Universität Stuttgart
Permanent URI for this communityhttps://elib.uni-stuttgart.de/handle/11682/1
Browse
4232 results
Search Results
Item Open Access ROSIE : RObust Sparse ensemble for outlIEr detection and gene selection in cancer omics data(2022) Jensch, Antje; Lopes, Marta B.; Vinga, Susana; Radde, NicoleThe extraction of novel information from omics data is a challenging task, in particular, since the number of features (e.g. genes) often far exceeds the number of samples. In such a setting, conventional parameter estimation leads to ill-posed optimization problems, and regularization may be required. In addition, outliers can largely impact classification accuracy. Here we introduce ROSIE, an ensemble classification approach, which combines three sparse and robust classification methods for outlier detection and feature selection and further performs a bootstrap-based validity check. Outliers of ROSIE are determined by the rank product test using outlier rankings of all three methods, and important features are selected as features commonly selected by all methods. We apply ROSIE to RNA-Seq data from The Cancer Genome Atlas (TCGA) to classify observations into Triple-Negative Breast Cancer (TNBC) and non-TNBC tissue samples. The pre-processed dataset consists of 16,600 genes and more than 1,000 samples. We demonstrate that ROSIE selects important features and outliers in a robust way. Identified outliers are concordant with the distribution of the commonly selected genes by the three methods, and results are in line with other independent studies. Furthermore, we discuss the association of some of the selected genes with the TNBC subtype in other investigations. In summary, ROSIE constitutes a robust and sparse procedure to identify outliers and important genes through binary classification. Our approach is ad hoc applicable to other datasets, fulfilling the overall goal of simultaneously identifying outliers and candidate disease biomarkers to the targeted in therapy research and personalized medicine frameworks.Item Open Access Heat transport from atmosphere through the subsurface to drinking‐water supply pipes(2023) Nissler, Elisabeth; Scherrer, Samuel; Class, Holger; Müller, Tanja; Hermannspan, Mark; Osmancevic, Esad; Haslauer, ClausDrinking‐water quality in supply pipe networks can be negatively affected by high temperatures during hot summer months due to detrimental bacteria encountering ideal conditions for growth. Thus, water suppliers are interested in estimating the temperature in their distribution networks. We investigate both experimentally and by numerical simulation the heat and water transport from ground surface into the subsurface, (i.e., above drinking‐water pipes). We consider the meteorological forcing functions by a sophisticated approach to model the boundary conditions for the heat balance at the soil-atmosphere interface. From August to December 2020, soil temperatures and soil moisture were measured dependent on soil type, land‐use cover, and weather data at a pilot site, constructed specifically for this purpose at the University of Stuttgart with polyethylene and cast‐iron pipes installed under typical in situ conditions. We included this interface condition at the atmosphere-subsurface boundary into an integrated non‐isothermal, variably saturated (Richards') the numerical simulator DuMux 3. This allowed, after calibration, to match measured soil temperatures with ±2°C accuracy. The land‐use cover influenced the soil temperature in 1.5 m more than the soil material used for back‐filling the trench above the pipe.Item Open Access Performance measurements for personalizable route planning for uncorrelated edge costs(2021) Bühler, FelixNowadays, ordinary route planners compute paths by choosing the shortest or fastest route. However, there exist additional metrics from which users with varying preferences could benefit. Personalized route planning offers the possibility to combine different metrics with personal preferences. Nevertheless, personalized route planning has mainly been tested with correlated metrics. But when including uncorrelated metrics, the computing time increases significantly. Previous work found that the speedup technique “Customizable Route Planning” can lead to feasible speedups for single metric calculations. Thus, in this work, we investigate how this speedup technique for Dijkstra improves the query performances of “Personalizable Route Planning” compared to “Personalizable Contraction Hierarchies”. Furthermore, we study the performances on uncorrelated metrics. We introduce a graph structure to compare the personalized speedup techniques “Personalizable Contraction Hierarchies”, “Personalizable Customizable Route Planning” and “Personalizable