Introduction: This study explores the application of the ChemSpaceAL pipeline, an AI-driven tool for molecular generation, in discovering drug candidates for various molecular targets, such as the HNH domain of Cas9, Fibroblast Activation Protein-alpha (FAP-alpha) and Trophoblast cell surface antigen 2 (TROP2). The goal was to evaluate the pipeline’s capacity to generate promising molecules and compare them with known inhibitors, including FDA-approved drugs for c-Abl kinase.

Methods: The ChemSpaceAL pipeline employs deep learning, specifically a Generative Pretrained Transformer (GPT)-based model, to generate novel molecules across multiple iterations. Active learning was used to refine the generated molecules by docking them to specific target proteins and scoring them based on predicted binding affinities. In the case of FAP-alpha, known patented inhibitors were scored to create a benchmark for the AI-generated molecules. The process was iteratively improved by adjusting learning parameters, such as the number of epochs and selection thresholds for active learning.

Results: The pipeline demonstrated the ability to generate molecules with a maximum score of 77 for c-Abl kinase, surpassing the highest score among FDA-approved inhibitors (67.5 for bafetinib), while the average score for the generated molecules was 48.5, compared to 53.1 for the FDA-approved inhibitors. In the case of FAP-alpha, known patented inhibitors scored between 10.5 and 21. AI-generated molecules produced comparable results, with an initial average score of 19.19 and a maximum score of 38.5 in the first iteration. Subsequent iterations saw fluctuations in performance, with iterative improvements stabilizing at an average score of 18.62 and a maximum score of 39 by the third iteration. Adjusting the active learning threshold from the top 10% to 20% of scored complexes yielded more substantial improvements in the molecular generation process.

Conclusion: The results suggest that ChemSpaceAL can explore chemical spaces beyond known inhibitors, occasionally identifying novel molecules with superior predicted binding affinity. However, the study highlights the limitations of relying solely on computational scoring methods, as aspects such as bio-availability and off-target effects are not captured. Future work will focus on reducing the number of active learning epochs to balance model performance and exploring more diverse chemical spaces. The development of a graphical user interface (GUI) and interdisciplinary collaboration for experimental validation will further enhance the pipeline’s accessibility and effectiveness, accelerating the drug discovery process

Our world has witnessed the devastating effects of fossil fuels extraction; the primary energy source for modern society. Transitioning to sustainable energy sources requires exploring alternative fuel production methods, focusing on utilizing captured anthropogenic CO2. Electrochemical reduction of CO2 (CO2RR) using renewable electricity presents an attractive option for storing energy in chemical bonds. While many electrocatalysts efficiently reduce CO2 to CO, only a few, such as copper, can produce high energy-dense molecules like ethanol or methanol with selectivities exceeding 60%. Molecular catalysts (MCs) hold promise as alternatives to metallic electrocatalysts, particularly when immobilized on the surface of gas diffusion electrodes (GDEs), where state-of-the-art CO2RR takes place. However, their efficiency and scalability must be addressed for industrial applications. Our work focuses on evaluating the impact of engineering the micro-environment at the interface where CO2RR occurs, for the purposes of optimizing the performance of MCs systems.

Specifically, the electrode active-site interaction for the CO2-to-ethanol pathway was investigated using a nickel iron-tetraphenylporphyrin (Ni-FeTPP) electrocatalyst. We explored the challenges of translating a molecular catalyst system from an unconventional 3D electrode design to a GDE-based electrolyzer. The difficulty lies in replicating the strong electronic coupling between Ni and FeTPP necessary for enabling the CO2-to-ethanol pathway; which led to two methods of coupling Ni to FeTPP: drop-casting and spray-coating. Although our system did not achieve CO2 reduction to ethanol, it provided significant insights into the impact of subtle micro-environment changes. These changes uncovered possible unintended interactions, suggesting the formation of nickel hydride species that could greatly influence the CO2-to-ethanol pathway catalyzed by Ni-FeTPP. Moreover, the competing behavior between CO2 and CO to enhance the CO2-to-methanol pathway was adressed by using cobalt phthalocyanine supported on multi-walled carbon nanotubes (CoPc/MWCNT). We engineered the local environment where molecular catalysts interact by introducing silver nanoparticles and PTFE particles to increase local CO availability near the active sites and promote further CO reduction. This approach resulted in a remarkable 4-fold and 17-fold increase in the Faradaic efficiency of methanol, from the addition of silver nanoparticles and PTFE particles, respectively. Our results demonstrate the CO2-to-methanol pathway is enhanced through the modification of the catalyst surface microenvironment of the GDE. Our research contributes to a deeper understanding of the prospect for improving the challenging electrochemical reduction of CO2 beyond CO.

