11163 research outputs found
Sort by
Outcome und Adhärenz moderierende Faktoren einer intensiven Sportintervention in drei psychiatrischen Störungsentitäten: subjektiver Gesundheitszustand, somatoforme Symptomlast, sportliche Vorerfahrung
No full textPsychische Erkrankungen tragen laut der Global Burden of Disease Study 2010 zu 7,4 % der weltweit verlorenen gesunden Lebensjahre bei (Whiteford et al. 2015). Die Identifikation und Erforschung wirksamer Therapieansätze ist daher von hoher Relevanz. Körperliche Aktivität gewinnt zunehmend an Bedeutung in der Behandlung psychischer Erkrankungen und wird bereits in den nationalen Leitlinien zur Depressionsbehandlung explizit empfohlen (Bundesärztekammer et al. 2022). Trotz zahlreicher Studien zur Effektivität von Sportinterventionen besteht ein erheblicher Forschungsbedarf hinsichtlich der Faktoren, die deren Wirksamkeit und Adhärenz beeinflussen.
Diese Studie untersucht, inwiefern der subjektive Gesundheitszustand (gemessen mit dem SF-36), die somatoforme Symptomlast (gemessen mit dem SOMS7-T) sowie die sportliche Vorerfahrung das Therapie-Outcome und die Adhärenz einer sechswöchigen intensiven Sportintervention, zusätzlich zur Standardbehandlung, bei Patienten mit Depressionen (55-75 Jahre), Schizophrenie oder Suchterkrankungen beeinflussen.
Das Studiendesign entsprach einer offenen, unkontrollierten Pilotstudie. Die Pateinten der drei Erkrankungsgruppen wurden innerhalb der ersten Woche ihrer teil- oder vollstationären Behandlung in die Studie eingeschlossen. Die Patienten machten an fünf Tagen die Woche für eine Stunde Sport. Das Sportprogramm beinhaltete Ausdauer- und Krafttraining, Aerobes Training und Yoga.
In die Studie wurden insgesamt 61 Patienten eingeschlossen, darunter 26 mit Depressionen, 22 mit Suchterkrankungen und 13 mit Schizophrenie. Die statistische Analyse erfolgte zur Untersuchung der Effekte der Intervention auf die subjektive psychische Gesundheit und auf die somatoforme Symptomlast, sowie zur Untersuchung eben dieser und zusätzlich der sportlichen Vorerfahrung als Faktoren, die mit der Adhärenz und dem Therapieerfolg assoziiert sind.
Die Intervention führte zu einer signifikanten Verbesserung der subjektiven psychischen Gesundheit über den Interventionszeitraum, ohne signifikante Unterschiede zwischen den drei Diagnosegruppen. Eine hohe somatoforme Symptomlast korrelierte mit einer geringeren globalen Verbesserung der Patienten (gemessen mit dem CGI). Bezüglich der Adhärenz zeigten Patienten mit einer sehr schlechten subjektiven psychischen Gesundheit sowie mit sportlicher Vorerfahrung eine besonders hohe Teilnahmebereitschaft. Im Gegensatz dazu war die Zugehörigkeit zur Suchtgruppe mit einer erhöhten Drop-out-Rate assoziiert.
Schlussfolgerung: Die Ergebnisse legen nahe, dass Patienten mit einer schlechten subjektiven psychischen Gesundheit und sportlicher Vorerfahrung besonders von einer sportlichen Intervention profitieren könnten, während eine hohe somatoforme Symptomlast als potenzieller Prädiktor für einen ausbleibenden Therapieerfolg betrachtet werden sollte. Um diese Zusammenhänge weiter zu validieren, sind zusätzliche randomisierte, kontrollierte Studien erforderlich.According to the Global Burden of Disease Study 2010, mental illnesses account for 7.4% of the globally lost healthy life years (Whiteford et al. 2015). Identifying and researching effective therapeutic approaches is therefore of great importance. Physical activity is gaining increasing significance in the treatment of mental illnesses and is already explicitly recommended in national guidelines for depression treatment (Bundesärztekammer et al. 2022). Despite numerous studies on the effectiveness of sports interventions, there remains a considerable research gap regarding the factors influencing their efficacy and adherence.
This study examines whether subjective health status (measured with the SF-36), somatoform symptom burden (measured with the SOMS7-T), and prior sports experience influence therapy outcomes and adherence to a six-week intensive sports intervention, in addition to standard treatment, in patients with depression (aged 55-75 years), schizophrenia, or substance use disorders.
The study followed an open, uncontrolled pilot design. Patients from the three diagnostic groups were included within the first week of their partial or full inpatient treatment. Participants engaged in physical activity for one hour, five days per week. The exercise program included endurance and strength training, aerobic exercise, and yoga.
A total of 61 patients were included in the study: 26 with depression, 22 with substance use disorders, and 13 with schizophrenia. Statistical analyses were conducted to examine the effects of the intervention on subjective mental health and somatoform symptom burden, as well as to investigate these and prior sports experience as factors associated with adherence and therapy success.
The intervention led to a significant improvement in subjective mental health over the intervention period, with no significant differences between the three diagnostic groups. A high somatoform symptom burden correlated with a lower overall improvement in patients (measured with the CGI). Regarding adherence, patients with very poor subjective mental health and prior sports experience showed particularly high participation rates. In contrast, belonging to the substance use disorder group was associated with a higher dropout rate.
The results suggest that patients with poor subjective mental health and prior sports experience may particularly benefit from a sports intervention, while a high somatoform symptom burden may serve as a potential predictor for a lack of therapeutic success. Further randomized, controlled studies are necessary to validate these findings.10000-01-0
Differentiation of histologic subtypes of renal cell carcinoma using multiphasic CT: A radiomics and machine learning study
No full textThe number of renal cell carcinoma (RCC) diagnoses has increased since the late 1990s, primarily due to incidental findings. Early detection complicates the radiological differentiation of the three most common RCC subtypes—clear cell renal cell carcinoma (ccRCC), papillary renal cell carcinoma (pRCC), and chromophobe renal cell carcinoma (chRCC)—as characteristic imaging features of later stages may be absent. At present, radiologists face diagnostic challenges in categorically distinguishing between these RCC subtypes. Given the distinct prognoses and treatment approaches associated with each subtype, an accurate diagnosis of the histological subtypes is crucial, with the gold standard being histopathological findings. The use of machine learning (ML) and Radiomics is being explored as a potential alternative, as biopsy confirmations are limited by tumor heterogeneity and only a small portion of the tumor tissue can be analyzed in comparison to cross-sectional imaging. However, the findings in current literature are constrained by monophasic or unrealistic imaging as well as small patient cohorts.
