Search

Cell-cell communication plays a significant role in shaping the functionality of the cells. Communication between the cells is also responsible for maintaining the physiological state of the cells and the tissue. Therefore, it is important to study the different ways by which cell-cell communication impacts the functional state of cells. Alterations in cell-cell communication can contribute to the development of disease conditions. In this thesis, we present two computational tools and a study to explore the different aspects of cell-cell communication. In the first manuscript, FunRes was developed to leverage the cell-cell communication model to investigate functional heterogeneity in cell types and characterize cell states based on the integration of interand intra-cellular signalling. This tool utilizes a combination of receptors and transcription factors (TFs) based on the reconstructed cell-cell communication network to split the cell types into functional states. The tool was applied to the TabulaMurisSenis atlas to discover functional cell states in both young and old mouse datasets. In addition, we compared our tool with state-of-the-art tools and validated the cell states using available markers from the literature. Secondly, we studied the evolution of gene expression in developing astrocytes under normal and inflammatory conditions. We characterized these cells using both transcriptional and chromatin accessibility data which were integrated to reconstruct the gene regulatory networks (GRNs) specific to the condition and timepoints. The GRNs were then topologically analyzed to identify key regulators of the developmental process under both normal and inflammatory conditions. In the final manuscript, we developed a computational tool that identified regulators of allergy and tolerance in a mouse model. The tool works by first reconstructing the cell-cell communication network and then analyzing the network for feedback loops. These feedback loops are impor

Biallelic mutations in PINK1/PRKN cause recessive Parkinson’s disease. Given the established role of PINK1/Parkin in regulating mitochondrial dynamics, we explored mitochondrial DNA (mtDNA) integrity and inflammation as disease modifiers in carriers of mutations in these genes. MtDNA integrity was investigated in a large collection of biallelic (n = 84) and monoallelic (n = 170) carriers of PINK1/PRKN mutations, idiopathic Parkinson’s disease patients (n = 67) and controls (n = 90). In addition, we studied global gene expression and serum cytokine levels in a subset. Affected and unaffected PINK1/PRKN monoallelic mutation carriers can be distinguished by heteroplasmic mtDNA variant load (AUC = 0.83, CI:0.74-0.93). Biallelic PINK1/PRKN mutation carriers harbor more heteroplasmic mtDNA variants in blood (p = 0.0006, Z = 3.63) compared to monoallelic mutation carriers. This enrichment was confirmed in iPSC-derived (controls, n = 3; biallelic PRKN mutation carriers, n = 4) and postmortem (control, n = 1; biallelic PRKN mutation carrier, n = 1) midbrain neurons. Lastly, the heteroplasmic mtDNA variant load correlated with IL6 levels in PINK1/PRKN mutation carriers (r = 0.57, p = 0.0074). PINK1/PRKN mutations predispose individuals to mtDNA variant accumulation in a dose- and disease-dependent manner.

Developing measures for rapid and early detection of disease re-emergence is important to perform science-based risk assessment of epidemic threats. In the past few years, several early warning signals (EWS) from complex systems theory have been introduced to detect impending critical transitions and extend the set of indicators. However, it is still debated whether they are generically applicable or potentially sensitive to some dynamical charac- teristics such as system noise and rates of approach to critical parameter values. Moreover, testing on empirical data has, so far, been limited. Hence, verifying EWS performance remains a challenge. In this study, we tackle this question by analyzing the performance of common EWS, such as increasing variance and autocorrelation, in detecting the emer- gence of COVID-19 outbreaks in various countries. Our work illustrates that these EWS might be successful in detecting disease emergence when some basic assumptions are sat- isfied: a slow forcing through the transitions and not-fat-tailed noise. In uncertain cases, we observe that noise properties or commensurable time scales may obscure the expected early warning signals. Overall, our results suggest that EWS can be useful for active moni- toring of epidemic dynamics, but that their performance is sensitive to certain features of the underlying dynamics. Our findings thus pave a connection between theoretical and empiri- cal studies, constituting a further step towards the application of EWS indicators for inform- ing public health policies.

Biallelic mutations in PINK1/PRKN cause recessive Parkinson’s disease. Given the established role of PINK1/Parkin in regulating mitochondrial dynamics, we explored mitochondrial DNA (mtDNA) integrity and inflammation as disease modifiers in carriers of mutations in these genes. MtDNA integrity was investigated in a large collection of biallelic (n = 84) and monoallelic (n = 170) carriers of PINK1/PRKN mutations, idiopathic Parkinson’s disease patients (n = 67) and controls (n = 90). In addition, we studied global gene expression and serum cytokine levels in a subset. Affected and unaffected PINK1/PRKN monoallelic mutation carriers can be distinguished by heteroplasmic mtDNA variant load (AUC = 0.83, CI:0.74-0.93). Biallelic PINK1/PRKN mutation carriers harbor more heteroplasmic mtDNA variants in blood (p = 0.0006, Z = 3.63) compared to monoallelic mutation carriers. This enrichment was confirmed in iPSC-derived (controls, n = 3; biallelic PRKN mutation carriers, n = 4) and postmortem (control, n = 1; biallelic PRKN mutation carrier, n = 1) midbrain neurons. Lastly, the heteroplasmic mtDNA variant load correlated with IL6 levels in PINK1/PRKN mutation carriers (r = 0.57, p = 0.0074). PINK1/PRKN mutations predispose individuals to mtDNA variant accumulation in a dose- and disease-dependent manner.