A radioisotope used in radionuclide therapy is Holmium-166 (Ho-166). The treatment effectiveness of Ho-166 could be improved by the use of a so called in vivo Dysprosium-166(Dy-166)/Ho-166 generator. The application of this generator is hindered by an effect called internal conversion (IC). This affect can arise after the decay of Dy-166 to Ho-166, which can cause separation of Ho-166 from its carrier.
Polymeric micelles might form a solution in the application of the Dy-166/Ho-166 generator in radionuclide therapy. The main goal of this thesis was to investigate and understand the loading mechanism of metallic species and polymeric micelles with a focus on the loading of Dy/Dy-166 and Ho-166.

It was found that it was not effective to load metallic species (Dy/Dy-166) as free ions or as solid precipitates. Loading metallic species as aqueous hydroxides showed to be crucial for achieving a good loading and high stability. The second goal was to study if polymeric micelles were able to retain Ho-166 inside their core under the effects of internal conversion. No additional losses of Ho-166 were found when Dy/Dy-166 and Ho-166 were loaded into the micelles. It was concluded that the PCL-PEO micelles prevented the loss of Ho-166 under internal conversion effects.

Cancer is one of the leading causes of death in the world. Because of this, there are many novel methods to treat it being currently researched. One of the answers from the nuclear medicine perspective is Radionuclide Therapy, where alpha or beta emitters are applied, with the goal to selectively irradiate tumours in the body. A promising radionuclide researched for Radionuclide Therapy is Holmium-166, which is a beta emitter with a short half-life of 26.8 hours, which is useful for the treatment of large metastases. A method to ensure an effective treatment with Holmium-166 is the use of a Dysprosium-166/Holmium-166 in vivo generator, as the dose delivered to patients per administered dose is two times higher with the in vivo generator rather than the direct use of Holmium-166. However, carriers used in Radionuclide Therapy, which usually involve the formation of a chelator-metal complex are not effective for the in vivo generator due to the release of auger electrons during the decay process leading to the destruction of the chelator-metal complex. A promising non-chelator method designed by Liu et al. at the Applied Radiation and Isotopes research team at TU Delft involves the radiolabelling of micelles. This thesis sought to evaluate the use of micelles, radiolabelled with this mechanism as an effective carrier for the Dysprosium-166/Holmium-166 in vivo generator. For this, micelles made of Polycaprolactone-block-Polyethylene Oxide and Polylactic Acidblock-Polyethylene oxide were evaluated based on the obtained radiolabelling efficiency and their stability when challenged with diethylenetriaminepentaacetic acid (DTPA). The results in this study show that the radiolabelling method used to encapsulate Dysprosium relies on the diffusion of Dysprosium hydroxides into the micelle core, followed by precipitation as the right concentration is reached inside the micelles. The best results obtained by this study occur when radiolabelling Polylactic Acid-block-Polyethylene Oxide (PLA-PEO) micelles, and then adding phosphate ions at a concentration of 3 × 10−8𝑀, 30 minutes after the addition of the active Dysprosium. Although more tuning is required on the radiolabelling mechanism to make micelles the most effective carriers of the 166Dy/166Ho in vivo generator, promising first steps were made to begin this process