The overarching aim of this study was to develop a robust statistical model capable of predicting RCC subtypes based on radiomic features and ML algorithms derived from routine multiphasic computed tomography (CT) scans of the kidneys. The study included patients who underwent surgical resection and histopathological confirmation of their renal tumors at the Universitätsmedizin Göttingen (UMG). A total of 145 patients with RCC were enrolled (64.1% male, mean age 65 ± 12), comprising 112 with clear cell RCC, 23 with papillary RCC, and 10 with chromophobe RCC.
Following preprocessing of the patient cohort (n = 145) using SMOTE (n = 218) and feature selection, 145 radiomic features were generated, and predictions of RCC subtypes were made using four machine learning algorithms (NNET, SVM, KNN, RF). Noteworthy is the comparison of diagnostic accuracy across datasets from different contrast agent phases and the inclusion of multicentrically generated, artifact-laden CT scans with varying slice thicknesses.
Limitations of the study include the retrospective design and the imbalance among RCC subtypes, which resulted in an underrepresentation of potentially characteristic imaging features and thus reduced diagnostic accuracy. When reviewing the diagnostic accuracy, we observed a wide spread of AUCs, with the highest value being 0.76. A neural network achieved the best diagnostic accuracy using data without preprocessing (no SMOTE, no RFE), while data from the SMOTE + RFE cohort resulted in the poorest diagnostic accuracy in 9 out of 12 cases. Concerning the differences in ML performance across various contrast agent phases, the best performance was observed in the corticomedullary and combined CM phases (combined phases AUC 0.76, corticomedullary phase AUC 0.74).
In conclusion, the results of the study demonstrate moderate diagnostic accuracy, which, however, can be considered more clinically robust than previously published studies, particularly in the context of heterogeneous image data and artifacts. The findings underscore that CT imaging in routine clinical practice is suitable for the application of radiomics, but reliable differentiation without pathological diagnosis is currently not feasible with the methods employed. Based on our results, future research should focus on neural networks. Furthermore, ML algorithms demonstrated higher diagnostic accuracy with data lacking RFE or SMOTE, as well as with less preprocessed data (SMOTE cohort, RFE cohort). Overfitting of the datasets, which could not be remedied by cross-validation, may serve as an explanation and could be avoided by equalizing cohort sizes, particularly for the underrepresented chRCC subtype. The similar best performances in the corticomedullary and combined CM phases suggest the relevance of the corticomedullary phase, with no apparent diagnostic benefit from the nephrogenic CM phase. This could be attributed to the differing contrast enhancement behavior, especially the hyperenhancement of ccRCC. Comparable results with larger cohorts and other malignant and benign renal masses are unlikely.
For artificial intelligence to be implemented in routine clinical practice, it is necessary to achieve good diagnostic accuracy on realistic imaging data. Future research should focus on further validation, especially with more balanced class distributions and a comparison of diagnostic accuracy with that of radiologists.2025-05-1
The Functional Role of the Membrane Bending Ability of the Yeast PROPPINs
No full textThe meticulously regulated mechanism, which overseas processes of cellular breakdown and recycling, is called macroautophagy. It has been preserved from yeast to more advanced eukaryotic organisms. As a response to various external and internal stimuli it operates continuously at a low baseline and degrades parts of the cytosol selectively as well as non-selectively. It is initiated with the de novo creation of the so- called isolation membrane or phagophore, a cup-shaped membrane structure, emerging at the pre-autophagosomal structure (PAS). The phagophore expands to subsequently form a closed, double-membrane structure called the autophagosome. Cellular cargo situated at the growing phagophore is completely encased before the structure is sealed off. The autophagosome eventually fuses with the vacuole to release its contents into the vacuolar lumen for degradation and recycling.
The formation of the autophagosome necessitates the production of phosphatidylinositol 3-phophate (PI3P) at the PAS. This facilitates the attachment of b-propellers that bind polyphosphoinositides (PROPPINs), which are a well-conserved group of WD40-repeat proteins that typically adopt a seven-bladed b-propeller structure. PROPPINs moderate protein-protein interactions, such as scaffolding or assembly and regulation of multi-subunit complexes, due to their WD40 domain. Additionally, PROPPINs feature a FRRG motif, which is highly conserved, at their fifth and sixth propeller blade. This motif enables the binding of phosphoinositides, namely two molecules of PI3P or PI(3,5)P2.
Three PROPPINs have been identified in Saccharomyces cerevisiae: Atg18, Atg21 and Hsv2. While they exhibit high homology, each fulfills distinct functions during different subtypes of autophagy. Atg18 is essential for all forms of autophagy depending on PI3P, which makes it a core autophagy protein. Furthermore, Atg18 contributes to non- autophagic functions at the vacuole, which are related to vacuolar morphology and fragmentation. In contrast, Atg21 is not essential for non-selective autophagy but carries out vital functions during selective autophagic processes, such as the Cvt pathway, which specializes in transport of vacuolar hydrolases like prApe1 to the vacuole. Opposite of these well-defined PROPPINs stands Hsv2, which is the least characterized of the three. It has small implications for piecemeal microautophagy of the nucleus, which marks the only known detection impacting autophagy.
This study focused on investigating the membrane bending ability of the yeast PROPPINs. This membrane bending is driven by the partial insertion of an amphipathic a-helix, which is formed by the loop 6CD of the PROPPINs upon contact with the bilayer. Since Hsv2 is the least characterized PROPPIN, this study put most emphasize on it and its potential autophagic and non-autophagic functions. Mutants of all three PROPPINS were designed and tested, which lack the ability to form an intact amphipathic a-helix and induce membrane curvature and fission.
The first experiments conducted in this study addressed the relevance of the membrane bending ability for the autophagic functions of Atg18 and Atg21. The autophagic assays showed that Atg21 needs its membrane bending ability for prApe1 maturation during the Cvt pathway and also to a lesser extent during bulk autophagy under starvation conditions. Atg18 and Hsv2 showed no involvement of their membrane bending ability during their autophagic functions. Nonetheless, we could show that the autophagy of large cargos depends on Hsv2. Here, Hsv2 binds to the growing edges of the phagophore and supports Atg18 in its function to facilitate growth of the isolation membrane.
Previous studies in our lab showed that Atg18 is able to bind to the retromer complex and regulate vacuolar morphology and vacuolar fragmentation. My experiments show, that Hsv2 is also able to bind to the retromer complex, especially in absence of Atg18 and mediate vacuolar fragmentation. Moreover, I could show that the membrane bending ability of Hsv2 is necessary for its retromer function. In addition, deletion of Hsv2 led to mislocalization of the t-SNARE Pep12 in diploid yeast cells. This indicates an additional function of Hsv2 with retromer that involves recycling of proteins.