Inflammatory responses are evolutionarily conserved reactions to pathogens, injury, or any form of a serious perturbation of a human organism. These mechanisms evolved together with us and, although capable of somewhat adapting, innate responses are gravely impacted by prolonged human lifespan. Better sanitary measures, health systems, food and medicine supply have prolonged human life expectancy to ~72 years. Aging is characterized by prolonged, chronic (often low-grade) inflammation. With tissue and cellular defense mechanisms becoming dysfunctional over time, this inflammation becomes detrimental and destructive to the human body. Aging is a major risk factor for Parkinson’s disease (PD), a movement disorder characterized by the loss of dopaminergic neurons. Even though the disease is predominantly idiopathic, genetic cases are contributing to a better understanding of the underlying cellular and neuropathological mechanisms. In comparison to neuronal demise, the contribution of microglia (the immune cells of the brain) to PD is relatively understudied. Initially studied in PD patient-derived post-mortem tissue, novel in vitro technologies, such as induced pluripotent stem cells (iPSCs), are permitting the generation of specific cell types of interest in order to study disease mechanisms. We derived microglia cells from iPSCs of patients and healthy or isogenic controls to explore (shared) pathological immune responses in LRRK2-PD and idiopathic PD. Our findings suggest a significant involvement of microglia cells in the pathogenesis of PD and highlight potential therapeutic targets in alleviating overactive immune responses.

The discovery of iPSC technology revolutionized the biomedical field, allowing the development of translatable and complex 2D and 3D cell culture systems. Organoids are 3D models containing multiple cell types that mimic complex microenvironments. This is highly advantageous to understand human development, physiology and disease, especially in inaccessible areas such as the brain. Human midbrain-specific organoids have been developed to study the midbrain (abundant in dopaminergic neurons). In Parkinson’s Disease (PD), dopaminergic neurons in the substantia nigra of the midbrain degenerate, causing a broad spectrum of clinical features. Midbrain organoids (MO) are rich in dopaminergic neurons, and contain spatially organized groups of neural cells and progenitors. MO generated from PD patients’ cells recapitulate dopaminergic neuron degeneration. In this thesis, we first demonstrated that dopaminergic neuron PD phenotypes and drug rescue effects were similar between MO and mice. After, we identified different neuronal clusters, progenitor cells, radial glia and mesenchymal cells in MO by scRNA-Seq. As expected, due to the neuro-ectodermal patterning of the MO’ starting cell population, we confirmed the absence of mesoderm-derived cell types, such as microglia and endothelial cells. This represents a limitation for the system in terms of cellular and molecular complexity. Microglia in the human brain perform surveillance, defence and homeostasis functions; they phagocytose metabolic waste products and cell debris. We successfully developed a novel protocol to integrate functional microglia into our MO model. SnRNA-Seq analysis and electrophysiological results suggested a reduction of stress levels and higher maturation of neurons in the presence of microglia, respectively. We then aimed to vascularise MO, which would better recapitulate the brain environment and improve oxygen and nutrient supply into the organoid core (a common 3D culture limitation). We integrated

Antimicrobial resistance (AMR) presents a global threat to public health due to the inability to comprehensively treat bacterial infections. Emerging resistant bacteria residing within human, animal and environmental reservoirs may spread from one to the other, at both local and global levels. Consequently, AMR has the potential to rapidly become pandemic whereby it is no longer constrained by either geographical or human-animal borders. Therefore, to enhance our understanding on the dissemination of AMR we systematically resolved different reservoirs of antimicrobial resistance, leveraging animal, environmental and human samples, to provide a One Health perspective. To identify antimicrobial resistance genes (ARGs) and compare their identity and prevalence across different microbial reservoirs, we developed the PathoFact pipeline which also contextualizes ARG localization on mobile genetic elements (MGEs). This methodology was applied to several metagenomic datasets covering microbiomes of infants, laboratory mice, a wastewater treatment plant (WWTP) and biofilms from glacier-fed streams (GFS). Investigating the infant gut resistome we found that the abundance of ARGs against (semi)-synthetic agents were increased in infants born via caesarian section compared to those born via vaginal delivery. Additionally, we identified mobile genetic elements (MGEs) encoding ARGs such as glycopeptide, diaminopyrimidine and multidrug resistance at an early age. MGEs are often pivotal in the accumulation and dissemination of AMR within a microbial population. Therefore, we assessed the effect of selective pressure on the evolution and consecutive dissemination of AMR within the commensal gut microbiome, utilizing a mouse model. While plasmids and phages were found to contribute to the spread of AMR, we found that integrons represented the primary factors mediating AMR in the antibiotic-treated mice. In addition to the above-described studies, we investigated the environmental res

With increasing prevalence, Parkinson’s disease presents a major challenge for medical research and public health. Despite years of investigation, significant knowledge gaps exist and Parkinson’s disease (PD) etiology remains unclear. A recent concept in the field is that neurodevelopmental aspects might contribute to the pathogenesis of neurodegenerative diseases such as PD. Our hypothesis is that mutations in PD-linked genes have an impact on the cells’ homeostasis at the neural precursor state, giving rise to vulnerable dopaminergic (DA) neurons, thereby increasing the degree of susceptibility for neurodegeneration with aging. In order to investigate this, we used a human midbrain organoid (hMO) model generated from iPSC-derived neural precursor cells. As part of the optimization of the model, we treated the organoids with the neurotoxin 6-OHDA to develop a neurotoxin-induced PD model and set up a high-content imaging pipeline coupled with machine learning classification to predict neurotoxicity. We then used these tools to derive PD patient-derived hMOs in order to investigate our main hypothesis. First, we focused on PD patients carrying a heterozygous mutation in the GBA gene. We developed a genome-scale metabolic model that predicted significant differences in lipid metabolism between patients and controls. Then, we validated the observations by performing a comprehensive lipidomics analysis confirming a dysregulated lipidome in mutant hMOs. Moreover, GBA-PD hMOs displayed PD-relevant phenotypes, impaired DA differentiation and an increased population of neural progenitor cells (NPCs) in cell cycle arrest, confirming the presence of neurodevelopmental defects. To further investigate the neurodevelopmental component of PD, we used patient-derived cell lines carrying PINK1 mutations. PINK1-PD neural precursors presented differences in their energetic profile, imbalanced proliferation, apoptosis, mitophagy, and an impaired differentiation efficiency to DA neuron

