effects more manifested in mixtures compared to single metal exposures. Furthermore, our study highlights that relying solely on metal accumulation levels as a reliable predictor of toxicity is insufficient; however, it can still serve as a valuable biomarker of exposure. The absence of a clear relationship between metal concentrations in tissues and observed effects emphasises the influence of internal compartmentalisation and the intricate molecular defence mechanisms involved in damage control and repair processes.

Romain Garrigues

• 27/05/2024
  
• 10 a.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisors: Matteo Campioli, Han Asard & Hamada Abd-Elgawad
• Department of Biology

Abstract

This thesis investigates the timing of phenological stages, the influence of environmental factors (temperature and daylength), and the molecular mechanisms underlying winter dormancy (endodormancy) in deciduous trees. The focus is on two deciduous species: Populus nigra (black poplars) and Fagus sylvatica (European beech). Three field experiments examine responses to current climate conditions, late fall warming, and modified winter-spring daylength. Transcriptome analysis sheds light on gene expression changes during endodormancy. While early successional species exhibit earlier phenological events, temperature variations play a significant role. Lack of chilling accumulation during late autumn affects dormancy depth and spring phenology, with elevated temperature treatments showing varied responses across years. Late autumn temperatures and daylength interaction affect bud dormancy and spring phenology in European beech and are showing that daylength is the last guardian to avoid early budburst. Transcriptome analysis identifies genes associated with dormancy transition, emphasizing DNA binding, cell cycle regulation, and phytohormone pathways. Overall, the thesis contributes to understanding deciduous tree behavior and implications for forest responses to climate change, offering insights into phenological events (dormancy depth and spring budburst) while presenting novel data on molecular mechanisms during endodormancy.

Duygu Gizem Sentürk

• 03/05/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, US.024
• ​Online PhD defence​
• Supervisors: Sandra Van Aert & Annick De Backer
• Department of Physics

Abstract

Nanomaterials play an essential role in modern technology from industry to life sciences. At the atomic scale, the properties and functionalities of materials strongly depend on their size and shape. Moreover, the functionalities of nanostructures are further enhanced when they are composed of multiple types of elements. Therefore, reliable characterisation of the atomic structure is essential to understand the structure-property relationship and to design nanostructures with tailored functionalities.

Scanning transmission electron microscopy (STEM) is one of the most powerful tools for characterising nanostructures. In STEM, electrons scattered from the sample are collected by an annular detector with predefined inner and outer angles, producing 2D atomic resolution image of a 3D nanostructure. Quantitative analysis of STEM images reveals the structural parameters such as the position, number and type of atoms. Until now, advanced statistics- and simulations-based methods have made it possible to count the number of atoms along the viewing direction for each atomic column in monotype crystalline nanostructures from a single STEM image. This allows the size and shape of nanomaterials to be estimated. However, for heterogeneous multi-element nanostructures, the atomic columns with varying thicknesses and elemental compositions often lead to indistinguishable measurements in the presence of noise. This complexity makes it challenging to determine the number of each type of atom. To overcome this problem, traditional quantification methods often rely on prior knowledge of the shape or size of the material, which is not always available.

To avoid the need for prior knowledge, we propose to use a combination of multiple STEM images. These images are obtained from non-overlapping annular detector collection regions revealing unique information about the thickness and composition of the atomic columns. For this purpose, the 4D STEM technique, providing a rich electron diffraction dataset, is beneficial as it allows the generation of multiple STEM images with arbitrary annular detector settings. ​ ​

In this research, we explore the extension of both statistics- and simulations-based methodologies for such a multimode STEM imaging approach and achieve atom counting of each type of element in heterogeneous nanostructures. Moreover, strategies that would reveal these unknown atom counts with the highest attainable accuracy and precision are investigated by deriving optimal statistical experimental settings. This thesis demonstrates the possibilities and limitations of using additional signals resulting from multiple STEM images and ultimately from a 4D dataset when characterising multi-element nanostructures, offering a dose-efficient alternative.

Laurens Van Hoecke

• 02/05/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, O.05
• ​Online PhD defence
• Supervisors: Patrice Perreault & Sammy Verbruggen
• Department of Bioscience Engineering

Abstract

Hydrogen (H₂) is expected to become a key molecule in the transition towards a society running on renewable energy. It can be used to store excess renewable energy at peak production moments and release this energy at a later stage when renewable energy production is less. However, storing H₂ is challenging due to the low density of this gas. As a solution, Liquid Organic Hydrogen Carriers or LOHC molecules have been proposed in the passed to increase volumetric energy density of H₂. LOHC are a class of molecules that have storage sites available, to which the H₂ gas can be chemically bounded. The LOHC molecule under investigation was dibenzyltoluene (DBT), which is an oil like liquid, that is easy to transport and poses little fire or explosion risks. To release the H₂ from the DBT carrier, via a so-called dehydrogenation reaction, efficient mass and heat transfer is required during the process, since a large volume increase is expected from H₂ release and the reaction is endothermic, i.e., a self – cooling process that takes place at temperatures around 300 °C. The heat has to be supplied specifically to the active sites of catalyst particles that are present inside the reactor and which enable the dehydrogenation to proceed. For heat transfer limited processes fluidized bed reactors are often used, which is a type of reactor where the particle phase is being agitated by the fluid flow. The research proposed in this work, was to explore via computational fluid dynamics (CFD) simulations the possibilities and challenges of using fluidized bed reactors for the dehydrogenation of LOHC. The model selection required for CFD simulations of a three-phase system was investigated in this work, with a main emphasis on the drag model selection. The CFD modelling study was focused on the use of swirling fluidized bed reactors, since it was hypothesised that the swirling effect could also aid in increased removal of the gas phase from the reaction medium to increase the efficiency of the process. Ultimately, it was shown that the main challenges in the design of fluidized bed reactors will be to create uniform particle distribution inside the reactor. A new design for a dehydrogenation reactor is proposed based on the insights gained in this thesis.