Since we found a function of Hsv2, restricted to diploid cells, we assumed an involvement of Hsv2 during sporulation, which was already hinted to in other studies. My sporulation experiments showed that hsv2∆ cells exhibit an impaired spore wall phenotype. Furthermore, we determined that Gsc2 and Chs3, two proteins involved in spore wall assembly, are mislocalizing in deletion cells as well as in mutated Hsv2 cells that no longer can form an intact a-helix in loop 6CD and induce membrane bending. This led to the conclusion, that Hsv2 retromer complex is involved in recycling of proteins during sporulation.2026-01-2
Kardiale Magnetresonanztomographie: Semiautomatische dreidimensionale Analyse und Visualisierung linksatrialer struktureller Veränderungen
Get PDFHintergrund: Aktuelle Studien und Leitlinien stärken die Rolle der Katheterablation als Therapieoption bei Vorhofflimmern. Der Therapieerfolg bei Erstablation ist jedoch eingeschränkt und abhängig von Ausmaß des linksatrialen strukturellen Remodelings und insbesondere der linksatrialen Fibrose. Die 3D-LGE-MRT ist die einzige nichtinvasive Methode mit dem Potential zur Patientenstratifikation, Festlegung möglicher individualisierter Ablationsstrategien sowie Kontrolle der Ablationsnarben. Für ihren objektiven, reproduzierbaren und zeiteffizienten Einsatz sind automatische Verfahren zur Segmentierung des linken Atriums und Quantifizierung der linksatrialen Fibrose in 3D-LGE-Bilddaten erforderlich. Kommerziell verfügbare Softwarelösungen zu diesem Zweck existieren aktuell nicht.
Methoden: Anhand von 34 Datensätzen, welche im Rahmen der FIND-AFRANDOMISED-Studie erhoben wurden, wurde in MATLAB ein Programm zur weitgehend automatischen Segmentierung des linken Atriums und der angrenzenden Strukturen entwickelt. Wesentliche Bausteine sind die Registrierung von 3D-MRA- und 3D-LGE-MRT-Bilddaten, die Multi-Atlas Whole Heart-Registrierung zur initialen Segmentierung aller vier Herzkammern und der angrenzenden großen Gefäße sowie die Level-Set-Segmentierung zur Verbesserung der Segmentierung lokaler Details wie insbesondere der proximalen Pulmonalvenen.
Ergebnisse: Das entwickelte Programm erlaubte in 30 von 34 untersuchten Datensätzen (88,2%) die weitgehend automatische Segmentierung des linken Atriums und der proximalen Pulmonalvenen in wenigen Schritten der Nutzerinteraktion und liefert strukturelle Parameter (LAV/LAVI, verschiedene Durchmesser, Sphärizität), eine 3D-Visualisierung fibrotischer Areale als Funktion verschiedener Intensitäts-Thresholds sowie die Durchmesser der Pulmonalvenenostien zur Ablationsplanung. Im Vergleich zur manuellen Segmentierung wurden bei einer Zeitreduktion von 40,4±15,3 auf 4,6±0,5 Minuten pro Datensatz ein mittlerer Dice-Koeffizient von 0,918±0,015 sowie eine mittlere Oberflächendistanz von 1,16±0,21 mm erzielt. Als erste Anwendung wurde die linksatriale Fibrose im Patientenkollektiv für verschiedene, in der Literatur existierende Thresholds verglichen. Je nach Wahl des Thresholds kann der ermittelte Anteil fibrotischen Myokards erheblich schwanken.
Schlussfolgerung: Die automatische Segmentierung des linken Atriums in 3D-LGE-MRT-Bilddaten erleichtert und standardisiert die Quantifizierung des linksatrialen strukturellen Remodelings und eröffnet neue Optionen zur Therapieplanung und -evaluation der Katheterablation bei Vorhofflimmern. Weitere Forschung ist insbesondere zur Standardisierung der Differenzierung von gesundem und fibrotischem linksatrialen Myokard erforderlich.Background: Catheter ablation is an established treatment for atrial fibrillation (AF), yet its efficacy, particularly for first-time procedures, is influenced by the degree of left atrial (LA) structural remodeling - most notably, LA fibrosis. Among current imaging modalities, three-dimensional late gadolinium enhancement MRI (3D LGE-MRI) remains the only non-invasive technique capable of stratifying patients, informing individualized ablation strategies, and assessing post-procedural scar formation. However, clinical adoption is hindered by the lack of standardized, automated tools for LA segmentation and fibrosis quantification.
Methods: Using data from 34 participants in the FIND-AFRANDOMISED study, we developed a semi-automated MATLAB-based pipeline for segmenting the LA and adjacent anatomical structures. The workflow integrates: (1) registration of 3D magnetic resonance angiography (MRA) and 3D LGE-MRI datasets; (2) multi-atlas-based whole-heart registration for initial segmentation of cardiac chambers and large vessels; and (3) level-set segmentation to refine anatomical details, particularly at the pulmonary vein ostia.
Results: The pipeline successfully performed automated LA and proximal pulmonary vein segmentation in 30 of 34 datasets (88.2%) with minimal user intervention. The tool computes structural parameters (e.g., LA volume and volume index, chamber diameters, sphericity), provides 3D visualization of fibrotic regions across multiple intensity thresholds, and estimates pulmonary vein ostial diameters to support ablation planning. Compared to manual segmentation, the method achieved a mean Dice coefficient of 0.918 ± 0.015 and a mean surface distance of 1.16 ± 0.21 mm, while reducing analysis time per dataset from 40.4 ± 15.3 minutes to 4.6 ± 0.5 minutes. As a first application, fibrosis quantification across different literature-reported intensity thresholds was performed, and demonstrated substantial variability in estimated fibrotic burden.