Most modern cancer treatments are a combination of therapies. In approximately 50% of these combinations radiation therapy plays a prominent role. Although external radiation therapy is very effective,
it still damages healthy tissue. Therefore, ways to increase radiation dose deposition at the tumor are sought. The interaction of radiation with the photocatalytic nanoparticle titanium dioxide, TiOኼ NPs, is investigated in effort to improve the effectiveness of radiation therapy, minimizing collateral damage to healthy tissue.
TiOኼ NPs, upon the absorption of light, are capable of producing reactive oxygen species, which are the main contributor to DNA strand breakage and cell damage through oxidative reactions in radiation therapy. Singlet oxygen is the most potent of these reactive oxygen species. This research aims to
determine whether TiOኼ NPs are capable of producing singlet oxygen when interacting with ionizing radiation. Efforts are made in uncovering the mechanisms and dependencies behind the production of singlet oxygen using the fluorescent probe Singlet Oxygen Sensor Green, SOSG. The oxidative capabilities of the TiOኼ NPs in combination with ionizing radiation are looked into with the probe Mythelene blue, MB, which degrades under oxidative reactions.
The results for aerobic conditions show a significant rise in fluorescent intensity by the probe SOSG when irradiated in the presence TiOኼ NPs. The rise in fluorescence appears to be linked to the presence of singlet oxygen but this could not be confirmed. SOSG has also shown a rise in fluorescence in
anaerobic circumstances not suitable for singlet oxygen production. This signal response is attributed to the increased presence of hydrated electrons, in line with hypothesized sensitivity of SOSG towards hydrated electrons. The mechanisms behind the production of singlet oxygen could not be uncovered
however the production does favour low photon energies and does not depend on dose rate.
The results from the MB degradation experiments indicate no significant amount of degradation is caused by the photocatalytic activity of TiOኼ NPs concentrations used in this thesis. The degradation that does occur is due to radiolysis, which favours low photon energies due to the increased absorbance.
The production of singlet oxygen through the oxidation of superoxide radicals, facilitated by the interaction of ionizing photon irradiation with TiOኼ NPs, is expected but could not be verified. Future research may verify the origin of the rise in fluorescencent SOSG to be hydrated electrons or singlet oxygen.

Synthesis of new perylene-3,4,9,10-tetra-carboxylic acid (PTCA) derivates is a research topic that has rapidly developed during the last decades. Current synthetic efforts are largely focussed on functionalizing the perylene core, via either bay- or ortho-substitution. The aim of this research project was to synthesize novel perylene mono-imide di-ester derivates by attaching different substituents at the bay-positions. The first synthetic route that was attempted in this research project was a route via a perylene mono-anhydride mono-imide derivate. The anhydride was attached to this molecule to activate the relevant bay-positions. It was found that the anhydride is not stable enough to withstand the reaction conditions that were used in an attempt to substitute the relevant bay positions. The next synthetic route that was attempted in this research project was a route via a perylene di-ester mono-imide derivate. Reactions were performed to explore whether the di-ester was able to activate the relevant bay-positions. It was found that at the used reaction conditions, phenol exchange takes place on the bay-positions of the perylene core. Due to the challenges that arose, the target compounds of this project have not been synthesized. However, three other new PTCA derivates have been synthesized during this project.

Since the industrial revolution the use of fossil fuels has increased exponentially, this increase caused a higher concentration of carbon dioxide in the atmosphere, which is responsible for climate change. To prevent a large-scale change of the earth’s climate humanity needs to switch to renewable energy sources. A large drawback of current renewable energy sources is that they almost all exclusively produce electricity which is hard to store for long periods of time. A possible solution to this energy storage issue is the electrochemical reduction of CO2 with electricity from renewable sources. Catalysts used for this process are transition metals but there are several problems that prevent large-scale implementation. Two fundamental problems with these catalysts are: Low activity causing low reaction rates, and low selectivity causing the formation of undesired products. A possible solution was proposed by Lim et al. (2014)1 where they suggested doping a silver electrode with sulphur to enhance the selectivity and activity for carbon monoxide production. This proposal was put the test in this thesis where silver was doped with sulphur by two different methods and were characterized by different techniques and tested at four different potentials. The method where a sample was immersed in a sulphur containing solution and subsequently annealed proofed to be the most effective way to dope a silver electrode. With the help of gas chromatography it was found out that at lower overpotentials the addition of sulphur is beneficial to its performance to produce CO but at higher potentials this advantage is lost. However if the results of this thesis are compared to literature it is clear that doping with sulphur is not effective. Furthermore it was seen that the sulphur doped Ag electrodes were not very stable and therefore not suited for large-scale implementation.