Hamid Ahmadi Eshtehardi

• 25/04/2024
  
• 10 a.m.
• Venue: Campus Drie Eiken, R.R4
• ​Online PhD defence
• Supervisors: Annemie Bogaerts & Marie-Paule Delplancke
• Department of Chemistry

Abstract

Despite the recent increasing interest in plasma technology for nitrogen fixation purposes, industrialization of this technology faces several challenges, including challenges of plasma catalysis for selective production of chemicals, the high energy cost of plasma-based nitrogen fixation compared to current industrial processes, and the design and development of scaled-up and energy-efficient plasma reactors for industrial purposes. In the framework of this thesis, we have tried to tackle these challenges and add to the state-of-the-art in plasma-based nitrogen fixation using a combination of experimental and modelling work.

Mees van Kampen

• 16/04/2024
  
• 1.30 p.m.
• Venue: Stadscampus, S.M.003
• ​Online PhD defence​
• Supervisors: Francesco Hautmann & Pierre Van Mechelen
• Department of Physics

Abstract

Studies on calculations for high-energy proton-proton collisions performed by the Parton Branching (PB) method for the evolution of transverse momentum dependent parton distribution functions (TMDs) will be presented. The PB method allows to perform both inclusive and exclusive calculations of collision final states by means of Monte Carlo techniques. Evolution of TMDs in the PB method allows for the resummation of soft gluons by the Sudakov form factor. The implementation of PB TMDs in the CASCADE Monte Carlo event generator allows for the calculation of a wide variety of particle collision processes in a wide kinematic range. ​ ​

I examine the PB method, focusing on the Sudakov form factor and the soft-gluon resolution scale. By extending the emission phase space with longitudinal splitting fractions z approaching one, we have achieved accurate perturbative resummation and a non-perturbative contribution to the evolution. A dynamical resolution scale separates resolvable and non-resolvable phase space regions, acting as a boundary between these perturbative and non-perturbative domains. We show that PB evolution with next-to-leading order (NLO) splitting functions achieves next-to-leading logarithmic accuracy in soft-gluon resummation. The implementation of the physical soft-gluon coupling enhances Sudakov resummation towards next-to-next-to-leading logarithmic accuracy. Non-perturbative contributions of the Sudakov are illustrated through the extraction of the Collins-Soper kernel. These extractions highlight the influence of both the emission phase space and the scale of the strong coupling in TMD evolution on the large-b behavior of the CS kernel. ​ ​

Combining higher order matrix element calculations with PB TMDs and TMD shower is done through matching and merging techniques. Azimuthal correlations of high transverse momentum jets in di-jet production and boson-jet production are calculated using PB TMDs matched to NLO matrix elements. QCD predictions for final states with multiple jets in hadron collisions make use of multi-jet merging methods. These methods consistently combine the contributions from hard scattering matrix elements with different parton multiplicities and parton showers. ​ ​

Calculations of jet transverse momentum and jet multiplicity distributions, as well as highly non-trivial jet event shapes, are performed with the recently developed TMD merging method. We investigate theoretical predictions for Z-boson plus jets production using multi-jet merging algorithms. Our analysis focuses on the differential jet rates and their discontinuities, which allows us to develop a method for quantitatively analyzing the merging algorithm and its dependence on the merging scale by varying invariant di-lepton masses.

Noelia Felipe Montiel

• 29/03/2024
  
•  p.m.
• Venue: Stadscampus, Hof Van Liere, F. de Tassiszaal
• ​Online PhD defence
• Supervisor: Karolien De Wael
• Department of Bioscience Engineering

Abstract

Illicit drug trafficking poses a significant threat to public health, safety and environment, requiring not only effective law enforcement measures, but also innovative harm reduction approaches. This PhD thesis aims to contribute to this multifaceted approach by developing electrochemical fingerprint (EF)-based sensors for the most prevalent illicit drugs and precursors. Throughout this thesis, the EF concept will be put to the test to ensure its practical applicability in real-world scenarios.

The lack of electroactivity in certain compounds may impose constraints on EF-based detection tools. This pivotal challenge is addressed by introducing a derivatization reaction to convert non-electrochemical active substances into electroactive ones. Specifically, BMK was used as a model due to its inherent inactive nature. This approach expands the scope of applicability of the EF also to non-electroactive species.

Recognizing that precursors are often exceptionally pure, the focus was then shifted to an opposite context, i.e. to heroin which is generally found together with cutting agents, impurities, and by-products from its manufacturing process. Oxidation signals of heroin were elucidated and further exploited to allow its detection in seizures. In this context, several analytical strategies were evaluated to enrich the EF of heroin. The development of these sensors makes clear the practical utility of the EF for highly impure substances in on-site settings.

Up to this point, the study of the EF focused exclusively on oxidative processes. However, reductive processes may also provide valuable insights on the target analytes. Therefore, the complete EF including oxidative and reductive processes was investigated to extract a richer and more detailed dataset, maximizing the information available. To achieve this, counterfeit Xanax emerged as a right candidate. This choice is based in the complex matrix of Xanax, which typically contains a diverse range of substances susceptible to both processes. By interrogating the complete EF, the research strives to disclose the full composition of complex matrixes.

When developing EF-based solutions for detecting illicit drugs, sensitivity arises as a critical point, especially in scenarios involving illicit drug smuggling. Consequently, the exploration of surfactants and nanomaterials on plant-based and synthetic cannabinoids becomes imperative to boost the analytical parameters.

Overall, by presenting electrochemical methods and practical solutions, this research paves the way for a more effective on-site EF-based devices to combat drug trafficking and associated challenges, positioning the electrochemical technology as a complementary tool to traditional devices.