Conclusion: The proposed automated segmentation approach enables efficient, reproducible assessment of LA remodeling in 3D LGE-MRI and supports advanced planning and evaluation of AF ablation procedures. Further standardization of thresholding criteria for differentiation of healthy and fibrotic left atrial myocardium is essential for broader clinical implementation and inter-study comparability.2025-06-1
Auswirkung der klinischen MRT-Auswertung auf die Detektionsrate des Prostatakarzinoms im Rahmen der MRT/TRUS-Fusionsbiopsie der Prostata
Get PDFDas Prostatakarzinom stellt in Deutschland das am häufigsten diagnostizierte Karzinom des Mannes dar. Es ist entscheidend, klinisch signifikante Karzinome frühzeitig zu erkennen. Die Prostatabiopsie ist hierfür die bevorzugte diagnostische Methode. Im Rahmen der MRTTRUS Fusionsbiopsie werden mitunter gezielte Proben entnommen. Um die gezielten Stanzzylinder planen und entnehmen zu können wird der Operateur im Vorfeld mit möglichst viele Informationen über die suspekten Läsionen unterstützt. Ziel dieser Arbeit ist es zu überprüfen, ob MRT-Befunde mit einer höheren Anzahl an Informationen dazu beitragen können eine gesteigerte Detektionsrate im Rahmen der MRT-TRUS Fusionsbiopsie zu erzielen.
Untersucht wurden 222 männliche Patienten, die im Zeitraum zwischen dem 01.01.2016 und 31.12.2017 eine MRT/TRUS-Fusionsbiospie der Prostata in der Klinik für Urologie in der Universitätsmedizin Göttingen erhielten. Die vorliegenden Befunde wurden anhand eines Scores evaluiert und anschließend gruppiert. Im Anschluss erfolgte eine gezielte Auswertung des Gesamtkollektives und der jeweiligen Befundgruppen.
In der Gruppe (N=192), mit den nach dem Score am höchsten bewerteten Befunden, konnte durch die gezielte Stanzbiopsie konnte bei 16 % ein PCa und in 37 % ein signifikantes PCa detektiert werden. Im paarweisen Vergleich war dieser Unterschied signifikant. In den Gruppen mit einer mittleren Informationsdichte (N=22) gemäß des Scores zeigte sich in den gezielten Stanzbiopsien, 22 % PCa und 14 % signifikante PCa. In der Gruppe mit weniger Informationen (N=8) erfolgte ein Nachweis von 13 % signifikanter PCa.
Die vorliegende Arbeit konnte zeigen, dass Befunde mit einem höheren Informationsgehalt, eine gesteigerte Detektion in den gezielten Stanzbiopsien an Prostatakarzinomen, vor allem auch signifikanten Prostatakarzinomen, aufweisen.Prostate carcinoma is the most frequently diagnosed cancer in men in Germany. It is crucial to recognise clinically significant carcinomas at an early stage. Prostate biopsy is the preferred diagnostic method for this. Targeted samples are sometimes taken as part of the MRITRUS fusion biopsy. In order to be able to plan and remove the targeted punch cylinders, the surgeon is supported in advance with as much information as possible about the suspicious lesions. The aim of this study is to examine whether MRI findings with a higher amount of information can contribute to an increased detection rate in the context of MRI-TRUS fusion biopsy.
The study examined 222 male patients who underwent MRI/TRUS fusion biopsy of the prostate at the Department of Urology at the University Medical Centre Göttingen between 1 January 2016 and 31 December 2017. The available findings were evaluated using a score and then grouped. Subsequently, the overall collective and the respective groups of findings were analysed.
In the group (N=192) with the highest score, the targeted punch biopsy detected PCa in 16% and significant PCa in 37%. This difference was significant in the paired comparison. In the groups with an average information density (N=22) according to the score, 22 % PCa and 14 % significant PCa were detected in the targeted punch biopsies. 13 % significant PCa were detected in the group with less information (N=8).
The present study was able to show that findings with a higher information content exhibit increased detection in targeted punch biopsies of prostate carcinomas, especially significant prostate carcinomas.2025-06-0
Pharmacotherapeutic approaches against Plasmodium falciparum
No full textMalaria tropica persists as one of the most impactful infectious diseases, threatening the health of nearly half of the world’s population every year. The disease is caused by Plasmodium falciparum (P. falciparum), a protozoan parasite transmitted
through the bite of female Anopheles mosquitoes. Currently available vaccines do not provide sufficient long-lasting immunity, prevention and eradication measures are only partially effective, and therapeutic options are limited. Furthermore, the use of
artemisinin-based combination therapies (ACTs), which represent the most successful antimalarial strategies to date, has been threatened by rapid emergence of drug resistance, necessitating the need to develop and implement novel approaches
against malaria.
The following thesis comprises two independent projects, which present findings on antimalarial strategies that warrant further exploration. As a model system, P. falciparum-infected human red blood cells (RBCs) were used.
First, we investigated membrane-permeable derivatives of 5-fluorodeoxyuridine triphosphate (FdUTP), which exhibited strong anti-parasitic activity against P. falciparum in vitro. Due to the lack of nucleoside kinases in this species, the use of nucleoside analogues has been precluded so far. Furthermore, phosphorylated nucleosides are too polar to permeate lipid membranes, which diminishes their efficacy. To overcome this challenge, we used cell-permeable analogues of FdUTP (cpFdUTP) with lipid-like modifications, which facilitated their entry into the red blood cell and the parasite. The membrane-permeable FdUTP
derivatives demonstrated rapid and potent anti-parasitic effects, interfering with DNA synthesis and arresting parasite development predominantly at the trophozoite-to-schizont transition. Moreover, the toxicity of cpFdUTP against human cells was ameliorated by the addition of thymidine or cell-permeable deoxythymidine triphosphate (cpdTTP) without restoring parasite proliferation. Thus, this opens a therapeutic window and highlights the potential application of cpFdUTP in combination with thymidine as a novel approach to treat malaria.
In addition to that, the combination of the commonly used antimalarial Pyrimethamine and the ferroptosis-inducing agent RAS-selective lethal 3 (RSL3) proved highly efficient in eliminating P. falciparum within human RBCs. Pyrimethamine is a well-known inhibitor of Plasmodium dihydrofolate reductase (DHFR), whereas RSL3 triggers lipid peroxidation-mediated cell death, and the combination of both exhibited strong synergy against the parasites. Initially, our hypothesis on the mechanism underlying this synergy was based on the ferroptosis-inducing activity of RSL3. Surprisingly, various combinations of DHFR inhibitors and alternative ferroptosis inducers did not achieve a comparable degree of synergy. Furthermore, Ferrostatin-1, an antagonist of lipid peroxidation, failed to rescue parasites from the deleterious impact of RSL3, pointing to a mechanism that is largely independent of oxidative stress and lipid peroxidation. In any case, the combination of RSL3 and Pyrimethamine holds promise as a novel antimalarial strategy that is worth investigating further.
