Biology subjects

Kurth, I.

Publications and source records attributed to Kurth, I..

18 recordsLinked to original sources

bioRxiv2026

A relational architecture of the human genome

Complex systems exhibit organization not only through their components but also via their interdependencies. Local nucleotide dependencies underpin regulatory motifs and RNA structures, yet principles governing a higher-order organization across the human genome remain unknown. Here we systematically perturbed every position of the com- plete telomere-to-telomere human genome using Nucleotide Transformer v3 with an ex- ascale compute cluster, generating nucleotide-resolution dependency maps that reveal a directed, weighted network of position-to-position genomic relationships. This network recapitulates known biological features and uncovers recurrent, directional architec- tures termed 'relons', spanning coding, non-coding, and unannotated regions. These re- lational modules reflect evolutionary and functional constraints and associate with hu- man genetic variation and pathogenic variants. Our findings suggest a higher-order rela- tional layer of genome organization, providing a framework to interpret DNA as a complex system of dependencies beyond linear and 3D structure, with implications for under- standing genetic variation and disease.

genomics↗

bioRxiv2026

A Framework for NGN1-Induced Sensory Neuron Differentiation for Disease Modelling and Drug Screening

Background: Neuropathic pain is a burdensome, difficult-to-treat, and highly heterogeneous condition with limited therapeutic options, underscoring the need for robust and reproducible human disease models. Human induced pluripotent stem cell (iPSC)-derived sensory neurons provide a promising platform for patient specific disease modelling and drug screening; however, their translational use is hampered by variability in differentiation efficiency, cellular composition, and functional maturation across protocols and cell lines. Methods: Here, we present a standardized and potentially scalable framework for NGN1 driven differentiation of human iPSCs into sensory neurons. Building on a previously published two step protocol (1), we systematically deconstructed and optimized each stage of differentiation across a large panel of genetically diverse iPSC lines. Results: We identified robust parameters for neural crest like cell (NCLC) generation, established a flow cytometry-based quality control strategy for NCLCs, and defined optimal combinations of seeding density and lentiviral multiplicity of infection to maximize sensory neuron progenitor yield. To improve culture homogeneity, we compared antimitotic selection strategies and demonstrated that tightly timed Ara-C treatment combined with low progenitor seeding density yields consistently pure sensory neuron cultures. We further evaluated maturation under physiologically relevant glucose conditions and performed a systematic review of media compositions to derive two defined maturation media. Morphological, immunocytochemical, transcriptomic, and electrophysiological analyses revealed that time in culture is a major determinant of maturation, while specific supplements such as prostaglandin E2; (PGE2) selectively enhance transcriptional signatures associated with nociceptor identity without substantially altering global network activity. Bulk RNA sequencing demonstrated broad expression of sensory neuron and pain related markers and gene programs across conditions, with long term maturation and PGE2; treatment showing the highest similarity to human dorsal root ganglion reference data. Functional assessment using multi electrode arrays enabled the detection of donor specific electrophysiological phenotypes, including reproducible hyperexcitability in small fiber neuropathy patient derived lines. Conclusions: This study establishes a modular, reproducible NGN1 based differentiation workflow with integrated quality checkpoints that accommodates iPSC line to line variability. The framework provides a practical foundation for translational sensory neuron research, patient specific disease modelling, and scalable drug screening applications.

neuroscience↗

bioRxiv2025

Genolator: A Multimodal Large Language Model Fusing Natural Language, Genomic, and Structural Tokens for Protein Function Interpretation