Deep brain stimulation (DBS) is a surgical therapy to alleviate symptoms of numerous movement and psychiatric disorders by electrical stimulation of specific neural tissues via implanted electrodes. Precise electrode implantation is important to target the right brain area. After the surgery, DBS parameters, including stimulation amplitude, frequency, pulse width, and selection of electrode’s active contacts, are adjusted during programming sessions. Programming sessions are normally done by trial and error. Thus, they can be long and tiring. The main goal of the thesis is to make the post-operative experience, particularly the programming session, easier and faster by using visual aids to create a virtual reconstruction of the patient’s case. This enables in silico testing of different scenarios before applying them to the patient. A quick and easy-to-use deep learning-based tool for deep brain structure segmentation is developed with 89% ± 3 accuracy (DBSegment). It is much easier to implement compared to widely-used registration-based methods, as it requires less dependencies and no parameter tuning. Therefore, it is much more practical. Moreover, it segments 40 times faster than the registration-based method. This method is combined with an electrode localization method to reconstruct patients’ cases. Additionally, we developed a tool that simulates DBS-induced electric field distributions in less than a seconds (FastField). This is 1000 times faster than standard methods based on finite elements, with nearly the same performance (92%). The speed of the electric field simulation is particularly important for DBS parameter initialization, which we initialize by solving an optimization problem (OptimDBS). A grid search method confirms that our novel approach convergences to the global minimum. Finally, all the developed methods are tested on clinical data to ensure their applicability. In conclusion, this thesis develops various novel user-friendly tools, enabling

From population collapses to cell-fate decision, critical phenomena are abundant in complex real-world systems. Among modelling theories to address them, the critical transitions framework gained traction for its purpose of determining classes of critical mechanisms and identifying generic indicators to detect and alert them (“early warning signals”). This thesis contributes to such research field by elucidating its relevance within the systems biology landscape, by providing a systematic classification of leading mechanisms for critical transitions, and by assessing the theoretical and empirical performance of early warning signals. The thesis thus bridges general results concerning the critical transitions field – possibly applicable to multidisciplinary contexts – and specific applications in biology and epidemiology, towards the development of sound risk monitoring system.

Systems biology is an interdisciplinary approach investigating complex biological systems at different levels by combining experimental and modelling approaches to understand underlying mechanisms of health and disease. Complex systems including biological systems are affected by a plethora of interactions and dynamic processes often with the aim to ensure robustness to emer- gent system properties. The need for interdisciplinary approaches became very evident in the recent COVID-19 pandemic spreading around the globe since the end of 2019. This pandemic came with a bundle of urgent epidemiological open questions including the infection and transmis- sion mechanisms of the virus, its pathogenicity and the relation to clinical symptoms. During the pandemic, mathematical modelling became an essential tool to integrate biological and healthcare data into mechanistic frameworks for projections of future developments and the assessment of different mitigation strategies. In this regard, systems biology with its interdisciplinary approach was a widely applied framework to support society in the COVID-19 crisis. In my thesis, I applied different mathematical modelling approaches as a tool to identify underlying mechanisms of the complex dynamics of the COVID-19 pandemic with a specific focus on the situation in Luxembourg. For this purpose, I analysed the COVID-19 pandemic at its different phases and from various perspectives by investigating mitigation strategies, consequences in the healthcare and economical system, and pandemic preparedness in terms of early-warning signals for re-emergence of new COVID-19 outbreaks by extended and adapted epidemiological Susceptible-Exposed-Infectious-Recovered (SEIR) models.

Real-world evaluations of metagenomic reconstructions are challenged by distinguishing reconstruction artifacts from genes and proteins present in situ. Here, we evaluate short-read-only, long-read-only and hybrid assembly approaches on four different metagenomic samples of varying complexity. We demonstrate how different assembly approaches affect gene and protein inference, which is particularly relevant for downstream functional analyses. For a human gut microbiome sample, we use complementary metatranscriptomic and metaproteomic data to assess the metagenomic data-based protein predictions. Our findings pave the way for critical assessments of metagenomic reconstructions. We propose a reference-independent solution, which exploits the synergistic effects of multi-omic data integration for the in situ study of microbiomes using long-read sequencing data.

In the summer of 2021, European governments removed most NPIs after experiencing prolonged second and third waves of the COVID-19 pandemic. Most countries failed to achieve immunization rates high enough to avoid resurgence of the virus. Public health strategies for autumn and winter 2021 have ranged from countries aiming at low incidence by re-introducing NPIs to accepting high incidence levels. However, such high incidence strategies almost certainly lead to the very consequences that they seek to avoid: restrictions that harm people and economies. At high incidence, the important pandemic containment measure ‘test-trace-isolate-support’ becomes inefficient. At that point, the spread of SARS-CoV-2 and its numerous harmful consequences can likely only be controlled through restrictions. We argue that all European countries need to pursue a low incidence strategy in a coordinated manner. Such an endeavour can only be successful if it is built on open communication and trust.

Background Pathogenic microorganisms cause disease by invading, colonizing, and damaging their host. Virulence factors including bacterial toxins contribute to pathogenicity. Additionally, antimicrobial resistance genes allow pathogens to evade otherwise curative treatments. To understand causal relationships between microbiome compositions, functioning, and disease, it is essential to identify virulence factors and antimicrobial resistance genes in situ. At present, there is a clear lack of computational approaches to simultaneously identify these factors in metagenomic datasets. Results Here, we present PathoFact, a tool for the contextualized prediction of virulence factors, bacterial toxins, and antimicrobial resistance genes with high accuracy (0.921, 0.832 and 0.979, respectively) and specificity (0.957, 0.989 and 0.994). We evaluate the performance of PathoFact on simulated metagenomic datasets and perform a comparison to two other general workflows for the analysis of metagenomic data. PathoFact outperforms all existing workflows in predicting virulence factors and toxin genes. It performs comparably to one pipeline regarding the prediction of antimicrobial resistance while outperforming the others. We further demonstrate the performance of PathoFact on three publicly available case-control metagenomic datasets representing an actual infection as well as chronic diseases in which either pathogenic potential or bacterial toxins are hypothesized to play a role. In each case, we identify virulence factors and AMR genes which differentiated between the case and control groups, thereby revealing novel gene associations with the studied diseases. Conclusion PathoFact is an easy-to-use, modular, and reproducible pipeline for the identification of virulence factors, bacterial toxins, and antimicrobial resistance genes in metagenomic data. Additionally, our tool combines the prediction of these pathogenicity factors with the identification of mobile genetic elements.