Taken together, the current studies present two different approaches for elimination of P. falciparum within human red blood cells, which have the potential to be translated into therapeutic strategies against malaria and warrant further exploration and development.2025-06-3
Fluorespirometrische Charakterisierung mitochondrialer Dysfunktion der Hirnregionen Cortex und Hippocampus eines Mausmodells für das Rett-Syndrom
No full textDas Rett-Syndrom, meist durch Mutationen im Methyl-CpG bindendes Protein 2 (Mecp2) verursacht, ist eine schwere neurologische Entwicklungsstörung. Eine mitochondriale Dysfunktion wird als wichtiger Faktor in der Pathogenese diskutiert. Ziel dieser Arbeit war die fluorespirometrische Charakterisierung mitochondrialer Dysfunktion im Cortex und Hippocampus eines Mausmodells für das Rett-Syndrom. Mittels hochauflösender Respirometrie und spezifischer Fluorophore wurden in etablierten Substrat-Entkoppler-Inhibitor Titrationsprotokollen der Sauerstoffverbrauch, die H2O2-Produktion sowie die ATP-Produktion untersucht, wobei eine neuartige kontinuierliche Darstellung der P»/O2-Effektivität entwickelt und angewandt wurde.
Es zeigte sich bei MeCP2-defizienten Tieren eine kortikal betonte, reduzierte Elektronentransferkapazität und eine verminderte maximale Kapazität der oxidativen Phosphorylierung. Die ATP-Produktionseffizienz (P»/O2) war jedoch unverändert. Extramitochondrial wurde eine reduzierte H2O2-Abgabe detektiert, während der relative Sauerstoffverbrauch im LEAK-Zustand (Respirationszustand in Abwesenheit von ADP) erhöht war.
Eine Variation der verfügbaren O2-Konzentration als zusätzlichen Stressor zeigte keine weitere Veränderung der Respirations- und H2O2-Abgaberaten. Eine Überexpression einer mitochondrialen Katalase (mCAT) zeigte keine signifikanten therapeutischen Effekte auf die dysfunktionale Atmungskettenfunktion. Eine Supplementation antioxidativer Stoffe in der Nahrung hatte geringe normalisierende Effekte auf den kortikalen Sauerstoffverbrauch der MeCP2-defizienten Tiere.
Zusammenfassend deuten die Ergebnisse auf eine reduzierte maximale ATP-Syntheseleistung bei erhaltener Effizienz in den Mitochondrien MeCP2-defizienter Mäuse hin, insbesondere im Cortex. Die verminderte extramitochondriale H2O2-Detektion, trotz Anzeichen für erhöhten oxidativen Stress, könnte auf eine gestörte intramitochondriale ROS-Verarbeitung hindeuten.Rett syndrome, typically caused by mutations in Methyl-CpG-binding protein 2 (Mecp2), is a severe neurodevelopmental disorder. Mitochondrial dysfunction is implicated as a significant factor in its pathogenesis. This study aimed to characterize mitochondrial dysfunction in the cortex and hippocampus of a Rett syndrome mouse model using fluorespirometry. Utilizing high-resolution respirometry and specific fluorophores with established substrate-uncoupler-inhibitor titration (SUIT) protocols, oxygen consumption, H2O2 production, and ATP production were investigated. A novel approach for the continuous representation of P»/O2 efficiency was developed and applied.
MeCP2-deficient animals exhibited a cortically pronounced reduction in electron transfer capacity and diminished maximal oxidative phosphorylation capacity. However, ATP production efficiency (P»/O2) remained unaltered. Extramitochondrial H2O2 release was found to be reduced, whereas relative oxygen consumption in the LEAK state (respiration in the absence of ADP) was elevated.
Varying available O2 concentration as an additional stressor did not induce further alterations in respiration or H2O2 release rates. Overexpression of mitochondrial catalase (mCAT) yielded no significant therapeutic effects on dysfunctional respiratory chain function. Dietary supplementation with antioxidants exerted slight normalizing effects on cortical oxygen consumption in MeCP2-deficient animals.
In conclusion, the findings suggest a reduced maximal ATP synthesis capacity with preserved efficiency in mitochondria from MeCP2-deficient mice, particularly in the cortex. The diminished extramitochondrial H2O2 detection, despite indications of elevated oxidative stress, may point towards impaired intramitochondrial ROS processing.2025-07-2
Agricultural Sustainable Development: Technology Adoption and Climate Change
No full textThe dissertation with the title “Agricultural Sustainable Development: Technology Adoption and Climate Change” aims to advance our understanding of how sustainable agricultural practices can be achieved through a multifaceted approach that addresses the growing challenges in agriculture. In an era marked by resource scarcity, rapid climate change, and the need for increased food production, sustainable development in agriculture has become critical not only to ensure food security but also to safeguard environmental health and resilience. This project aims to 1) examine the role of technology adoption in enhancing agricultural productivity and profitability for producers, 2) investigate the impacts of climate change, economic growth, government policies, and land quality on cropland loss and food security, and 3) assess the evolution of policies that support long-term sustainability goals. By integrating these perspectives, the project seeks to provide actionable insights to guide policymakers, farmers, and stakeholders in developing effective solutions for sustainable agriculture.
In the first study, we examine whether technology adoption benefits agricultural producers in China. Using nationally representative survey data, we draw on evidence from product certification and e-commerce adoption and apply multinomial endogenous switching regression (MESR) to analyze production and profitability outcomes for grain and cash crop producers. We find that agricultural producers who adopt e-commerce, product certification, or both strategies have significant price premiums, increased sales revenue, and greater profitability, despite the increased production costs, with cash crop producers and large-scale producers gaining more benefits. Additionally, we find a synergistic effect: agricultural producers benefit more from adopting both product certification and e-commerce simultaneously than the sum of the effects of adopting the two strategies separately. This positive reinforcement between the two strategies is more pronounced among cash crop producers compared to grain crop producers, as well as among larger-scale producers compared to their small-scale counterparts. These findings provide valuable policy implications for assisting agricultural producers in adopting suitable strategies, thereby enhancing their competitiveness and profitability within the global value chains of agri-food. While these findings emphasize producer-centric strategies for sustainable agriculture, we also recognize the need to investigate broader production-level dynamics and environmental pressures.
