Karina Hettwer

Non-targeted methods (NTMs) using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) show promise in bacterial resistance detection, yet temporal variations in spectral features pose significant challenges. These proteomic patterns, which characterize bacterial phenotypes and pathological functions, may vary over time due to bacterial adaptation, virulence, or resistance mechanisms, resulting in large prediction uncertainties and potentially degrading NTM performance. We present a comprehensive screening method to detect temporal changes in MALDI-TOF spectral patterns, demonstrated using methicillin-resistant and -susceptible Staphylococcus aureus (MRSA/MSSA) isolates collected over several years. Our approach combines convolutional neural networks (CNNs) with statistical methods, including significance testing, kernel density estimation, and receiver operating characteristics for dataset shift detection. We employ Gradient-weighted Class Activation Mapping (Grad-CAM) for post hoc feature description, enabling biochemical characterization of temporal changes. This analysis reveals crucial insights into the dynamic relationship between spectral data patterns over time, addressing key challenges in developing robust NTMs for routine applications.

Non-targeted methods (NTMs) require robust methods for comparing spectral data for reliable classification and identification. Traditional approaches using match factors reduce complex spectral relationships to single values, limiting their utility in quality assurance. This study presents an evaluation of spectral comparison methodologies, contrasting classical Mahalanobis distance (MD) with neural network approaches, namely, neural classification distance (NCD). Using matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry data from bacterial isolates, we systematically assessed these methods across varying levels of spectral variability. The MD approach exhibited consistent performance under controlled conditions but showed limitations with increasing spectral complexity. In contrast, the NCD demonstrated adaptability across all scenarios, revealing its capability in handling complex spectral relationships. Through this exemplary example, we present the mathematical framework for quantifying spectral variations and establish criteria for method selection in different analytical scenarios. This work provides a foundation for proposing data quality metrics in NTMs and offers practical implementations for routine quality assurance. The methodology developed here extends beyond mass spectrometry applications and contributes to the broader field of analytical quality control in complex spectral analysis.

This study investigates the potential of data-driven modeling to monitor and predict mechanical properties of extruded film using recyclates. The first step is to investigate whether a process parameter can indicate the input material quality of the material, which can vary significantly due to the use of recyclate. The die pressure was shown to be a key indicator due to its strong correlation with viscosity and material degradation. The second step is to explore the ability of machine learning models—Generalized Additive Models, Linear Regression, and Random Forest—to predict film tensile strength and modulus based on extrusion process and material parameters. The results demonstrate that including Melt Flow Rate and shear viscosity in addition to pressure and feedstock type (Virgin, PIR, and PCR) significantly improves model accuracy, with Generalized Additive Models achieving the highest R² of 85.7% for tensile strength prediction. Additionally, the observed variability between different recyclate streams highlights the need for a more detailed classification of recyclates to better predict and optimize the mechanical performance of the film. While data-driven approaches for predicting properties show promise, their effectiveness remains limited by data availability and feedstock variability. Expanding datasets and improving process stability will be critical to refining predictive models for industrial application.

Food fraud, even when not in the news, is ubiquitous and demands the development of innovative strategies to combat it. A new non-targeted method (NTM) for distinguishing spelt and wheat is described, which aids in food fraud detection and authenticity testing. A highly resolved fingerprint in the form of spectra is obtained for several cultivars of spelt and wheat using liquid chromatography coupled high-resolution mass spectrometry (LC-HRMS). Convolutional neural network (CNN) models are built using a nested cross validation (NCV) approach by appropriately training them using a calibration set comprising duplicate measurements of eleven cultivars of wheat and spelt, each. The results reveal that the CNNs automatically learn patterns and representations to best discriminate tested samples into spelt or wheat. This is further investigated using an external validation set comprising artificially mixed spectra, samples for processed goods (spelt bread and flour), eleven untypical spelt, and six old wheat cultivars. These cultivars were not part of model building. We introduce a metric called the D score to quantitatively evaluate and compare the classification decisions. Our results demonstrate that NTMs based on NCV and CNNs trained using appropriately chosen spectral data can be reliable enough to be used on a wider range of cultivars and their mixes.

Background: Despite the significance of colonoscopy for early diagnosis of colorectal adenocarcinoma (CRC), population-wide screening remains challenging, mainly because of low acceptance rates. Herein, exosomal (exo-miR) and free circulating microRNA (c-miR) may be used as liquid biopsies in CRC to identify individuals at risk. Direct comparison of both compartments has shown inconclusive results, which is why we directly compared a panel of 10 microRNAs in this entity. Methods: Exo-miR and c-miR levels were measured using real-time quantitative PCR after isolation from serum specimens in a cohort of 69 patients. Furthermore, results were compared to established tumor markers CEA and CA 19-9. Results: Direct comparison of exo- and c-miR biopsy results showed significantly higher microRNA levels in the exosomal compartment (p < 0.001). Exo-Let7, exo-miR-16 and exo-miR-23 significantly differed between CRC and healthy controls (all p < 0.05), while no c-miR showed this potential. Sensitivity and specificity can be further enhanced using combinations of multiple exosomal miRNAs. Conclusions: Exosomal microRNA should be considered as a promising biomarker in CRC for future studies. Nonetheless, results may show interference with common comorbidities, which must be taken into account in future studies.

Adverse effects of drug combinations and their underlying mechanisms are highly relevant for safety evaluation, but often not fully studied. Hydroxychloroquine (HCQ) and azithromycin (AZM) were used as a combination therapy in the treatment of COVID-19 patients at the beginning of the pandemic, leading to higher complication rates in comparison to respective monotherapies. Here, we used human-induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) to systematically investigate the effects of HCQ, AZM, and their combination on the structure and functionality of cardiomyocytes, and to better understand the underlying mechanisms. Our results demonstrate synergistic adverse effects of AZM and HCQ on electrophysiological and contractile function of iPSC-CMs. HCQ-induced prolongation of field potential duration (FPDc) was gradually increased during 7-day treatment period and was strongly enhanced by combination with AZM, although AZM alone slightly shortened FPDc in iPSC-CMs. Combined treatment with AZM and HCQ leads to higher cardiotoxicity, more severe structural disarrangement, more pronounced contractile dysfunctions, and more elevated conduction velocity, compared to respective monotreatments. Mechanistic insights underlying the synergistic effects of AZM and HCQ on iPSC-CM functionality are provided based on increased cellular accumulation of HCQ and AZM as well as increased Cx43- and Nav1.5-protein levels.

Detection of food fraud and geographical traceability of ingredients is a continually sought goal for government institutions, producers, and consumers. Herein we explore the use of non-target high-resolution mass spectrometry approaches and demonstrate its utility through a particularly challenging case study – to distinguish wheat and spelt cultivars. By employing a data-independent acquisition (DIA) approach for sample measurement, the spectra are of considerable size and complexity. We utilize artificial intelligence (AI) algorithms (artificial neural networks) to evaluate the extensive proteomic footprint of several wheat and spelt cultivars. The AI model thus obtained is used to classify newer varieties of spelt, processed flour, and bread samples. Additionally, we discuss the validation of such a method coupling DIA and AI approaches. The novel framework for method validation enables calculation of precision parameters for facile comparison of the discriminatory power of the method and in the development of a reliable decision rule.