Tom Eilers

• 29/03/2024
  
• 3 p.m.
• Venue: Campus Drie Eiken, O.006
• Supervisor: Sarah Lebeer
• Department of Bioscience Engineering

Abstract

Fermentation of foods and beverages has been an important biological preservation method for humanity for millennia. In this PhD project, we focused on lactic acid fermentation, and more specifically vegetable fermentations. In vegetable fermentations, a salt-brine is generally added to reduce the growth of spoilage organisms. In most salt-brine vegetable fermentations studied to date, lactic acid bacteria (LAB), particularly members of Lactobacillaceae are the most important fermentation bacteria.

This PhD explored three major aspects of LAB-dominated vegetable fermentations: (1) The effect of external factors studied, including vegetable type, salt concentration, and addition of CO₂, on the microbial community dynamics. (2) the genomic diversity of the Lactobacillaceae as determined by pangenome analysis. (3) The role of cellulase and carotenoid production by Lactobacillaceae as adaptation factors in vegetable fermentations and other environments.

Firstly, we found that most fermentations were characterized by a LAB-dominated community with the genus Leuconostoc as the most abundant genus. Root-associated vegetables, such as beet, carrot, parsnip, and sunroot had the most stable and robust fermentations. Additionally, lower salt concentrations was linked to the presence of less favorable genera of the Enterobacterales order. CO₂-addition could mitigate the effect of the low salt concentration, as most Enterobacterales disappeared faster.

Secondly, the genomic diversity of the Lactobacillaceae was explored and various unique orthogroups for each species could be detected within their pangenome. These unique orthogroups were used to develop selective and specific strain-specific primers and were validated in situ by tracking starter cultures during carrot juice fermentations. Additionally, this exploration led to the discovery of a novel species called Lactiplantibacillus carotarum, isolated from fermented vegetables. Phenotypically, it was able to metabolize more different carbohydrates compared to the closest relatives, which can be highly present in some fermented vegetable substrates.

Lastly, the functional potential of cellulase and carotenoids by Lactobacillaceae isolated from fermented vegetables was explored. These factors were hypothesized to be important for adaptation of the fermented food microorganisms to the vegetable fermentation environment and were verified in vitro.

In conclusion, this PhD work contributed to a new understanding of the role of plant substrates, salt, and CO₂ in the microbial ecology of vegetable fermentations. We also revealed the potential of the pangenome for vegetable fermentation research. This work incrementally advances our understanding of these man-made ecosystems and might facilitate the development of novel functional foods in a more targeted approach.

Tim Leys

• 28/03/2024
  
• 4 p.m.
• Venue: Campus Middelheim, G.010
• Supervisor: Guillermo A. Pérez
• Department of Computer Science

Abstract

In the present day, complex and dynamic systems have become ubiquitous. Their development remains intrinsically complex and, in safety-critical environments, verification of these systems is necessary to guarantee their safety and reliability. With the rise in popularity of generative software such as ChatGPT and GithubCopilot, there is a need for automatic fast verification of software.

Although counter automata are, generally speaking, a Turing complete model of computation, subclasses exist for which reachability (and its complement, safety) is decidable. Two such classes are one-counter automata, that limit the amount of counters to 1, and vector addition systems with states (VASS, for short), which have no conditions on the counter values besides that they have to be positive. In both cases, tight complexity bounds for the reachability problem have been proven: Reachability for one-counter automata is NP-complete and Ackermann-complete for VASS.

To provide more efficient solutions, we present conservative approximations for both classes. We present continuous one-counter automata, which allow partial updates to be scaled down before applying them to the counter. Then, we give a P-time algorithm for verifying reachability in continuous one-counter automata, as well as an NP algorithm for verifying reachability in parametric one-counter automata. These algorithms are even more powerful in that they compute the full set of reachable values. We present a geometrical interpretation of reachability sets for continuous VASS—here, again, counter updates can be scaled down. We prove that the reachability problem for so-called flat VASS and linear path schemes is in P and that linear path schemes of polynomial length are sufficient to witness reachability. Additionally, we provide an implementable algorithm by encoding the reachability sets of linear path schemes as a solution of set (in)equalities.

Finally, we look at a practical case study of verifying properties on stochastic compartment models. We establish that certain stochastic compartment models can be encoded as probabilistic counter machines where the configurations are bounded. 

Based on the latter, we obtain simple descriptions of the Markov chains induced by such models in the PRISM language. This enables the analysis of such compartmental models via probabilistic model checkers. We report on experimental results where we analyze the efficiency of probabilistic model checkers on a Belgian COVID-19 model.

Evgenii Vlasov

• 27/03/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, S.207
• ​Online PhD defence
• Supervisors: Sara Bals & Johan Verbeeck
• Department of Physics

Abstract

Electron tomography (ET) is an indispensable tool for determining the three-dimensional (3D) structure of nanomaterials in (scanning) transmission electron microscopy ((S)TEM). ET enables 3D characterization of a variety of nanomaterials across different fields, including life sciences, chemistry, solid-state physics, and materials science down to atomic resolution. However, the acquisition of a conventional tilt series for ET is a time-consuming process and thus cannot capture fast transformations of materials in realistic conditions. Moreover, only a limited number of nanoparticles (NPs) can be investigated, hampering a general understanding of the average properties of the material. Therefore, alternative characterization techniques that allow for high-resolution characterization of the surface structure without the need to acquire a full tilt series in ET are required which would enable a more time-efficient investigation with better statistical value.

In the first part of this work, an alternative technique for the characterization of the morphology of NPs to improve the throughput and temporal resolution of ET is presented. The proposed technique exploits surface-sensitive secondary electron (SE) imaging in STEM employed using a modification of electron beam-induced current (EBIC) setup. The time- and dose efficiency of SEEBIC are tested in comparison with ET and superior spatial resolution is shown compared to conventional scanning electron microscopy. Finally, contrast artefacts arising in SEEBIC images are described, and their origin is discussed.

