Medical real-world data stored in clinical systems represents a valuable knowledge source for medical research, but its usage is still challenged by various technical and cultural aspects. Analyzing these challenges and suggesting measures for future improvement are crucial to improve the situation. This comment paper represents such an analysis from the perspective of research.

The balance of excitation and inhibition is essential for cortical information processing, relying on the tight orchestration of the underlying subcellular processes. Dynamic transcriptional control by DNA methylation, catalyzed by DNA methyltransferases (DNMTs), and DNA demethylation, achieved by ten–eleven translocation (TET)-dependent mechanisms, is proposed to regulate synaptic function in the adult brain with implications for learning and memory. However, focus so far is laid on excitatory neurons. Given the crucial role of inhibitory cortical interneurons in cortical information processing and in disease, deciphering the cellular and molecular mechanisms of GABAergic transmission is fundamental. The emerging relevance of DNMT and TET-mediated functions for synaptic regulation irrevocably raises the question for the targeted subcellular processes and mechanisms. In this study, we analyzed the role dynamic DNA methylation has in regulating cortical interneuron function. We found that DNMT1 and TET1/TET3 contrarily modulate clathrin-mediated endocytosis. Moreover, we provide evidence that DNMT1 influences synaptic vesicle replenishment and GABAergic transmission, presumably through the DNA methylation-dependent transcriptional control over endocytosis-related genes. The relevance of our findings is supported by human brain sample analysis, pointing to a potential implication of DNA methylation-dependent endocytosis regulation in the pathophysiology of temporal lobe epilepsy, a disease characterized by disturbed synaptic transmission.

The Eph receptor tyrosine kinases and their respective ephrin-ligands are an important family of membrane receptors, being involved in developmental processes such as proliferation, migration, and in the formation of brain cancer such as glioma. Intracellular signaling pathways, which are activated by Eph receptor signaling, are well characterized. In contrast, it is unknown so far whether ephrins modulate the expression of lncRNAs, which would enable the transduction of environmental stimuli into our genome through a great gene regulatory spectrum. Applying a combination of functional in vitro assays, RNA sequencing, and qPCR analysis, we found that the proliferation and migration promoting stimulation of mouse cerebellar granule cells (CB) with ephrinA5 diminishes the expression of the cancer-related lncRNA Snhg15. In a human medulloblastoma cell line (DAOY) ephrinA5 stimulation similarly reduced SNHG15 expression. Computational analysis identified triple-helix-mediated DNA-binding sites of Snhg15 in promoters of genes found up-regulated upon ephrinA5 stimulation and known to be involved in tumorigenic processes. Our findings propose a crucial role of Snhg15 downstream of ephrinA5-induced signaling in regulating gene transcription in the nucleus. These findings could be potentially relevant for the regulation of tumorigenic processes in the context of glioma.

Data quality deficiencies significantly limit the applicability of real-world data in data-driven medical research. In this study, using an oncological use case, we report and discuss common quality deficiencies in real-world medical datasets, such as missing data, class imbalances, and timeliness issues. We compiled a multi-departmental real-world dataset comprising 13861 cancer cases diagnosed at University Hospital Cologne and examined data quality throughout the data integration process.

The Cancer of Unknown Primary (CUP) syndrome is characterized by identifiable metastases while the primary tumor remains hidden. In recent years, various data-driven approaches have been suggested to predict the location of the primary tumor (LOP) in CUP patients promising improved diagnosis and outcome. These LOP prediction approaches use high-dimensional input data like images or genetic data. However, leveraging such data is challenging, resource-intensive and therefore a potential translational barrier. Instead of using high-dimensional data, we analyzed the LOP prediction performance of low-dimensional data from routine medical care. With our findings, we show that such low-dimensional routine clinical information suffices as input data for tree-based LOP prediction models. The best model reached a mean Accuracy of 94% and a mean Matthews correlation coefficient (MCC) score of 0.92 in 10-fold nested cross-validation (NCV) when distinguishing four types of cancer. When considering eight types of cancer, this model achieved a mean Accuracy of 85% and a mean MCC score of 0.81. This is comparable to the performance achieved by approaches using high-dimensional input data. Additionally, the distribution pattern of metastases appears to be important information in predicting the LOP.