The second study shifts the focus to cropland loss and its implications for agricultural production and food security in China—key components of sustainable development beyond the individual farm level. Cropland loss poses significant threats to agricultural production and food security in China, yet knowledge of its recent trajectories and underlying drivers remains limited. This study investigates the impact of climate (particularly temperature extremes), economic growth, government policies, and land quality on cropland loss, using county-level data from 2016 to 2019. By differentiating cropland abandonment, urbanization-driven conversion, and reforestation, we provide a nuanced analysis of cropland dynamics. We find that, first, cropland loss is increasingly pronounced, with abandonment accounting for nearly 90% of the annual reduction. Second, climate change, especially the increasing high heat, accelerates cropland loss through (1) reduced agricultural profitability in low-productivity, grain-focused areas with rigid pricing policies, and (2) a shift toward off-farm employment in economically underdeveloped regions. Third, southern counties, major grain-producing zones, and agriculturally dependent areas remain particularly vulnerable to extreme heat, especially in high-productivity flatlands. Finally, cropland expansion does not offset these losses, but better mechanization enhances resilience to heat-induced pressures. Several policy implications can be drawn from this study, including promoting high-standard cropland construction to enhance soil quality and yields, reforming grain pricing mechanisms and subsidies to stabilize incomes, and modernizing agricultural technologies and land-management practices such as land transfer and mechanization.
While the second study emphasizes cropland loss as a critical aspect of agricultural sustainable development, the third study addresses the methodological challenges in accurately assessing climate change’s impact on agricultural production by examining measurement errors in climate data. Climate change and its effects on economic growth, particularly agricultural production, are key topics in academic research and policy-making. Accurate assessment of climate impacts on agriculture requires precise climate data, yet weather data often contain measurement errors, posing challenges for studies on agriculture's economic implications. The third study examines how measurement errors, particularly urban heat island (UHI) biases in temperature data, affect estimates of climate change's impact on agricultural production. Using five years (2015-2019) of unbalanced panel data covering 2,288 counties in mainland China, we confirm higher temperatures in built-up areas compared to cropland areas, leading to a 21.94% underestimation of climate impact on grain yield, and the bias is more pronounced in developed regions. Further, this study analyzes heterogeneity at two levels: crop types and economic geographical regions. Heterogeneous analysis based on cropland temperature reveals that potatoes are the most sensitive to climate change, followed by wheat, rice, and corn. The impact of temperature on grain yield is positive in northern China and negative, though not significant, in southern China, suggesting environmental injustice under global warming. Furthermore, correcting these errors reveals previous studies understated the temperature impact on yield by over 50%, emphasizing the need for precise temperature data and refined data processing methods for future research.
In the final study of this dissertation, we shift our focus to a policy-centered analysis, evaluating the evolution of policies designed to achieve long-term sustainability goals. The Common Agricultural Policy (CAP) of the European Parliament (EP) serves as a cornerstone of European public policy studies and represents the world's leading agricultural support system. This study examines whether government support policies in the agricultural sector have evolved toward a fairer, greener, and more flexible framework. To guide the analysis, we develop the SSCIR-3F analytical framework, which integrates five key CAP policy domains—Food Security (S), Sustainability (S), Agricultural Competitiveness (C), Farmer Income (I), and Rural Development (R)—with the triphasic progression of agricultural development: Food, Farm, and Future (3F). This framework captures CAP’s policy priorities, budgetary dynamics, and the theoretical stages of agricultural development. Using machine learning techniques, including text mining and clustering, this study analyzes legislative texts ratified by the EP to identify patterns of harmony, conflict, and evolution within the CAP. The findings reveal a clear prioritization in recent EP discussions: sustainability has emerged as the dominant focus, followed by farmer incomes, food security, rural development, and competitiveness. This hierarchy aligns closely with the SSCIR-3F framework. Clustering results further highlight key themes such as biodiversity, ecosystems, forests, and climate change, reinforcing the increasing emphasis on environmental sustainability. Additionally, a congruence between CAP expenditures and the evolving SSCIR priorities was identified, suggesting that policy articulation increasingly aligns with tangible implementation. These findings underscore the CAP's trajectory toward a more ecologically focused, balanced, and adaptable policy paradigm.
Overall, this thesis attempts to explain the multifaceted challenges and opportunities in achieving agricultural sustainable development, emphasizing the roles of technology, climate, land use, and policy evolution. By combining empirical analysis, methodological refinement, and policy evaluation, this thesis provides insights into advancing sustainable agricultural practices and adapting agricultural systems to environmental and socioeconomic pressures. It offers implications for stakeholders to address pressing global challenges in food security and sustainability.2025-07-0
Circulating Methylated DNA Fragments of the Genes ATP2A2, B9D1 and RASAL1 in Blood as Biomarkers for Heart Failure with Preserved Ejection Fraction (HFpEF) and Aortic Stenosis
No full textSowohl Patienten mit HFpEF als auch Patienten mit Aortenklappenstenose weisen im Allgemeinen eine hohe Mortalität und schlechte Prognose auf. Mit den aktuell zur Verfügung stehenden klinisch etablierten Biomarkern gelingt keine ausreichende Identifikation von Hochrisikopatienten und keine ausreichende Prognostizierung des individuellen Krankheitsprogresses, sodass für beide Erkrankungen weiterhin eine schlechte Prognose besteht. Die Prävalenz der Erkrankungen HFpEF und Aortenklappenstenose zeigt eine altersabhängige Verschiebung, im Sinne einer hohen Prävalenz insbesondere bei älteren Patienten, sodass die Relevanz der Erkrankungen in der weltweit alternden Gesellschaft voraussichtlich zunehmen wird. Beide Erkrankungen teilen sich die pathomechanistische Bedeutung der Herzfibrose, die zudem eine prognostische Relevanz zeigt. Die DNA-Methylierung gilt als einer der Hauptmechanismen der epigenetischen Modifikation menschlicher DNA. Die aberrante Promotormethylierung spezifischer Gene zeigt einen kausalen Einfluss auf die pathologische Fibrogenese im Herzen. In Vorarbeiten ergab sich, entsprechend des Prinzips der Liquid Biopsy, eine Korrelation der im Blut gemessenen aberranten Promotormethylierung spezifischer Gene mit der im Herzgewebe gemessenen Promotormethylierung dieser Gene. Anhand des Prinzips der Liquid Biopsy ergeben sich damit Hinweise auf die Eignung der aberranten Promotormethylierung spezifischer Gene im Blut als Biomarker für kardiale Erkrankung mit pathomechanistischer Gemeinsamkeit der Herzfibrose. Aufgrund eines Mangels an Myokardbiopsien diente in der vorliegenden Arbeit die Mortalität als Korrelat für Herzfibrose und den Erkrankungsprogress. Eine Vorarbeit der Arbeitsgruppe Zeisberg ergab bereits Hinweise auf die Eignung der im Blut gemessenen aberranten Promotormethylierung der drei Gene ATP2A2, B9D1 und RASAL1 als prognostische Biomarker für Patienten mit der Erkrankung HFpEF. Auch für das Krankheitsbild Aortenklappenstenose konnte eine Vorarbeit der Arbeitsgruppe Zeisberg Hinweise auf die Eignung der aberranten Promotormethylierung spezifischer Gene im Blut als prognostischen Biomarker finden.