In the second part of my presentation, I will focus on real applications of the proposed technique and introduce a high-throughput methodology that combines images acquired by SEEBIC with quantitative image analysis to retrieve information about the helicity of gold nanorods. I will show that SEEBIC imaging overcomes the limitation of ET providing a general understanding of the connection between structure and chiroptical properties.

Irene Ramirez Rojas

• 22/03/2024
  
• 3 p.m.
• Venue: Campus Drie Eiken, O.04
• ​Online PhD defence​
• Supervisors: Erik Verbruggen & Heidy Schimann
• Department of Biology

Abstract

Tropical rainforests host an exceptional biodiversity and play a fundamental role in the regulation of global climatic cycles. Soil fungi and bacteria are key players in the transformation and processing of nutrients in terrestrial ecosystems while having an essential role as tree mutualists or antagonists. Still, there are gaps in our understanding of the main variables driving soil microbes on these forests and it is unclear how future climate change scenarios may impact soil microbes and further affect the ecosystem.

In this thesis, we first explored the drivers of the microbial community composition in two pristine forests in French Guiana by using amplicon DNA sequencing. The neighboring tree species were found to be a crucial factor influencing the fungal and bacterial community composition at our sites regardless of the season. Additionally, within the environmental factors explored, soil moisture, phosphorus (P) and nitrogen (N) availability were consistently the main soil properties controlling the composition of soil microbial communities.

Secondly, as increased nutrient deposition due to anthropogenic activities are expected to affect tropical forests ecosystems N and P availability, a factorial N and P nutrient addition experiment in the same sites was used to assess the effects of changes in the soil nutrient stoichiometry on the soil microbial communities. These results showed that after 3 years of nutrient additions, the bacterial and fungal community composition was affected by both the N and P additions. Besides, the fungal community composition had a stronger response to the nutrient addition, especially when P was added. Moreover, when the nutrient addition effect was assessed in bacteria and fungi with different life strategies, we found different nutrient optima between them.

Furthermore, to study the effect of the connection to an existing mycorrhizal mycelium on tree seedlings, I established a mycelium exclusion experiment. Interestingly, we could not detect an effect of the mycorrhizal mycelium exclusion on the seedling N uptake, performance, or fungal community composition in roots after one year.

All together this work provides a deeper understanding of the factors influencing the soil microbial communities on these lowland tropical forests, demonstrating that the tree community composition exerts a higher influence on the soil microbial community composition than previously expected. Moreover, our results show that the fungal and bacterial community composition and its relationship with trees in the vicinity is highly dependent on the ecosystem nutrient availability.

Lize Delanghe

• 05/03/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, O.01
• ​Online PhD defence
• Supervisors: Sarah Lebeer, Ingmar Claes, Julie Leysen & Margo Hagendorens
• Department of Bioscience Engineering

Abstract

The human skin microbiome is an open ecosystem that is influenced by external and personal factors and forms a key barrier against pathogens and other foreign substances. Different inflammatory skin conditions, such as atopic dermatitis (AD), are associated with a disturbed skin microbiome. The etiology of AD is complex, with roles for genetics, the immune system, environmental factors, and the skin microbiome, with a key role for Staphylococcus aureus as a skin pathogen. A high need for alternative treatments for AD has sparked the interest in probiotic interventions. This PhD thesis aimed to achieve in-depth insights into the relation between the skin microbiome and AD and to contribute to the development of a new probiotic product for AD. To develop microbiome-based treatments, we should first thoroughly understand the healthy skin microbiome. For this, we collected skin samples from the inner elbow of healthy volunteers and identified the core bacterial community. Age showed the be the major driver defining the skin microbiome composition and longitudinal stability. In addition, significant associations were found between specific skin taxa and season, hygiene, supplements, and the number of household member. Next, we compared the skin microbiome of our healthy cohort with AD patients. Here, we showed that S. aureus did not appear to have a prominent role in mild AD. We also assessed other taxa associated with mild AD and identified six skin species with an increased and 15 genera with a decreased abundance. To elucidate the potential of lactobacilli-based probiotics, we performed an exploratory study investigating the application of three lactobacilli strains on mild AD skin. The lactobacilli showed a good engraftment in the skin niche and the formulated creams reduced erythema and itch on the patients skin. These results already showed that probiotics have a high potential for AD, but there exists a large variety in strain-specific modes of action wherefore a careful selection of the most effective strains is important. Therefore, we selected 76 bacterial isolates and based on in vitro characterization of these strains, with the focus on safety, applicability and functional activity, we could finally select seven probiotic strains.

In conclusion, this work contributed to the microbiome research field and the development of new and effective treatments for AD by gaining insights into the skin microbiome in health and disease. In addition, we could select seven strains showing high promising modes of action, next to good safety and growth conditions.

Falk Mielke

• 01/03/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.04
  
• Supervisors: Peter Aerts & Chris Van Ginneken
• Department of Biology

Abstract

One crucial characteristic which sets animals apart from most other forms of life is the capability of quick locomotion.

Yet moving around does not always come easy: some animals are better at it than others, and common research tools in the field of locomotor biomechanics occasionally fail to quantify subtle differences.

For example, during early development, variability in locomotor patterns is high, faults and instabilities are common, but maturation is quick, which is a challenging context for comparative biomechanics.

In this project, I set out to expand our methodological capabilities for developmental, comparative biomechanics.

The thesis covers kinematics (how animals move) and kinetics (why they move). Based on the observation that locomotor patterns are often repetitive, and applying probabilistic, predictive models which incorporate variability, I spotlight the most fragile of newborn animals: those who are born with particularly low weight.

My primary model system are piglets, where low birth weight is common due to increasing litter sizes in commercial breeding.