Against the COVID-19 pandemic, non-pharmaceutical interventions have been widely applied and vaccinations have taken off. The upcoming question is how the interplay between vaccinations and social measures will shape infections and hospitalizations. Hence, we extend the Susceptible-Exposed-Infectious-Removed (SEIR) model including these elements. We calibrate it to data of Luxembourg, Austria and Sweden until 15 December 2020. Sweden results having the highest fraction of undetected, Luxembourg of infected and all three being far from herd immunity in December. We quantify the level of social interaction, showing that a level around 1/3 of before the pandemic was still required in December to keep the effective reproduction number Refft below 1, for all three countries. Aiming to vaccinate the whole population within 1 year at constant rate would require on average 1,700 fully vaccinated people/day in Luxembourg, 24,000 in Austria and 28,000 in Sweden, and could lead to herd immunity only by mid summer. Herd immunity might not be reached in 2021 if too slow vaccines rollout speeds are employed. The model thus estimates which vaccination rates are too low to allow reaching herd immunity in 2021, depending on social interactions. Vaccination will considerably, but not immediately, help to curb the infection; thus limiting social interactions remains crucial for the months to come.

Machine learning and data-driven solutions open exciting opportunities in many disciplines including healthcare. The recent transition to this technology into real clinical settings brings new challenges. Such problems derive from several factors, including their dataset origin, composition and description, hampering their fairness and secure application. Considering the potential impact of incorrect predictions in applied-ML healthcare research is urgent. Undetected bias induced by inappropriate use of datasets and improper consideration of confounders prevents the translation of prediction models into clinical practice. Therefore, in this work, the use of available systematic tools to assess the risk of bias in models is employed as the first step to explore robust solutions for better dataset choice, dataset merge and design of the training and validation step during the ML development pipeline.

Background: High resolution mass spectrometry (HRMS) is being used increasingly in the context of suspect and non-targeted screening for the identification of bioorganic molecules. There is correspondingly increasing awareness that higher confidence identification will require a systematic, group effort to increase the fraction of compounds with tandem mass spectra available in central, publicly available resources. While typical suspect screening efforts will only result in tentative annotations with a moderate level of confidence, library spectral matches will yield higher confidence or even full confirmation of the identity if the reference standards are available. Results: This article first explores representative percent coverage of measured tandem mass spectra in selected major environmental suspect databases of interest in the context of human biomonitoring, demonstrating the current extensive gap between the number of potential substances of interest (up to hundreds of thousands) and measured spectra (0.57–3.6% of the total chemicals have spectral information available). Furthermore, certain datasets are benchmarked, based on previous efforts, to show the extent to which acquired experimental data were comparable between laboratories, even with HRMS instruments based on different technologies (i.e., quadrupole–quadrupole-time of flight versus ion trap/quadrupole-Orbitrap). Instruments and settings that are less comparable are also revealed, primarily linear ion trap instruments, which show distinctly lower comparability. Conclusions: Based on these efforts, harmonization guidelines for the acquisition and processing of tandem mass spectrometry data are proposed to enable European (and ideally worldwide) laboratories to contribute to common resources, without requiring extensive changes to their current in house methods.

Parkinson's disease (PD) is a neurodegenerative disorder caused by a complex interplay of genetic and environmental factors. For the stratification of PD patients and the development of advanced clinical trials, including causative treatments, a better understanding of the underlying genetic architecture of PD is required. Despite substantial efforts, genome-wide association studies have not been able to explain most of the observed heritability. The majority of PD-associated genetic variants are located in non-coding regions of the genome. A systematic assessment of their functional role is hampered by our incomplete understanding of genotype-phenotype correlations, for example through differential regulation of gene expression. Here, the recent progress and remaining challenges for the elucidation of the role of non-coding genetic variants is reviewed with a focus on PD as a complex disease with multifactorial origins. The function of gene regulatory elements and the impact of non-coding variants on them, and the means to map these elements on a genome-wide level, will be delineated. Moreover, examples of how the integration of functional genomic annotations can serve to identify disease-associated pathways and to prioritize disease- and cell type-specific regulatory variants will be given. Finally, strategies for functional validation and considerations for suitable model systems are outlined. Together this emphasizes the contribution of rare and common genetic variants to the complex pathogenesis of PD and points to remaining challenges for the dissection of genetic complexity that may allow for better stratification, improved diagnostics and more targeted treatments for PD in the future.

Background: The role of schools and children in the transmission of SARS-CoV-2 remains to be determined. Following a first wave in spring and gradual easing of lockdown, Luxembourg experienced an early second epidemic wave before the start of summer school holidays on 15th July. This provided the opportunity to study the role of school-age children and school settings in SARS-CoV-2 transmission. More specifically, we compared the incidence in school-age children, teachers and the general working population, and estimated the number of secondary transmissions occurring at schools using contact tracing data. Findings: While SARS-CoV-2 incidence was much higher in adults aged 20 and above than in children aged 0 to 19 during the first wave in spring, no significant difference was found during the second wave in early summer. The incidence during the second wave was similar for pupils, teachers and the general working population. Based on a total of 424 reported confirmed COVID-19 cases in school-age children and teachers, we estimate that 179 index cases caused 49 secondary transmissions in schools. While some small clusters of mainly student-to-student transmission within the same class were identified, we did not observe any large outbreaks with multiple generations of infection. Interpretation: Transmission of SARS-CoV-2 within Luxembourg schools was limited during the early summer epidemic wave in 2020. Precautionary measures including physical distancing as well as easy access to testing, systematic contact tracing appears to have been successful in mitigating transmission within educational settings. Funding Statement: LV is supported by the Luxembourg National Research Fund grant COVID-19/2020- 1/14701707/REBORN, LM is supported by Luxembourg National Research Fund grant COVID19/14863306/PREVID, PW is supported by the European Research Council (ERC-CoG 863664). Declaration of Interests: No competing interests. Ethics Approval Statement: The Health Directorate ha