Das Ziel der vorliegenden Arbeit war es, anhand des Prinzips der Liquid Biopsy zu untersuchen, ob sich die aberrante Promotormethylierung der drei Gene ATP2A2, B9D1 und RASAL1 im Blut als Biomarker für die klinische Progression und zur Risikostratifizierung von Patienten mit HFpEF bzw. Aortenklappenstenose eignet, die sich die pathomechanistische Gemeinsamkeit der Herzfibrose teilen. Zudem sollte überprüft werden, ob die Messung der aberranten Promotormethylierung der drei Gene im Blut Vorteile gegenüber der Erfassung anderer klinischer Parameter bietet.
Es wurden Blutproben und verschiedene klinische Parameter von Patienten aus drei Kohorten analysiert. Die DIAST- und HFpEF-Kohorte dienten der Betrachtung der Grunderkrankung HFpEF. Die TAVI-Kohorte diente der Betrachtung der Grunderkrankung Aortenklappenstenose. Zur Messung der aberranten Promotormethylierung der drei Gene ATP2A2, B9D1 und RASAL1 wurde die DNA zunächst aus Blutproben isoliert und aufgereinigt. Die methylierte DNA konnte anschließend antikörpervermittelt per MeDIP separiert werden. Der Anteil der Promotormethylierung der drei Gene wurde daraufhin per qRT-PCR ermittelt.
Nach einer allgemeinen Charakterisierung der untersuchten Kohorten folgte zunächst die Untersuchung von Zusammenhängen der Promotormethylierung der drei Gene im Blut mit klinischen Parametern. Einzig die Promotormethylierung des Gens B9D1 korrelierte hierbei signifikant mit dem bereits etablierten kardialen Biomarker BNP. Eine höhere B9D1-Promotormethylierung im Blut ging mit höheren Werten von BNP im Blut einher. Dieses Ergebnis kann als Hinweis darauf verstanden werden, dass das Gen B9D1 eine Rolle am Herzen spielt und die aberrante B9D1-Promotormethylierung wie BNP eine kardiale Pathologie widerspiegelt. Höhere Werte von BNP im Blut gehen im Allgemeinen mit einer verstärkten kardialen Pathologie einher. Eine aberrante Promotormethylierung führt in der Regel zu einer verminderten Genexpression. In Zusammenschau lassen die Ergebnisse der vorliegenden Arbeit daher eine protektive Genfunktion des Gens B9D1 am Herzen vermuten, die im Rahmen kardialer Pathologien durch eine aberrante Promotormethylierung des Gens B9D1 verloren geht. Für alle weiteren klinischen Parameter ergab sich keine signifikante Korrelation mit der Promotormethylierung der drei untersuchten Gene im Blut. Diese insgesamt fehlende Korrelation kann als Hinweis darauf verstanden werden, dass die Messung der Promotormethylierung im Blut der Erfassung anderer klinischer Parameter, einschließlich bereits etablierter Biomarker, überlegen ist und einen Mehrwert bieten kann.
Im zweiten Teil der Arbeit wurde die Möglichkeit zur Nutzung der aberranten Promotormethylierung der drei Gene ATP2A2, B9D1 und RASAL1 im Blut als prognostischer Biomarker für die Krankheitsbilder HFpEF und Aortenklappenstenose evaluiert. In Überlebensmodellen an der DIAST-Kohorte zeigte die aberrante Promotormethylierung der drei Gene im Blut einen signifikanten Einfluss auf das Gesamtüberleben der Patienten mit der Grunderkrankung HFpEF. Diese Überlebensmodelle wurden in Form von Cox-Regressionen auf die HFpEF- und TAVI-Kohorte übertragen. Die ATP2A2-Promotormethylierung im Blut konnte für beide Krankheitsbilder nicht als prognostischer Biomarker validiert werden. Am Datensatz der TAVI-Kohorte zur Untersuchung des Krankheitsbildes Aortenklappenstenose gelang ebenfalls keine Validierung der Promotormethylierung der drei Gene im Blut als prognostischer Biomarker. Am Datensatz der HFpEF-Kohorte ergaben sich dagegen Hinweise auf die Eignung der Promotormethylierung der Gene B9D1 und RASAL1 im Blut als prognostische Biomarker für Patienten mit der Grunderkrankung HFpEF. Die im ersten Teil der Arbeit herausgestellte fehlende Korrelation der aberranten Promotormethylierung der drei Gene im Blut mit klinischen Parametern kann in Zusammenschau mit den Ergebnissen des zweiten Teils dieser Arbeit als Zeichen der Spezifität und Unabhängigkeit und damit des Mehrwertes der neuen potentiellen prognostischen Biomarker B9D1- und RASAL1-Promotormethylierung im Blut interpretiert werden.
Resümierend konnte im Rahmen der vorliegenden Arbeit die Promotormethylierung der zwei Gene B9D1 und RASAL1 im Blut als potentieller prognostischer Biomarker für die Erkrankung HFpEF bestätigt und der Mehrwert hinsichtlich der Erhebung anderer klinischer Parameter herausgestellt werden. Es ergeben sich damit anhand der Ergebnisse der vorliegenden Arbeit Hinweise auf die Eignung der aberranten Promotormethylierung spezifischer Gene im Blut für kardiale Erkrankungen, die sich die pathomechanistische Gemeinsamkeit der Herzfibrose teilen. Hier müssen sich Folgestudien anschließen.Both patients with HFpEF and those with aortic stenosis generally have a high mortality and a poor prognosis. With the currently available clinically established biomarkers, it is not possible to sufficiently identify high-risk patients or adequately predict the individual progression of the disease, resulting in a persistently poor prognosis for both conditions. The prevalence of HFpEF and aortic stenosis shows an age-dependent shift, with a particularly high prevalence among elderly patients, suggesting that the relevance of these diseases will likely increase in the context of an aging global population. Both diseases share the pathomechanistic significance of cardiac fibrosis, which also has prognostic relevance. DNA methylation is considered one of the main mechanisms of epigenetic modification of human DNA. Aberrant promoter methylation of specific genes has been shown to causally influence pathological fibrogenesis in the heart. In accordance with the principle of liquid biopsy, previous studies have demonstrated a correlation between the aberrant promoter methylation of specific genes measured in the blood and the promoter methylation of these genes measured in heart tissue. These findings, rooted in the principle of liquid biopsy, indicate the potential suitability of aberrant promoter methylation of specific genes in the blood as a biomarker for cardiac diseases sharing the pathomechanistic feature of cardiac fibrosis. Due to a lack of myocardial biopsies, mortality served as a correlate for cardiac fibrosis and disease progression in the present study. A preliminary study by the Zeisberg research group has already provided indications for the suitability of aberrant promoter methylation of the three genes ATP2A2, B9D1 and RASAL1 measured in blood as prognostic biomarkers for patients with HFpEF. Another preliminary study by the Zeisberg research group found indications of the suitability of aberrant promoter methylation of specific genes in the blood as prognostic biomarkers for aortic stenosis.