The applied methods (Fourier analysis, probabilistic statistics) are exapted from other fields, thus not novel, but rarely applied to quantitative biomechanics to date.

I also highlight limitations and quantify commonly accepted inaccuracies of existing inverse dynamic methods. The case of piglet locomotion can demonstrate how the exapted tools enable unprecedented detail in the analysis of locomotor biomechanics.

In particular, I can confirm that low birth weight piglets are fully capable of normal locomotion, and I precisely quantify how their development is halted shortly after birth.

This provides some important constraints for the evaluation of coping strategies in commercial farming.

Besides, the methodological advances which I present in detail enable a whole new set of research questions for different contexts, within the field of locomotor biomechanics and beyond.

Jonas Schram

• 23/02/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.4
• ​Online PhD defence
• Supervisors: Karolien De Wael & Marc Parrilla Pons
• Department of Bioscience Engineering

Abstract

Today, illicit drugs are omnipresent in society. Clandestine markets are growing faster than ever before, record amounts of cocaine are seized in seaports and airports, while the associated violence is spiralling out of control. In addition, drug monitoring centres worldwide are warning for the increasing complexity of the drug markets, as the traditionally popular drugs are joined by countless new synthetic variants, while medical drugs are also increasingly being abused. In order to provide services confronted with illicit drug samples (police, customs, forensic scientists, first responders, …) with important information on the identity of an unknown sample, suitable analytical tests are required. While these exist for laboratory environments, on-site applicable tests are important to accelerate the decision-making process. Electrochemical sensors have all the advantages required for such on-site tests: they are fast, portable, easy-to-use and reliable. Furthermore, they are not influenced by colours, which are frequently added to drug samples to deceive the existing tests.

Previous work has mainly focussed on the detection of a single drug per analysis. However, many drugs could be encountered due to the diversity of the drug markets. Therefore, this project developed electrochemical strategies for the detection of multiple drugs simultaneously. First, the electrochemical behaviour of the individual drugs was studied in different measuring conditions (assessing the influence of pH, concentration and temperature). Then, all findings and strategies were combined to detect multiple targets simultaneously. An electrochemical sensor was developed for the four most popular drugs at music festivals: cocaine, MDMA, amphetamine and ketamine. This sensor generates a so-called ‘superfingerprint’ of the sample, which is then automatically interpreted by a developed algorithm in order to produce a straightforward output.

Finally, a pill analysis sensor was developed in the context of drug checking services, where a consumer can anonymously have a sample chemically analysed to obtain information on the composition, dose and potentially harmful additives. The sensor achieved an outstanding accuracy in identifying the main component and provided the option to quantify, as well as an indication on the presence of other substances in the sample. The project’s findings demonstrate the potential for electrochemistry in illicit drug detection and provide a basis for the development of new sensors, targeting other drug combinations.

Fardokht Rezayi

• 22/02/2024
  
• 4 p.m.
• Venue: Campus Drie Eiken, O.01
• ​Online PhD defence​
• Incoming Joint PhD Cardiff University - Universiteit Antwerpen
• Supervisors: Sabine Van Doorslaer & Damien Murphy
• Department of Chemistry

Abstract

Cu(II) coordination chemistry is of significant importance due to copper's widespread applications, particularly in chemical catalysis. This thesis explores the molecular structure, electronic properties, and variable coordination geometry of trigonal bipyramidal complexes of Cu(II) with tripodal ligands, more specifically different tripodal tetraamines. While square planar and square pyramidal Cu(II) complexes are commonly studied, less attention is given to trigonal bipyramidal Cu(II) centres. A variety of Electron Paramagnetic Resonance (EPR) techniques is used as a unique analytical tool to probe Cu(II) complex chemistry.

While the counter ions had only a negligible effect on coordination through outer sphere interactions, the effect of the type of tetraamine, pH and their concentration was significant, revealing subtle and strong variations in the coordination chemistry upon change of these conditions and thus emphasizing the importance of understanding the solution-based structures when aiming for specific applications.

The performance of different trigonal bipyramidal Cu(II)-tetraamine complexes for the selective oxidation of glycerol was further explored. The interest in glycerol oxidation is growing, since glycerol is a valuable bio-renewable compound formed during biomass conversion. Through a combination of different techniques, the catalytic behaviour could be fit to the faith of the Cu(II) complex during reaction. Attempts were made to heterogenise the Cu(II) complexes into Y zeolites in order to allow easy removal of the catalyst from the reaction mixture after glycerol oxidation. Though the correlation between Cu(II)-complex encapsulation, Si:Al ratio, and proton count in the zeolitic structure was identified, the heterogeneous material proved unsuitable for glycerol oxidation. Nevertheless, it holds promise for exploring alternative catalytic reactions.

Gunter Flipkens

• 19/02/2024
  
• 1 p.m.
• Venue: Campus Drie Eiken, O.1
• ​Online PhD defence
• Supervisors: Ronny Blust & Raewyn Town
• Department of Biology

Abstract

Drastic greenhouse gas emission reductions and gigaton-scale atmospheric carbon dioxide (CO2) removal are needed to keep global warming below 2°C. Silicate rock weathering has regulated climate on earth over geological time scales. Coastal enhanced silicate weathering (CESW) aims to accelerate this process by distributing gigatons of finely ground olivine-rich rock in dynamic coastal environments. Olivine is a proposed candidate mineral due to its abundance, relatively fast weathering rate, and theoretically high CO2 sequestration potential. However, the in situ CO2 sequestration potential remains uncertain and olivine’s high nickel (Ni) and chromium (Cr) content could be of potential ecological concern. This thesis aimed to advance our understanding of olivine dissolution and CO2 sequestration kinetics under the influence of hydrodynamics and assess potential ecotoxicological effects of CESW.