Route Planning”. Three graph partitioning algorithms have been implemented to realize “Customizable Route Planning”: K-means, Gonzales, and Merge. Our experiments show that Merge works well in combination with “Personalizable Contraction Hierarchies” preprocessing. We found that “Personalizable Customizable Route Planning” is a good alternative, as it uses much fewer edges for finding the costs of the shortest path. For uncorrelated metrics, “Personalizable Customizable Route Planning” and “Personalizable Route Planning” achieved speedups higher than “Personalizable Contraction Hierarchies”. Our contribution comprises a novel graph structure for comparing different Dijkstra variants. With our experiments, we provide a deeper understanding of the personalized route planning problem. Additionally, we propose improvements for “Personalizable Contraction Hierarchies” for less contracted graphs with uncorrelated metrics.Item Open Access Towards improved targetless registration and deformation analysis of TLS point clouds using patch-based segmentation(2023) Yang, Yihui; Schwieger, Volker (Prof. Dr.-Ing. habil. Dr. h.c.)The geometric changes in the real world can be captured by measuring and comparing the 3D coordinates of object surfaces. Traditional point-wise measurements with low spatial resolution may fail to detect inhomogeneous, anisotropic and unexpected deformations, and thus cannot reveal complex deformation processes. 3D point clouds generated from laser scanning or photogrammetric techniques have opened up opportunities for an area-wise acquisition of spatial information. In particular, terrestrial laser scanning (TLS) exhibits rapid development and wide application in areal geodetic monitoring owing to the high resolution and high quality of acquired point cloud data. However, several issues in the process chain of TLS-based deformation monitoring are still not solved satisfactorily. This thesis mainly focuses on the targetless registration and deformation analysis of TLS point clouds, aiming to develop novel data-driven methods to tackle the current challenges. For most deformation processes of natural scenes, in some local areas no shape deformations occur (i.e., these areas are rigid), and even the deformation directions show a certain level of consistency when these areas are small enough. Further point cloud processing, like stability and deformation analyses, could benefit from the assumptions of local rigidity and consistency of deformed point clouds. In this thesis, thereby, three typical types of locally rigid patches - small planar patches, geometric primitives, and quasi-rigid areas - can be generated from 3D point clouds by specific segmentation techniques. These patches, on the one hand, can preserve the boundaries between rigid and non-rigid areas and thus enable spatial separation with respect to surface stability. On the other hand, local geometric information and empirical stochastic models could be readily determined by the points in each patch. Based on these segmented rigid patches, targetless registration and deformation analysis of deformed TLS point clouds can be improved regarding accuracy and spatial resolution. Specifically, small planar patches like supervoxels are utilized to distinguish the stable and unstable areas in an iterative registration process, thus ensuring only relatively stable points are involved in estimating transformation parameters. The experimental results show that the proposed targetless registration method has significantly improved the registration accuracy. These small planar patches are also exploited to develop a novel variant of the multiscale model-to-model cloud comparison (M3C2) algorithm, which constructs prisms extending from planar patches instead of the cylinders in standard M3C2. This new method separates actual surface variations and measurement uncertainties, thus yielding lower-uncertainty and higher-resolution deformations. A coarse-to-fine segmentation framework is used to extract multiple geometric primitives from point clouds, and rigorous parameter estimations are performed individually to derive high-precision parametric deformations. Besides, a generalized local registration-based pipeline is proposed to derive dense displacement vectors based on segmented quasi-rigid areas that are corresponded by areal geometric feature descriptors. All proposed methods are successfully verified and evaluated by simulated and/or real point cloud data. The choice of proposed deformation analysis methods for specific scenarios or applications is also provided in this thesis.Item Open Access Alumina and zirconia-reinforced polyamide PA-12 composites for biomedical additive manufacturing(2021) Nakonieczny, Damian S.; Kern, Frank; Dufner, Lukas; Antonowicz, Magdalena; Matus, KrzysztofThis work aimed to prepare a composite with a polyamide (PA) matrix and surface-modified ZrO2 or Al2O3 to be used as ceramic fillers (CFs). Those composites contained 30 wt.