Parkinson’s disease (PD) exhibits systemic effects on human metabolism with emerging roles for the gut microbiome. Here, we integrated longitudinal metabolome data from 30 drug-naïve, de-novo PD patients and 30 matched controls with constraint-based modeling of gut microbial communities derived from an independent, drug-naïve PD cohort, and prospective data from a general population. Our key results are i) longitudinal trajectory of metabolites associated with the interconversion of methionine and cysteine via cystathionine differed between PD patients and controls, ii) dopaminergic medication showed strong lipidomic signatures, iii) taurine-conjugated bile acids correlated with the severity of motor symptoms, while low levels of sulfated taurolithocholate were associated with incident PD in the general population, and iv) computational modeling predicted changes in sulfur metabolism, driven by A. muciniphila and B. wadsworthia, consistent with the changed metabolome. In conclusion, the multi-omics integration revealed PD-specific patterns in microbial-host sulfur co-metabolism that may contribute to PD severity.

With increasing prevalence, neurodegenerative disorders present a major challenge for medical research and public health. Despite years of investigation, significant knowledge gaps exist, which impede the development of disease-modifying therapies. The development of tools to model both physiological and pathological human brains greatly enhanced our ability to study neurological disorders. Brain organoids, derived from human induced pluripotent stem cells (iPSCs), hold unprecedented promise for biomedical research to unravel novel pathological mechanisms of a multitude of brain disorders. As brain proxies, these models bridge the gap between traditional 2D cell cultures and animal models. Owing to their human origin, hiPSC-derived organoids can recapitulate features that cannot be modeled in animals by virtue of differences in species. Parkinson’s disease (PD) is a human-specific neurodegenerative disorder. The major manifestations are the consequence of degenerating dopaminergic neurons (DANs) in the midbrain. The disease has a multifactorial etiology and a multisystemic pathogenesis and pathophysiology. In this thesis, we used state-of-the-art technologies to develop a human midbrain organoid (hMO) model with a great potential to study PD. hMOs were generated from iPSC-derived neural precursor cells, which were pre-patterned to the midbrain/hindbrain region. hMOs contain multiple midbrain-specific cell types, such as midbrain DANs, as well as astrocytes and oligodendrocytes. We could demonstrate features of neuronal maturation such as myelination, synaptic connections, spontaneous electrophysiological activity and neural network synchronicity. We further developed a neurotoxin-induced PD organoid model and set up a high-content imaging platform coupled with machine learning classification to predict neurotoxicty. Patient-derived hMOs display PD-relevant pathomechanisms, indicative of neurodevelopmental deficits. hMOs as novel in vitro models open up new avenues t

The study of 3D cell culture models not only bridges the gap between traditional 2D in vitro experiments and in vivo animal models, it also addresses processes that cannot be recapitulated by these traditional models. Therefore, it offers an opportunity to better understand complex biology, for instance brain development, where conventional models have not proven successful. The so{called brain organoid technology provides a physiologically relevant context, which holds great potential for its application in modelling neurological diseases. To obtain these highly specialised structures, resembling specifically key features of the human midbrain, we derived a human midbrain-specific organoid (hMO) system from regionally patterned neural stem cells (NSCs). The resulting neural tissue exhibited abundant neurons with midbrain dopaminergic neuron (mDAN) identity, as well as astroglia and oligodendrocyte di erentiation. Within the hMOs, we could observe neurite myelination and the formation of synaptic connections. Regular fire patterning and neural network synchronicity were determined by multielectrode array (MEA) recordings. In addition to electrophysiologically functional mDANs producing and secreting dopamine (DA), we also detected responsive neuronal subtypes, like GABAergic and glutamatergic neurons. To investigate Parkinson's disease (PD)-relevant pathomechanisms, we derived hMOs from PD patients carrying the LRRK2-G2019S mutation and compared them to healthy control hMOs. In addition to a reduced number and complexity of mDANs, we determined a signi cant increase of the stem cell marker FOXA2 in the patient-derived hMOs. This suggests a neurodevelopmental defect induced by a PD-specific mutation and emphasises the importance of advanced three-dimensional (3D) stem cell-based in vitro models. The in this thesis described hMOs are suitable to reveal PD{relevant phenotypes, thus constitute as a powerful tool for human-specific in vitro disease modelling of neurologi

By modulating the human gut microbiome, prebiotics and probiotics (combinations of which are called synbiotics) may be used to treat diseases such as colorectal cancer (CRC). Methodological limitations have prevented determining the potential combina- torial mechanisms of action of such regimens. We expanded our HuMiX gut-on-a-chip model to co-culture CRC-derived epithelial cells with a model probiotic under a simulated prebiotic regimen, and we integrated the multi-omic results with in silico metabolic modeling. In contrast to individual prebi- otic or probiotic treatments, the synbiotic regimen caused downregulation of genes involved in procarci- nogenic pathways and drug resistance, and reduced levels of the oncometabolite lactate. Distinct ratios of organic and short-chain fatty acids were produced during the simulated regimens. Treatment of primary CRC-derived cells with a molecular cocktail reflecting the synbiotic regimen attenuated self-renewal ca- pacity. Our integrated approach demonstrates the potential of modeling for rationally formulating synbi- otics-based treatments in the future.