First, we investigated the effect of continuous grain-grain collisions on olivine weathering rates in seawater. Physical agitation enhanced olivine dissolution by 8 to 19 times compared to stagnant conditions, likely due to advective pore water flushing. Therefore, olivine should be supplied in coastal areas with sufficiently high bed shear stress and pore water exchange rates. Subsequently, a flume experiment was conducted to investigate the effect of current on olivine dissolution in permeable sediment. Olivine dissolution was more than one order of magnitude lower than expected, for reasons that could not be identified, highlighting the need for studies under environmentally realistic conditions.

Next, a first assessment of the safe olivine deployment scale was made based on existing marine Ni and Cr environmental quality standards. Results indicated that 0.059 to 1.4 kg of olivine per m2 of seabed could be supplied without posing metal toxicity risks for benthic biota. Changes in sediment physicochemical properties may also lead to avoidance of olivine rich sediments by marine organisms. Short choice experiments with the gastropod Littorina littorea and amphipod Gammarus locusta indicated avoidance of pure olivine but tolerance to environmentally relevant olivine concentrations of 30% w/w and lower. Metal bioaccumulation and chronic olivine toxicity testing with Gammarus locusta further revealed concentration and grain size dependent effects, with adverse reproductive outcomes at olivine concentrations of 10% w/w and higher. These findings underscore the necessity for additional olivine toxicity data to derive accurate, site-specific olivine application guidelines. Overall, our work provides novel insights into the stimulating effect of hydrodynamics on olivine reactivity and shows possible trace metal-related adverse ecological impacts of CESW.

Mikhail Mychinko

• 12/02/2024
  
• 4 p.m.
• Venue: Campus Groenenborger, US.024
• ​Online PhD defence​
• Supervisor: Sara Bals
• Department of Physics

Abstract

During the past decades, metallic nanoparticles (NPs) have attracted great attention in materials science due to their specific optical properties based on surface plasmon resonances. Because of these phenomena, plasmonic NPs (or nanoplasmonics) are very promising for application in biosensing, photocatalysts, medicine, data storage, solar energy conversion, etc. Currently, colloidal synthesis techniques enable scientists to routinely produce mono and bimetallic NPs of various shapes, sizes, composition, and elemental distribution, with superior properties for plasmonic applications. Two primary directions for further advancing nanoplasmonic-based technologies include synthesizing novel morphologies, such as highly asymmetric chiral NPs, and gaining deeper insights into the factors affecting the stability of produced nanoplasmonics. With the increasing complexity of nanoplasmonics’ morphologies and higher stability requirements, there is a pressing need for thorough investigations into their 3D structures and their evolution under different conditions, with high resolution. Electron tomography (ET) emerges as an ideal tool to retrieve shape and element-sensitive information about individual nanoparticles in 3D, achieving resolution down to the atomic level. Moreover, ET techniques can be combined with in situ holders, enabling detailed studies of processes mimicking real applications of nanoplasmonic-based devices. The first part of this defense will focus on structural and morphological characterization of chiral Au NPs, promising for spectroscopy techniques based on the differential absorption of left- and right-handed circularly polarized light. Specifically, I will discuss the primary strategies for wet-colloidal growth of the various types of intrinsically chiral Au NPs. Advanced ET methods will be demonstrated as powerful tools for characterizing the final helical morphologies of the produced Au NPs and for studying the chiral growth mechanisms by examining intermediate structures obtained during chiral growth.

The second part will focus on the stability under heating of various Au@Ag core-shell NPs. Operating in real conditions, such as elevated temperatures, may cause particle reshaping and redistribution of metals between the core and shell, gradually altering nanoplasmonics properties. Hence, a thorough understanding of the influence of size, shape, and defects on these processes is crucial for further developments. I will show how recently developed techniques, combining fast ET with in-situ heating holders, have allowed me to evaluate the influence of various parameters (size, shape, defect structure) on heat-induced elemental redistribution in Au@Ag core-shell nanoparticles qualitatively and quantitatively. Additionally, I will discuss the prospects of high-resolution ET for visualizing the diffusion of individual atoms within complex nanostructures.

Jan Clavel

• 05/02/2024
  
• 2 p.m.
• Venue: Campus Drie Eiken, Q0.02
• Supervisors: Ivan Nijs, Erik Verbruggen & Jonas Lembrechts
• Department of Biology

Abstract

Non-native species invasions are one of the most impactful drivers of biodiversity and ecosystem services loss worldwide and their occurrence is increasing rapidly as a consequence of ever-growing anthropogenic activities. One aspect of plant species invasion, which is only recently being recognized as a significant determinant of invasion success, is the symbiosis between plants and mycorrhizal fungi. Here, I focus on anthropogenic disturbance in mountain ecosystems and its impact on plant communities and mycorrhizal fungi to answer how these communities are impacted by said disturbance and whether non-native plants can benefit from these altered conditions. Therefore, I used three different approaches: 1) repeated surveys of plants and arbuscular mycorrhizal fungi along disturbed roadsides in the mountains of Norway, 2) combining a global plant dataset from along mountain roads with a database associating plants with their mycorrhizal types, and 3) an in-situ seed addition experiment measuring non-native plant success and changes in fungal community following different types of disturbance treatments. Through these methods, I assessed the effects of anthropogenic disturbance on mycorrhizal symbiosis and non-native plant species at multiple scales and resolutions.

I found that road disturbance has a globally consistent effect on mycorrhizal types in mountain systems: plants associated with arbuscular mycorrhizal (AM) fungi were more abundant following disturbance, and vegetation associated with ectomycorrhizal- or ericoid-mycorrhizal fungi was conversely less abundant. In the Norwegian regional study, AM fungi were similarly more abundant and diverse in the roots of plant communities affected by roadside disturbance. Experimental results found that physical disturbance and nutrient addition facilitate non-native plant success, have negative effects on EcM fungi and positive effects on fungal pathogens.