% ceramic powder to 70 wt.% polymer. Possible applications for this type of composite include bioengineering applications especially in the fields of dental prosthetics and orthopaedics. The ceramic fillers were subjected to chemical surface modification with Piranha Solution and suspension in 10 M sodium hydroxide and Si3N4 to achieve the highest possible surface development and to introduce additional functional groups. This was to improve the bonding between the CFs and the polymer matrix. Both CFs were examined for particle size distribution (PSD), functional groups (FTIR), chemical composition (XPS), phase composition (XRD), and morphology and chemical composition (SEM/EDS). Filaments were created from the powders prepared in this way and were then used for 3D FDM printing. Samples were subjected to mechanical tests (tensility, hardness) and soaking tests in a high-pressure autoclave in artificial saliva for 14, 21, and 29 days.Item Open Access Designing covalent organic framework‐based light‐driven microswimmers toward therapeutic applications(2023) Sridhar, Varun; Yildiz, Erdost; Rodríguez‐Camargo, Andrés; Lyu, Xianglong; Yao, Liang; Wrede, Paul; Aghakhani, Amirreza; Akolpoglu, Birgul M.; Podjaski, Filip; Lotsch, Bettina V.; Sitti, MetinWhile micromachines with tailored functionalities enable therapeutic applications in biological environments, their controlled motion and targeted drug delivery in biological media require sophisticated designs for practical applications. Covalent organic frameworks (COFs), a new generation of crystalline and nanoporous polymers, offer new perspectives for light‐driven microswimmers in heterogeneous biological environments including intraocular fluids, thus setting the stage for biomedical applications such as retinal drug delivery. Two different types of COFs, uniformly spherical TABP‐PDA‐COF sub‐micrometer particles and texturally nanoporous, micrometer‐sized TpAzo‐COF particles are described and compared as light‐driven microrobots. They can be used as highly efficient visible‐light‐driven drug carriers in aqueous ionic and cellular media. Their absorption ranging down to red light enables phototaxis even in deeper and viscous biological media, while the organic nature of COFs ensures their biocompatibility. Their inherently porous structures with ≈2.6 and ≈3.4 nm pores, and large surface areas allow for targeted and efficient drug loading even for insoluble drugs, which can be released on demand. Additionally, indocyanine green (ICG) dye loading in the pores enables photoacoustic imaging, optical coherence tomography, and hyperthermia in operando conditions. This real‐time visualization of the drug‐loaded COF microswimmers enables unique insights into the action of photoactive porous drug carriers for therapeutic applications.Item Open Access Counterion effects on the mesomorphic and electrochemical properties of guanidinium salts(2024) Ebert, Max; Lange, Alyna; Müller, Michael; Wuckert, Eugen; Gießelmann, Frank; Klamroth, Tillmann; Zens, Anna; Taubert, Andreas; Laschat, SabineIonic liquid crystals (ILCs) combine the ion mobility of ionic liquids with the order and self-assembly of thermotropic mesophases. To understand the role of the anion in ILCs, wedge-shaped arylguanidinium salts with tetradecyloxy side chains were chosen as benchmark systems and their liquid crystalline self-assembly in the bulk phase as well as their electrochemical behavior in solution were studied depending on the anion. Differential scanning calorimetry (DSC), polarizing optical microscopy (POM) and X-ray diffraction (WAXS, SAXS) experiments revealed that for spherical anions, the phase width of the hexagonal columnar mesophase increased with the anion size, while for non-spherical anions, the trends were less clear cut. Depending on the anion, the ILCs showed different stability towards electrochemical oxidation and reduction with the most stable being the PF6 based compound. Cyclic voltammetry (CV) and density functional theory (DFT) calculations suggest a possible contribution of the guanidinium cation to the oxidation processes.Item Open Access Wertschöpfung und Erlösgenerierung etablierter Unternehmen mit digitalen Innovationen : eine explorative Analyse anhand digitaler Services für das vernetzte Automobil(2021) Bosler, Micha; Burr, Wolfgang (Prof. Dr.)Digitale Innovationen implizieren für etablierte Unternehmen, deren Kerngeschäft ehemals auf rein physischen Angeboten basierte, radikale Veränderungen der Wertschöpfung und Erlösgenerierung. Neuartige Kombinationen