While genetic advances have successfully defined part of the complexity in Parkinson’s disease (PD), the clinical characterization of phenotypes remains challenging. Therapeutic trials and cohort studies typically include patients with earlier disease stages and exclude comorbidities, thus ignoring a substantial part of the real-world PD population. To account for these limitations, we implemented the Luxembourg PD study as a comprehensive clinical, molecular and device-based approach including patients with typical PD and atypical parkinsonism, irrespective of their disease stage, age, comorbidities, or linguistic background. To provide a large, longitudinally followed, and deeply phenotyped set of patients and controls for clinical and fundamental research on PD, we implemented an open-source digital platform that can be harmonized with international PD cohort studies. Our interests also reflect Luxembourg-specific areas of PD research, including vision, gait, and cognition. This effort is flanked by comprehensive biosampling efforts assuring high quality and sustained availability of body liquids and tissue biopsies. We provide evidence for the feasibility of such a cohort program with deep phenotyping and high quality biosampling on parkinsonism in an environment with structural specificities and alert the international research community to our willingness to collaborate with other centers. The combination of advanced clinical phenotyping approaches including device-based assessment will create a comprehensive assessment of the disease and its variants, its interaction with comorbidities and its progression. We envision the Luxembourg Parkinson’s study as an important research platform for defining early diagnosis and progression markers that translate into stratified treatment approaches.

There is a global increase in the incidence of non-communicable diseases associated with unhealthy food intakes. Conditions such as diabetes, heart disease, high blood pressure, and strokes represent a high societal impact and an economic burden for health-care systems around the world. To understand these diseases, one needs to account the several factors that influence how the human body processes food, some of which are determined by the genome and patterns of gene expression that translate to the ability - or lack of - to degrade and absorb certain nutrients. Other factors, like the gut microbiota, are more volatile because its composition is highly moldable by diet and lifestyle. Multi-omics technologies can support the comprehensive collection of dietary intake data and monitoring of the health status of individuals. Also, a correct analysis of this data could lead to new insights about the complex processes involved in the digestion of dietary components and their involvement in the prevention or the appearance of health problems, but its integration and interpretation are still problematic. Thus, in this thesis, we propose the utilization of Constraint-Based Reconstruction and Analysis (COBRA) methods as a framework for the integration of this complex data. To achieve this goal, we have created a knowledge-base, the Virtual Metabolic Human (VMH), that combines information from large-scale models of metabolism from the human organism and typical gut microbes, with food composition information, and a disease compendium. VMH’s unique combination of resources leverages the exploration of metabolic pathways from different organisms, the inclusion of dietary information into in-silico experiments through its own diet designer tool, visualization and analysis of experimental and simulation data, and exploring disease mechanisms and potential treatment strategies. VMH is a step forward in providing the necessary tools to investigate the mechanisms behind the influen

Mitochondrial impairment is a well-established pathological pathway implicated in Parkinson’s disease (PD). Defects of the complex I of the mitochondrial respiratory chain have been found in post mortem brains from sporadic PD patients. Furthermore, several disease-related genes are linked to mitochondrial pathways, such as PRKN, PINK1, DJ-1 and HTRA2 and are associated to mitochondrial impairment. This phenotype can be caused by the dysfunction of mitochondrial quality control machinery at different levels: molecular, organellar or cellular. Mitochondrial unfolded protein response represents the molecular level and implicates various chaperones and proteases. If the molecular level of quality control is not sufficient, the organellar level is required and involves mitophagy and mitochondrial derived vesicles to sequester whole dysfunctional organelle or parts of it. Only when the impairment is too severe, it leads to cell death via apoptosis which defines the cellular level of quality control. Here we review how currently known PD-linked genetic variants interfere with the different levels of mitochondrial quality control. We discuss the graded risk concept of the most recently identified PARK loci (PARK 17-23) and some susceptibility variants such as GBA, LRRK2 and SNCA. Finally, the emerging concept of rare genetic variants as candidates for PD, such as HSPA9, TRAP1 and RHOT1 complete the picture of the complex genetic architecture of PD that will direct future precision medicine approaches.

This paper introduces a new lifting method for analyzing convergence of continuous-time distributed synchronization/consensus systems on the unit sphere. Points on the d-dimensional unit sphere are lifted to the (d+1)-dimensional Euclidean space. The consensus protocol on the unit sphere is the classical one, where agents move toward weighted averages of their neighbors in their respective tangent planes. Only local and relative state information is used. The directed interaction graph topologies are allowed to switch as a function of time. The dynamics of the lifted variables are governed by a nonlinear consensus protocol for which the weights contain ratios of the norms of state variables. We generalize previous convergence results for hemispheres. For a large class of consensus protocols defined for switching uniformly quasi-strongly connected time-varying graphs, we show that the consensus manifold is uniformly asymptotically stable relative to closed balls contained in a hemisphere. Compared to earlier projection based approaches used in this context such as the gnomonic projection, which is defined for hemispheres only, the lifting method applies globally. With that, the hope is that this method can be useful for future investigations on global convergence.

The human gastrointestinal tract is home for trillions of bacteria that influence homeostasis and health in a complex biological system: the gut microbiome. Accumulating evidence suggests that a state of pathological imbalance in the microbiome (dysbiosis) is present in patients suffering from colorectal cancer (CRC). To date, microbiome studies identified specific bacteria being associated with dysbiosis in CRC. Some of these bacteria (e.g. Fusobacteria) directly or indirectly interact with cancer and immune cells of their host. However, current studies only focused on certain microbes in detail, hence, their role in the etiology of the disease remains elusive. Accordingly, my project investigates the role of CRC-associated bacteria in tumor initiation and progression while addressing the question: which and what kind of microbes interact with, favor, or can cause CRC? In a first step, we identified CRC-associated bacteria, enriched at the tumor site of Luxembourgish CRC patients. By using Fusobacterium nucleatum as our study model, we predicted and optimized bacterial growth (media) in silico by using a genome-scale metabolic reconstruction model for a constraint-based modelling approach. Next, we assessed bacterial growth and metabolism in the optimized growth medium by using flow cytometry and mass spectrometry. Finally, we co-cultured the bacteria together with primary patient-derived cultures in the recently developed, microfluidics-based, human-microbial cross-talk model (HuMiX) [1]. As part of our ongoing validations, we infected patient-derived, healthy and cancerous 3D colonic organoids with our bacterial candidate. This workflow enables us to analyze pro-tumorigenic capacities of CRC-associated bacteria on healthy and cancerous colonocytes. It will serve as a promising tool for future analysis of host-microbial interaction mechanisms of various CRC-associated bacteria on a transcriptomic, proteomic, and metabolomic level. [1] Shah P, Fritz JV, Glaab E, De