Our results show that anthropogenic disturbance has an effect on mycorrhizal fungi and in turn impacts the distribution of plant species in disturbed mountain systems. The resulting shift in mycorrhizal fungi towards AM can facilitate non-native plant success through disruption of the native fungal communities, especially so in high elevation and cold climate regions which are naturally less dominated by AM plants. I believe that these conclusions highlight the role of mycorrhizal symbiosis in understanding plant invasions trajectories and in turn emphasize the importance of closely monitoring sources of anthropogenic disturbance in mountain systems in order to prevent future establishment of non-native plants.

Haifa Rejeb Sfar

• 29/01/2024
  
• 2 p.m.
• Venue: Campus Groenenborger, T.105
• ​Online PhD defence​
• Supervisors: Albert De Roeck & Nick Van Remortel
• Department of Physics

Abstract

This thesis presents a search for long-lived heavy neutral leptons (HNLs) using proton-proton collision events, with a focus on the νMSM model for HNL production. The νMSM model is a theoretical framework that extends the Standard Model of particle physics to include right-handed neutrinos and provides a possible explanation for the observed neutrino masses and mixing angles. We have analyzed a data sample containing two leptons (electron or muon) and jets, with an integrated luminosity of 138 fb1 collected from 2016 to 2018, which corresponds to the full RunII dataset, and have developed a novel jet tagger based on a deep neural network to identify displaced jets from the HNL decay.

To estimate the contribution from background processes, we used an ABCD method, which is a data-driven technique that relies on the correlation between two independent variables to separate signal and background events. We applied this method to the data in sideband regions and determined the expected background in the signal region. No excess in data over the expected background is observed. Limits on the HNL production cross section are derived as a function of the HNL mass and the three coupling strengths (VlN ) to each lepton generation (l).

Our results provide the best limit on the coupling strength for pure muon coupling scenarios, excluding values of |VμN |2 > 5(4) × 107 for Dirac (Majorana) HNLs with a mass of 10 GeV at 95% CL. This has important implications for the viability of the νMSM model and other theoretical models that propose the existence of HNLs. Our methodology, including the use of the jet tagger and the ABCD method, can be applied to future searches for HNLs at higher energies and luminosities.

However, our study has limitations, such as the assumption of a specific HNL production mechanism and the use of simplified background models. Future research could focus on improving the sensitivity of the jet tagger, exploring alternative HNL production scenarios, developing more sophisticated background estimation techniques, and combining the results from existing HNL searches to improve the sensitivity and coverage of the parameter space.

Tong Zhang

• 26/01/2024
  
• 3 p.m.
• Venue: Campus Groenenborger, V.008
• Supervisors: Shoubhik Das & Bert Maes
• Department of Chemistry

Abstract

Due to the climate change, pollutions, energy shortage and other interrelated global crises, there is always an increasing demand for the development of environmentally friendly processes in the chemical industries. In the last two decades, the field of photochemistry has emerged as a potent methodology across diverse domains, enabling the synthesis of numerous intricate compounds through environmentally sustainable means. This thesis elucidates four distinct methodologies concerning the generation of valuable products across diverse domains through the utilization of photoredox and photochemical reactions. The thesis is divided into five chapters:

• Chapter 1: An overview and introductory exposition of the fundamental principles and concepts pertaining to photochemistry are provided.

• Chapter 2: We have enhanced the generation of hydrogen peroxide by introducing an aryl amino group in polymeric carbon nitrides via visible light-mediated photocatalysis. In addition to increasing the efficiency of photocatalytic system, the description of the whole reactive scenario for the polymeric carbon nitrides has been depicted by combining diverse characteristic methods and theoretical calculations. Futhermore, the possible active catalytic sites are identified with the aid of 15N and 19F solid state NMR without using any expensive labeling reagent.

• Chapter 3: We have developed a unique methodology for the generation of α-amino radicals under the irradiation of visible light under a metal-free condition. This strategy is induced by π–π stacking and ion-pairing interactions and facilitated the synthesis of functionalized amines through three-component coupling reactions.

• Chapter 4: We have designed an efficient method for the red light-mediated sulfonyltrifluoromethylation of olefins which provide remarkable regioselectivity. This reaction system has been thoughtfully designed, and excellent substrate compatibility and functional group tolerance exhibits the industrial potential, thus validating the significance of this strategy.

• Chapter 5: We have developed a metal-free photocatalytic system for the transformation of biomass into formic acid. Compared to previous strategies, our method can work efficiently at room temperature and atmospheric pressure. Notably, real biomass and even daily-life-based-materials such as waste papers and oak cork stoppers of wine bottles are also smoothly converted to formic acid.

Lin Zi

• 22/01/2024
  
• 9 a.m.
• Venue: Campus Drie Eiken, Q.002
• Supervisors: Han Asard, Hamada Abd Elgawad & Kris Laukens
• Department of Biology

Abstract

One aspect of climate change is the increased persistence of precipitation regimes (PRs), characterized by alternated longer dry and wet periods. While the ecological impacts of singular extreme events like drought and flood have been extensively studied, the immediate and legacy effects of the evolving more persistent PR, particularly at the molecular level in plants, remain underexplored.

This doctoral thesis aims to bridge this knowledge gap by conducting a large-scale outdoor experiment, applying a range of PR from short to long dry/wet cycles to grassland mesocosms. Ecometabolomics analysis revealed that the metabolome of a relatively sensitive species, Centaurea jacea, shifted under mild PR (10-day dry/wet cycle), while the metabolome of other less sensitive species changed only from a 20-day PR onwards. Accumulation of amino acids, lignin, and decreased non-structural sugar levels are universal responses across several species to increasing PR extremity, while changes in other metabolite classes are exhibited in a more species-specific manner. The sensitive species are less capable of inducing sufficient changes in important molecules such as lignin and phenylalanine, which may partly explain its sensitivity in PR responses.