digitaler und physischer Ressourcen bieten einerseits Innovationspotenziale, andererseits wandeln sich die wirtschaftlichen Rahmenbedingungen fundamental. Da bisherige Veröffentlichungen zu digitalen Innovationen meist auf grundlegenden, konzeptionellen Überlegungen basieren, zeigt sich sowohl hinsichtlich empirischer Untersuchungen als auch theoretischer Erklärungen ein Forschungsdefizit. Daran setzt die vorliegende Arbeit an. Als konkreter Untersuchungsgegenstand dienen digitale Services für vernetzte Fahrzeuge. Mittels einer multiplen Fallstudienanalyse von vier Automobilherstellern trägt die explorative Studie zu einem tieferen Verständnis des Innovationsumfelds, den wertschöpfenden Aktivitäten, der damit intendierten Erlösgenerierung und den auftretenden Herausforderungen bei. Als primäre Datenquelle dienen 23 Experteninterviews. Die fallübergreifend generalisierten Erkenntnisse verdeutlichen, welche vielseitigen Veränderungen und Schwierigkeiten für etablierte Unternehmen entstehen, wenn digitale Innovationen ergänzend zu bestehenden Produkten veröffentlicht werden. Gerade das Kerngeschäft der Automobilhersteller zeichnet sich eigentlich durch sehr lange Entwicklungs- und Produktlebenszyklen aus. Entscheidungen über die Umfänge künftiger Baureihen, einschließlich der technologischen Innovationen, werden traditionell mit mehrjähriger Vorlaufzeit getroffen, um Risiken zu minimieren und zum Produktionsbeginn ein perfektioniertes Fahrzeug bei höchsten Qualitätsstandards zu gewährleisten. Das erweist sich als völlig konträr zum digitalen Geschäft, wo insbesondere die Schnelligkeit zum ausschlaggebenden Kriterium für Wettbewerbsvorteile wird. Es braucht deutlich kürzere Innovationsprozesse, hohe Investitionen, mehr Risikobereitschaft und eine veränderte Denkweise, die sich von ehemals bewährten Praktiken löst. Die erfolgreiche Veröffentlichung digitaler Innovationen geht - neben der langwierigen Transformation der eigenen Ressourcenbasis - mit einer Ausdehnung in Wertschöpfungsnetzwerke sowie der intensiven Integration von Partnern einher. Ohnehin endet das Geschäftsmodell der Hersteller nicht mehr mit dem Verkauf des Fahrzeugs an den Händler. Die Unternehmen agieren erstmals direkt im Endkundengeschäft und erhoffen sich kontinuierliche Einnahmen in Form kostenpflichtiger Services. Angesichts der Möglichkeit der nachträglichen Einführung digitaler Innovationen werden die vernetzten Automobile über ihren gesamten Lebenszyklus veränderbar. Allerdings sind die Automobilhersteller nicht nur mit einem veränderten Produktverständnis, sondern auch mit neuen Anforderungen der Kunden an digitale Dienste konfrontiert, die zunächst überhaupt erkannt und daraufhin berücksichtigt werden müssen. Abschließend richtet sich der Fokus auf die Theoretisierung digitaler Innovationen. Ausgehend von den empirisch identifizierten Zusammenhängen werden Eignung und Anpassungsbedarf ausgewählter Ansätze - wie Dynamic Capabilities, der Practice-based View oder das Dominante Design - diskutiert.Item Open Access Construction of robust Escherichia coli strains for large-scale production(2022) Ziegler, Martin; Takors, Ralf (Prof. Dr.-Ing.)The biotechnical production of many fine chemicals, proteins or pharmaceuticals depends on large-scale microbial cultivations. Due to limited mixing, heterogeneities in process relevant parameters such as nutrient concentrations arise in such fermentations. Escherichia coli (E. coli) is a model organism frequently used in the biotechnological industry. If E. coli is cultivated under heterogeneous conditions, biological reactions of the microorganism result in reduced process performance. Since large-scale fermentations are not economically feasible in academic settings, scale-down reactors that mimic aforementioned heterogeneities are used to investigate heterogenous fermentations. Previous studies in scale-down reactors unraveled that, depending on the process strategy, the unstable supply of a limiting primary carbon or nitrogen source such as glucose or ammonium is one of the underlying causes of process performance loss. Low concentrations of glucose or ammonium elicit the stringent response as a biological starvation reaction which comprises extensive transcriptional reactions. In the first project that contributes to this thesis, the regulatory and transcriptional reactions of the strains E. coli MG1655 and E. coli SR to repeated exposure to ammonium starvation zones were examined in a scale-down reactor. The scale-down reactor followed a two-compartment approach and consisted of a stirred tank reactor and a plug-flow reactor simulating passage through a starvation zone. E. coli SR is a strain with modulated stringent response. It was observed that short-term starvation stimuli do not trigger