Objective: The purpose of this study was to profile cerebrospinal fluid (CSF) from early-stage PD patients for disease-related metabolic changes and to determine a robust biomarker signature for early-stage PD diagnosis. Methods: By applying a non-targeted and mass spectrometry-driven approach, we investigated the CSF metabolome of 44 early-stage sporadic PD patients yet without treatment (DeNoPa cohort). We compared all detected metabolite levels with those measured in CSF of 43 age- and gender-matched healthy controls. After this analysis, we validated the results in an independent PD study cohort (T€ubingen cohort). Results: We identified that dehydroascorbic acid levels were significantly lower and fructose, mannose, and threonic acid levels were significantly higher (P <.05) in PD patients when compared with healthy controls. These changes reflect pathological oxidative stress responses, as well as protein glycation/glycosylation reactions in PD. Using a machine learning approach based on logistic regression, we successfully predicted the origin (PD patients vs healthy controls) in a second (n518) as well as in a third and completely independent validation set (n536). The biomarker signature is composed of the three markers—mannose, threonic acid, and fructose—and allows for sample classification with a sensitivity of 0.790 and a specificity of 0.800. Conclusion: We identified PD-specific metabolic changes in CSF that were associated with antioxidative stress response, glycation, and inflammation. Our results disentangle the complexity of the CSF metabolome to unravel metabolome changes related to earlystage PD. The detected biomarkers help understanding PD pathogenesis and can be applied as biomarkers to increase clinical diagnosis accuracy and patient care in early-stage PD.

Metabolomic markers associated with incident central adiposity gain were investigated in young adults. In a 9-mo prospective study of university freshmen (n = 264). Blood samples and anthropometry measurements were collected in the first 3 d on campus and at the end of the year. Plasma from individuals was pooled by phenotype [incident central adiposity, stable adiposity, baseline hemoglobin A1c (HbA1c) > 5.05%, HbA1c < 4.92%] and assayed using GC-MS, chromatograms were analyzed using MetaboliteDetector software, and normalized metabolite levels were compared using Welch’s t test. Assays were repeated using freshly prepared pools, and statistically significant metabolites were quantified in a targeted GC-MS approach. Isotope tracer studies were performed to determine if the potential marker was an endogenous human metabolite in men and in whole blood. Participants with incident central adiposity gain had statistically significantly higher blood erythritol [P < 0.001, false discovery rate (FDR) = 0.0435], and the targeted assay revealed 15-fold [95% confidence interval (CI): 13.27, 16.25] higher blood erythritol compared with participants with stable adiposity. Participants with baseline HbA1c > 5.05% had 21-fold (95% CI: 19.84, 21.41) higher blood erythritol compared with participants with lower HbA1c (P < 0.001, FDR = 0.00016). Erythritol was shown to be synthesized endogenously from glucose via the pentose-phosphate pathway (PPP) in stable isotope-assisted ex vivo blood incubation experiments and through in vivo conversion of erythritol to erythronate in stable isotope-assisted dried blood spot experiments. Therefore, endogenous production of erythritol from glucose may contribute to the association between erythritol and obesity observed in young adults.

Metabolome analyses of body fluids are challenging due pre-analytical variations, such as pre-processing delay and temperature, and constant dynamical changes of biochemical processes within the samples. Therefore, proper sample handling starting from the time of collection up to the analysis is crucial to obtain high quality samples and reproducible results. A metabolomics analysis is divided into 4 main steps: 1) Sample collection, 2) Metabolite extraction, 3) Data acquisition and 4) Data analysis. Here, we describe a protocol for gas chromatography coupled to mass spectrometry (GC–MS) based metabolic analysis for biological matrices, especially body fluids. This protocol can be applied on blood serum/plasma, saliva and cerebrospinal fluid (CSF) samples of humans and other vertebrates. It covers sample collection, sample pre-processing, metabolite extraction, GC–MS measurement and guidelines for the subsequent data analysis.

The “dark matter” in metabolomics (unknowns) represents an exciting frontier with significant potential for discovery in relation to biochemistry, yet it also presents one of the largest challenges to overcome. This focussed review takes a close look at the current state-of-the-art and future challenges in tackling the unknowns with specific focus on the human gut microbiome and host-microbe interactions. Metabolomics, like metabolism itself, is a very dynamic discipline, with many workflows and methods under development, both in terms of chemical analysis and post-analysis data processing. Here, we look at developments in the mutli-omic analyses and the use of mass spectrometry to investigate the exchange of metabolites between the host and the microbiome as well as the environment within the microbiome. A case study using HuMiX, a microfluidics-based human-microbial co-culture system that enables the co-culture of human and microbial cells under controlled conditions, is used to highlight opportunities and current limitations. Common definitions, approaches, databases and elucidation techniques from both the environmental and metabolomics fields are covered, with perspectives on how to merge these, as the boundaries blur between the fields. While reflecting on the number of unknowns remaining to be conquered in typical complexsamples measured with mass spectrometry (often ordersof magnitude above the “knowns”), we provide an outlook on future perspectives and challenges in elucidating the relevant “dark matter”.

Time-series of high throughput gene sequencing data intended for gene regulatory network (GRN) inference are often short due to the high costs of sampling cell systems. Moreover, experimentalists lack a set of quantitative guidelines that prescribe the minimal number of samples required to infer a reliable GRN model. We study the temporal resolution of data vs.quality of GRN inference in order to ultimately overcome this deficit. The evolution of a Markovian jump process model for the Ras/cAMP/PKA pathway of proteins and metabolites in the G1 phase of the Saccharomyces cerevisiae cell cycle is sampled at a number of different rates. For each time-series we infer a linear regression model of the GRN using the LASSO method. The inferred network topology is evaluated in terms of the area under the precision-recall curve (AUPR). By plotting the AUPR against the number of samples, we show that the trade-off has a, roughly speaking, sigmoid shape. An optimal number of samples corresponds to values on the ridge of the sigmoid.