The aim of the present study was to investigate, in accordance with the principle of liquid biopsy, whether aberrant promoter methylation of the three genes ATP2A2, B9D1 and RASAL1 in blood is suitable as a biomarker for clinical progression and for risk stratification of patients with HFpEF or aortic stenosis who share the pathomechanistic feature of cardiac fibrosis. Furthermore, the study aimed to determine whether measuring the aberrant promoter methylation of these three genes in the blood offers advantages over the measurement of other clinical parameters.
Blood samples and various clinical parameters were analyzed from patients in three cohorts. The DIAST and HFpEF cohorts were used to study the underlying disease HFpEF. The TAVI cohort was used to investigate the underlying disease aortic stenosis. To measure the aberrant promoter methylation of the three genes ATP2A2, B9D1 and RASAL1, the DNA was first isolated from blood samples and subsequently purified. In the following step methylated DNA was separated using antibody-mediated MeDIP. Finally, the proportion of promoter methylation of the three genes was determined by qRT-PCR.
Following a general characterization of the studied cohorts, the investigation initially focused on correlations between the promoter methylation of the three genes in blood and clinical parameters. Only the promoter methylation of the B9D1 gene correlated significantly with the already established cardiac biomarker BNP. Higher B9D1 promoter methylation in the blood was associated with higher levels of BNP in the blood. This result can be interpreted as an indication that the B9D1 gene plays a role in the heart and that aberrant B9D1 promoter methylation, like BNP, reflects cardiac pathology. Higher levels of BNP in the blood are generally associated with increased cardiac pathology. Aberrant promoter methylation typically leads to reduced gene expression. Taken together, the results of the present study therefore suggest a protective gene function of the B9D1 gene in the heart, which is lost in the context of cardiac pathologies due to aberrant promoter methylation of the B9D1 gene. For all other clinical parameters, no significant correlation was found with the promoter methylation of the three genes examined in the blood. This overall lack of correlation can be taken as an indication that measuring promoter methylation in the blood is superior to measuring other clinical parameters, including already established biomarkers, and may offer added value.
In the second part of this study, the potential use of aberrant promoter methylation of the three genes ATP2A2, B9D1 and RASAL1 in the blood as a prognostic biomarker for the diseases HFpEF and aortic stenosis was evaluated. In survival models based on the DIAST cohort, aberrant promoter methylation of the three genes in the blood showed a significant influence on the overall survival of patients with the underlying disease HFpEF. These survival models were applied to the HFpEF and TAVI cohorts in the form of Cox regressions. ATP2A2 promoter methylation in the blood could not be validated as a prognostic biomarker for either of the two diseases. Using the data set of the TAVI cohort to study the disease aortic stenosis, it was also not possible to validate the promoter methylation of the three genes in the blood as a prognostic biomarker. In contrast, the data set of the HFpEF cohort provided indications of the suitability of promoter methylation of the genes B9D1 and RASAL1 in the blood as prognostic biomarkers for patients with the underlying disease HFpEF. The lack of correlation between aberrant promoter methylation of the three genes in the blood and clinical parameters, as highlighted in the first part of the study, can be interpreted in conjunction with the results of the second part of this study as a sign of the specificity and independence and thus the added value of the promoter methylation of the genes B9D1 and RASAL1 in the blood as new potential prognostic biomarkers.
In summary, this study confirmed that promoter methylation of the two genes B9D1 and RASAL1 in blood could represent a potential prognostic biomarker for HFpEF, demonstrating added value over the measurement of other clinical parameters. Based on these findings, there is evidence to suggest that aberrant promoter methylation of specific genes in blood could represent a suitable biomarker for cardiac diseases sharing the pathomechanistic feature of cardiac fibrosis. Further studies are warranted to validate these findings.2025-09-1
Analysis of Transcription Factor Binding and Cancer Stage-Specific Regulatory Dynamics Using DNA-Binding Domain Annotations
No full textTranscription Factors (TFs) regulate gene expression by binding to specific DNA sequences
through their DNA-Binding Domain (DBD)s. Accurate modeling of TF–DNA
interactions is essential for uncovering the mechanisms of gene regulation. This thesis introduces
TFClassPredict, a deep learning model built on a fine-tuned Bidirectional Encoder
Representations from Transformers (BERT) architecture that classifies DNA sequences into
TF groups defined by shared DBD features across different hierarchical levels. TFClassPredict
integrates TF structural properties, derived from their DNA-binding domains, with
DNABERT’s DNA language modeling capabilities to yield robust binding predictions.
The first part of the thesis covers the design, training, and evaluation of the model architecture.
TFClassPredict is trained on DNA sequences annotated with structural DBD
groupings and demonstrates strong classification performance across both broad and more
specific hierarchical levels. This structure-based approach reduces redundancy and enables
more reliable prediction of TF activity across diverse genomic contexts. The model’s strong
performance highlights that TFs sharing similar DNA-binding domains also tend to exhibit
similar binding site preferences.
In the second part of the thesis, TFClassPredict is applied to chromatin accessibility profiles
derived from cancer samples. The model is used to analyze regulatory activity in promoter
and enhancer regions and relate TF activity patterns to clinical features such as cancer
stage and survival. Several TFClasses show distinct patterns between early and late-stage
tumors, relating predicted regulatory activity to stages of cancer progression. The analysis
revealed distinct regulatory patterns between promoter and enhancer regions, highlighting
differences in TF activity across these elements. This analysis also uncovered distinct TF activity
profiles associated with early- and late-stage cancers, suggesting stage-specific shifts
in regulatory dynamics. These results demonstrate that TFClass-based predictions can distinguish
regulatory programs linked to tumor development, highlighting their potential for
characterizing disease progression.
The thesis offers a framework for predicting regulatory activity based on shared structural
features of TFs and their sequence preferences. The approach presented in this thesis offers
potential value in clinical research. By linking structural TF activity profiles to disease
states, the model may support biomarker discovery, patient stratification, and deciphering
regulatory patterns from clinical sequencing data. TFClassPredict’s predictive capabilities
could be applied to build tools for early-stage diagnostics or tailored treatment approaches.2025-08-0