this regulatory program in E. coli SR and the transcriptional reaction was noticeably reduced. Long-term adaptation of the strain to repeated cycles of limitation and starvation also clearly differed from E. coli MG1655. Despite lack of the stringent response, E. coli SR showed no deficits in the assimilation of the limiting ammonium or in biomass yield on ammonium. In the second project of this thesis, a series of deletion strains with robust phenotype against glucose starvation zones were constructed. Candidate genes were identified and successively removed from the genome of E. coli MG1655 by Recombineering. The fundamental growth parameters of the strains were determined in shaking flask fermentations and no noticeable differences compared to E. coli MG1655 were found. Chemostat cultivations in a scale-down reactor with glucose as the limiting nutrient source revealed that the final strain of the deletion series, E. coli RM214, had a significantly lower maintenance coefficient under heterogeneous conditions than E. coli MG1655. Moreover, in an exemplary heterologous protein productionscenario E. coli RM214 rhaB- pJOE4056.2_tetA proved to be more robust to heterogeneities and showed a significantly higher product yield than E. coli MG1655 rhaB- pJOE4056.2_tetA. In the third project of this thesis, the production of pyruvate in E. coli MG1655 by inhibition of pyruvate dehydrogenase through CRISPR interference was investigated. A central goal was to achieve the stable production in nitrogen-limited conditions. For this, different target sequences in the operon pdhR-aceEF-lpd were tested and the strains cultivated in shaking flask fermentations. All tested target sequences were generally suitable to trigger the accumulation of pyruvate. Combined CRISPR interference against two target sequences did not lead to an increased pyruvate yield in most cases. In addition, the strains E. coli MG1655 pdCas9 psgRNA_aceE_234 and E. coli MG1655 pdCas9 psgRNA_aceE_234_pdhR_329 were characterized in two phase fermentations in lab-scale reactors. The initial phase was an unlimited exponential growth phase and was followed by an ammonium-limited production phase. E. coli MG1655 pdCas9 psgRNA_aceE_234 only produced pyruvate during the exponential phase, and reuptake of pyruvate occurred in the second phase. In contrast, E. coli MG1655 pdCas9 psgRNA_aceE_234_pdhR_329 stably produced pyruvate during the exponential and the ammonium-limited phase and is a potential chassis strain for the growth-decoupled production of pyruvate derived bioproducts. The overarching research issues of the projects were the characterization of strains in heterogeneous conditions and the development of new strategies to improve their performance. The collected data leads me to conclude that the construction of robust microbial strains for large-scale applications is both expedient and feasible. Tailored genetic modifications are the method of choice to achieve this goal. Furthermore, suitable genetic constructs offer promising possibilities for the stable growth-decoupled production of chemicals in nitrogen-limited conditions.Item Open Access High-resolution spatio-temporal measurements of the colmation phenomenon under laboratory conditions(Stuttgart : Eigenverlag des Instituts für Wasser- und Umweltsystemmodellierung der Universität Stuttgart, 2022) Mayar, Mohammad Assem; Wieprecht, Silke (Prof. Dr.-Ing.)The fine sediment infiltration and accumulation into the gravel bed of rivers, the so-called colmation phenomenon, is a pernicious process exacerbated by anthropogenic activities. Owing to the importance and complexity of this phenomenon, it has been widely studied over the last decades. Various devices and methods have been developed to assess this phenomenon, where most of them are destructive and sample-based, resulting in an alteration of the natural conditions. Therefore, non-intrusive techniques, which provide spatial and temporal details with a high-resolution, are required to discretize the mechanisms involved in the colmation process. To address these issues, investigations under laboratory conditions may simplify the complexity of nature and enable individual and exactly defined boundary conditions to be investigated. Therefore, this thesis aims at (i) developing a non-intrusive and undisturbed measurement method for the high-resolution spatio-temporal measurements of the sediment infiltration processes and the development of sediment accumulation in an artificial river bed under laboratory conditions, (ii) applying this method to certain experiments for the assessment of the effects of different boundary conditions on sediment infiltration, and (iii) investigating the colmation phenomenon (also known as clogging) of gravel beds. For