With technological advances in culture-independent molecular methods, we are uncovering a new facet of our natural history by accounting for the vast diversity of microbial life which colonizes the human body. The human microbiome contributes functional genes and metabolites which affect human physiology and are, therefore, considered an important factor for maintaining health. Much has been described in the past decade based primarily on 16S rRNA gene amplicon sequencing regarding the diversity, structure, stability and dynamics of human microbiota in their various body habitats, most notably within the gastrointestinal tract (GIT). Relatively high levels of variation have been described across different stages of life and geographical locations for the GIT microbiome. These observations may prove helpful for the future contextualization of patterns in other body habitats especially in relation to identifying generalizable trends over human lifetime. Given the large degree of complexity and variability, a key challenge will be how to define baseline healthy microbiomes and how to identify features which reflect deviations therefrom in the future. In this context, metagenomics and functional omics will likely play a central role as they will allow resolution of microbiome-conferred functionalities associated with health. Such information will be vital for formulating therapeutic interventions aimed at managing microbiota-mediated health particularly in the GIT over the course of a human lifetime.

In metabolomics-based biomarker studies, the monitoring of pre-analytical variations is crucial and requires quality control tools to enable proper sample quality evaluation. In this dissertation work, biospecimen research and machine learning algorithms are applied (1) to develop sample quality assessment tools and (2) to develop disease-specific diagnostic models. In this regard, a novel plasma sample quality assessment tool, the LacaScore, is presented. The LacaScore plasma quality assessment is based on the plasma levels of ascorbic acid and lactic acid. The biggest challenge in metabolomics analyses is that the sample quality is often not known. The presented tool enhances the knowledge and importance of the monitoring of pre-analytical variations, such as pre-centrifugation time and temperature, prior to sample analysis in the emerging field of metabolomics. Based on the LacaScore, decisions on the suitability/fit-for-purpose of a given sample or sample cohort can be made. In this dissertation work, the knowledge on sample quality was applied in a biomarker discovery study based on cerebrospinal fluid (CSF) from early-stage Parkinson’s disease (PD) patients. To date, no markers for the diagnosis of Parkinson’s disease are available. In this work, a non-targeted GC-MS approach is presented and shows significant changes in the metabolic profile in CSF from early-stage PD patients compared to matched healthy control subjects. Based on these findings, a biomarker signature for the prediction of earlystage PD has been developed by the application of sophisticated machine learning algorithms. This disease-specific signature is composed of metabolites involved in inflammation, glycosylation/glycation and oxidative stress response. In summary, this dissertation illustrates the importance of sample quality monitoring in biomarker studies that are often limited by small amounts of human body fluids. The monitoring of sample quality enhances the robustness and reproducib

Metabolic enzymes are very specific. However, most of them show weak side activities toward compounds that are structurally related to their physiological substrates, thereby producing side products that may be toxic. In some cases, ‘metabolite repair enzymes’ eliminating side products have been identified. We show that mammalian glyceraldehyde 3-phosphate dehydrogenase and pyruvate kinase, two core glycolytic enzymes, produce 4-phosphoerythronate and 2-phospho-L-lactate, respectively. 4-Phosphoerythronate strongly inhibits an enzyme of the pentose phosphate pathway, whereas 2-phospho-L-lactate inhibits the enzyme producing the glycolytic activator fructose 2,6-bisphosphate. We discovered that a single, widely conserved enzyme, known as phosphoglycolate phosphatase (PGP) in mammals, dephosphorylates both 4-phosphoerythronate and 2-phospho-L-lactate, thereby preventing a block in the pentose phosphate pathway and glycolysis. Its yeast ortholog, Pho13, similarly dephosphorylates 4-phosphoerythronate and 2-phosphoglycolate, a side product of pyruvate kinase. Our work illustrates how metabolite repair enzymes can make up for the limited specificity of metabolic enzymes and permit high flux in central metabolic pathways.

Genome-scale metabolic models derived from human gut metagenomic data can be used as a framework to elucidate how microbial communities modulate human metabolism and health. We present AGORA (assembly of gut organisms through reconstruction and analysis), a resource of genome-scale metabolic reconstructions semi-automatically generated for 773 human gut bacteria. Using this resource, we identified a defined growth medium for Bacteroides caccae ATCC 34185. We also showed that interactions among modeled species depend on both the metabolic potential of each species and the nutrients available. AGORA reconstructions can integrate either metagenomic or 16S rRNA sequencing data sets to infer the metabolic diversity of microbial communities. AGORA reconstructions could provide a starting point for the generation of high-quality, manually curated metabolic reconstructions. AGORA is fully compatible with Recon 2, a comprehensive metabolic reconstruction of human metabolism, which will facilitate studies of host–microbiome interactions.

Aging-related biomolecular changes in the human brain are thought to be associated with an increased risk for neurodegenerative diseases. In particular, aging and Parkinson’s disease (PD) share various molecular hallmarks, including a gradual decline in dopamine synthesis and increased levels of deleted mitochondrial DNA. While some specific mechanistic links between brain aging and PD have been proposed and investigated previously, systematic analyses of shared molecular alterations at a genome-scale level are required to obtain a better understanding of the affected cellular processes and their interrelations. We present a joint analysis of high-throughput brain transcriptomics data from PD patients and unaffected individuals from different adult age groups using a statistical meta-analysis and a recently published pathway and network analysis approach. Our analyses provide statistical evidence for specific functional associations between molecular network changes in PD and aging, identify new significant joint pathway deregulations and suggest mechanistic explanations for the observed age-dependence of PD risk.

The dependence-testing (DT) bound is one of the strongest achievability bounds for the binary erasure channel (BEC) in the finite block length regime. In this paper, we show that maximum likelihood decoded regular low-density parity-check (LDPC) codes with at least 5 ones per column almost achieve the DT bound. Specifically, using quasi-regular LDPC codes with block length of 256 bits, we achieve a rate that is less than 1% away from the rate predicted by the DT bound for a word error rate below 10^−3. The results also indicate that the maximum-likelihood solution is computationally feasible for decoding block codes over the BEC with several hundred bits.