Beyond immediate effects, my research found that previous exposure to more persistent PR resulted in acclimated grassland communities in the following year. These communities showed increased aboveground productivity and structural sugar content, reduced molecular stress responses and reduced diversity. Furthermore, soil inoculum from more persistent PR promoted the upregulation of several pathways, such as hormone synthesis (e.g., jasmonic acid, abscisic acid, salicylic acid, ethylene), oxidative stress, cell wall modification (e.g., lignin deposition, callose synthesis, cell wall thickening, pectin metabolic process), and chitin catabolic processes, which may provide potential beneficial effects for plants.

In conclusion, this thesis demonstrates that more persistent PR induces significant changes in plant biochemical and transcriptional levels. While these changes may enhance the acclimation of grassland species, they may also decrease nutritive value, potentially altering their role in the feeding of organisms. Species or individuals unable to induce sufficient protective changes may be excluded from the community, leading to a loss of diversity in the ecosystem.

Nathanaël Six

• 18/01/2023
  
• 4 p.m.
• Venue: Campus Drie Eiken, S1
• Supervisors: Jan Sijbers & Jan De Beenhouwer
• Department of Physics

Abstract

X-ray computed tomography (CT) is a powerful and non-invasive technique to visualise the internal structure of an object from a set of X-ray radiographs. Reconstruction algorithms are used to map projection data to a 3D volume. A model of the X-ray acquisition process is used by reconstruction algorithms and the algorithms require a large number of projections to function well. However, in certain applications, the number of projections has to be limited, to reduce total delivered dose, lower acquisition time or because of geometrical constraints. Furthermore, the most commonly used algorithms have a simple linear forward model for X-ray attenuation that does not model the real acquisition accurately. Finally, conventional reconstruction algorithms in CT are not efficient with respect to computation time. In this thesis, we will develop improved reconstruction algorithms for CT by investigating more accurate non-linear forward models, and different numerical optimisation approaches for these models.

Eline Cauwenberghs

• 16/01/2024
  
• 5 p.m.
• Venue: Campus Drie Eiken, Q0.02
• Supervisors: Sarah Lebeer & Kim van Hoorenbeeck
• Department of Bioscience Engineering

Abstract

For several decades, it was widely believed that the healthy respiratory tract was sterile. However, advancements in next-generation sequencing techniques have enabled us to unveil the microbial communities residing within the respiratory tract, both in health and disease. However, there is a lack of in-depth functional insights into these communities to design innovative microbiome therapies for the respiratory tract. Nevertheless, such therapies show great promise as they can act via multifactorial mechanism of action improving respiratory health. Specifically for people with cystic fibrosis (CF), microbiome therapies could complement current treatments by preventing chronic colonization of pathogens at early age and by enhancing the effectiveness of therapies such as antibiotics and modulators through creation of a more stable microbial ecosystem. This PhD thesis aimed to expand the knowledge on the microbial communities within various respiratory tract niches and explore the potential of microbiome therapies as a preventive or complementary treatment strategy for CF. First, we studied the salivary microbiome of 246 women using 16S rRNA amplicon sequencing to investigate whether we could detect bacterial biomarkers linked with respiratory disease. We found that the salivary microbiome was highly preserved among healthy women and lifestyle and host-related parameters had only subtle effects on specific taxa. For the detection of bacterial biomarkers linked with respiratory disease, more targeted sampling methods and insights at a lower taxonomic level were needed. Therefore, we next performed shallow metagenomic shotgun sequencing on respiratory tract samples, including nasopharyngeal, oropharyngeal and sputum samples. The host DNA depletion step was found to reduce the presence of Gram-negative taxa, which play an important role in the human airways. Taken this bias into account, we could show the potential of shallow shotgun sequencing in a clinical setting by identifying important CF pathogens at species level in oropharyngeal and sputum samples providing valuable information about the patient’s disease status. Lastly, we aimed to evaluate the potential of probiotics for CF. Lacticaseibacillus casei AMBR2 demonstrated important adaptation and multifactorial probiotic characteristics in vitro with a focus on CF pathogenesis. Moreover, we identified four putative bacteriocin gene clusters in the genome of AMBR2 of which one exhibited high activity against Gram-negative pathogens.

This work combined microbiome research (16S rRNA amplicon and shallow metagenomic shotgun sequencing) with functional studies to achieve the research goals. It has enriched the understanding regarding microbe-microbe and microbe-host interactions in the context of CF, paving the way for future therapeutic applications.

Lennart Fernandes

• 10/01/2024
  
• 4 p.m.
• Venue: CST, Promotiezaal van de Grauwzusters, Lange Sint-Annastraat 7, 2000 Antwerpen
• ​Online PhD defence
• Supervisors: Michiel Wouters & Jacques Tempere
• Department of Physics

Abstract

In this thesis the behaviour of quantum fluids out of equilibrium is studied. These are ultracold gases in which quantum mechanical effects - usually only visible at the atomic scale - determine the macroscopic properties. The work is composed of several related projects, in which the central role of quantum fluctuations is highlighted.

In the Gaussian theory for quantum systems, fluctuations are approximated as corrections to a classical description of the fluid. A central contribution of this work is the development and application of a new approach that enables this established method to describe quantum states with large, nonlinear fluctuations. Inspired by the study of open quantum systems, we do this by introducing a virtual environment, which continuously measures the system to suppress the growth of quantum fluctuations. This method is successfully applied to the dynamics of a spinor fluid, in which restrictions on the interactions between atoms lead to the formation of a non-equilibrium state with a high degree of quantum entanglement. ​ ​

Finally, the acquired knowledge on spinor fluids is used for their application as a platform for analog gravity. We introduce Hawking radiation emitted by black holes, show that the underlying mechanism is a general property of quantum fields in a flowing background, and apply this analogy to spin waves in a spinor gas.