this purpose, the gamma-ray attenuation method is used together with an artificial gravel bed arranged from the spheres with various diameters and placed in a laboratory flume. This new method works based on the gamma radiation that passes through the infiltrated sediments, water, and bed spheres, in which the gamma-ray attenuation is linked to the variations of the infiltrated sediments’ quantity. The main simplification of this approach is that gravel beds are represented by the combinations of different-sized spheres. This gives the opportunity to fully distinguish infiltrating sediments from the bed material, reduce the complexity of the natural environment, and allows for repetitive measurements of the same position with different boundary conditions. From the results of this study, first, the gamma-ray attenuation measurement method was optimized to resolve the inconsistencies in the measurements. Subsequently, the concept of the non-intrusive and undisturbed measurement is proved through box experiments. Additional reproducibility experiments in the laboratory flume, for a similar bed structure, showed only small deviations between two experiments with the same setup. Consequently, the established technique was used in a series of experiments to evaluate the effects of different supply rates, total supply masses, and sediment particle size boundary conditions on the sediment infiltration and colmation processes. Vertical profiles of the infiltrated sediment were quantified through high spatial resolution measurements. Furthermore, to evaluate the infiltrating sediment accumulation development, and the temporal variations of the infiltrated sediments, the vertical profile measurements were first repeated after a specific time-period to track interval-averaged variations in all positions of the vertical axis. Next, a specific position of the vertical axis was measured continuously during the entire experiment in a high temporal resolution. The measured vertical profiles illustrate the vertical distribution, colmation, and unimpeded percolation of the infiltrated sediments. The dynamic one-point measurement precisely identifies the three phases (the start of the pore-filling, the required time to fill the pore, and the final amount of infiltrated sediments including natural fluctuation during the ongoing experiments) of the sediment infiltration or the possible clogging. As a limitation, the gamma-ray attenuation system’s current configuration only works in artificial gravel beds because of the given density difference between infiltrated sediments and the artificial bed structure. Intense radiations that pass through the natural bed's thickness are capable of detecting a significant amount of infiltrated sediments. However, small amounts of infiltrated sediments will create only a minimal shift in attenuation, which might be confused with the statistical error. In addition, the legal restriction against using radioactive material in the natural environment is another reason for not applying it in the field. Furthermore, the gamma-ray attenuation method cannot resolve the sediment distribution in the measurement horizon and provides an integrative result for each measurement position. In addition, if a mixture of silt, clay, and sand is supplied to the experiment, the gamma-ray attenuation system will produce a bulk result of all the infiltrated materials. To conclude, despite the limitations mentioned above, the gamma-ray attenuation method offers a unique opportunity for the non-intrusive and undisturbed measurements of the sediment infiltration or the special case of colmation, with a high spatio-temporal resolution. This method has the potential to quantify the investigated processes on a millimetric spatial scale, if the measurement time is not a constraint, or vice versa, in a high temporal resolution (seconds) for a specific position, if spatial scale is not important. Moreover, the gamma-ray attenuation approach can simultaneously measure the longitudinal distribution of the sedimentological processes, if multiple instruments or a single device with several radiation-emitting-holes is in operation. Last, but not least, rather than the spheres, artificial gravel beds could be made of any substance with a composition significantly different from the infiltrating sediments, and the boundary conditions of the experiments can be improved in order to attain conditions close to nature. Finally, the gamma-ray attenuation method can be integrated with advanced flow measurement instruments such as Particle Image Velocimetry (PIV) and other high-resolution endoscopic devices to track the behavior of fine sediment infiltration and its clogging process in the porous gravel beds as it occurs in nature.