BackgroundDecoding the genetic code to unveil its genome functionality is a monumental task which would greatly advance the understanding of disease mechanisms and development of targeted treatment approaches. Although large language models (LLMs) have transformed natural language processing across diverse domains, translating the complex language of DNA into human-readable form remains challenging due to genomic data complexity and unexplored regions of the human genome. Current (genomic) language models either have a solid understanding of natural language or of the genomic code. Models fusing both aspects are largely lacking. ResultsHere we present Genolator, a multimodal large language model that integrates embeddings from DNA sequences, amino acid sequences, and protein structures with natural language queries. Fine-tuned on over 370,000 question-answer pairs generated using abstracted Gene-Ontology (GO) terms, Genolator effectively answers queries regarding protein subcellular localization, molecular function, and biological processes. Evaluation demonstrates high accuracy in confirming or denying protein function associations, outperforming baseline models such as openly available allrounder LLMs like GPT 4.1 as well as smaller domain specific models integrating knowledge from foundation models like Evo2 and ESM-2. Explorations of the Genolators hidden states unveil a biologically and linguistically plausible organization of its learned representations. ConclusionGenolator enhances accessibility to genomic information by enabling natural language interaction with protein data, facilitating biological discovery, and clinical research. It represents a step towards bridging genomic code and human language through the integration of a multimodal LLM.

bioinformatics↗

bioRxiv2025

A Unified Single-Cell Atlas of HNSCC: Uncovering HPV and Sex Variability in the Tumor Microenvironment

Head and neck squamous cell carcinoma (HNSCC) is highly heterogeneous, with variations driven by HPV status and sex. However, existing single-cell RNA sequencing (scRNA-seq) studies are often limited in sample size and lack standardized methodologies, limiting cross-study comparisons. To address this, we integrated scRNA-seq data from 78 patients (274,911 cells) across multiple studies, creating a unified HNSCC atlas that harmonizes annotations and enables robust tumor microenvironment (TME) analyses. Using STACAS for semi-supervised integration and automated annotation tools such as Ikarus and scGate, we improved tumor and immune cell classification. Leveraging our atlas, we identified HPV- and sex-specific shifts in immune and stromal composition, with HPV+ tumors enriched in adaptive immune cells and HPV- tumors showing more stromal and myeloid populations. Preliminary sex-stratified analyses suggested distinct microenvironmental patterns, warranting further investigation. This publicly available atlas provides a comprehensive framework for reproducibly studying HNSCC biology, improving patient stratification, and may help informing personalized therapies.

cancer biology↗

bioRxiv2025

Tumor Genotype Dictates Mitochondrial and Immune Vulnerabilities in Liver Cancer

Although oncogenic alterations influence tumor metabolism, how they impose distinct metabolic programs within a shared tissue context remains poorly defined. Here, we developed a rapid mitochondrial profiling platform to compare metabolites and proteins in genetic models of primary liver cancer (PLC). Analyses of six genetically distinct PLCs revealed that mitochondrial energy metabolism is largely dictated by oncogene identity. Kras-driven tumors required creatine metabolism to buffer energy demands during early tumorigenesis, whereas c-MYC-driven tumors relied on oxidative phosphorylation. Among c-MYC-driven PLCs, Pten-deficient tumors accumulated mitochondrial phosphoethanolamine, a precursor for phosphatidylethanolamine (PE) synthesis. Inhibition of PE synthesis selectively impaired the growth of Pten-deficient tumors and extended survival, in part through enhanced infiltration of CD8 T cells and sensitization to TNF-mediated cytotoxicity. Mechanistically, loss of PE elevated surface TNF receptor 2 (TNFR2), promoting TNF signaling and pro-inflammatory response. These findings uncover genotype-specific mitochondrial metabolic liabilities and establish PE synthesis as a tumor-intrinsic mechanism of immune evasion in PLC.

cancer biology↗

bioRxiv2025

Deep Learning Predicts Survival Across Squamous Tumor Entities From H&E Stains: Insights from Head and Neck, Esophagus, Lung and Cervical Cancer

Computational pathology-based models are becoming increasingly popular for extracting biomarkers from images of cancer tissue. However, their validity is often only demonstrated on a single unseen validation cohort, limiting insights into their generalizability and posing challenges for explainability. In this study, we developed models to predict overall survival using haematoxylin and eosin (H&E) slides from FFPE samples in head and neck squamous cell carcinoma (HNSCC). By validating our models across diverse squamous tumor entities, including, head and neck (HR = 1.58, 95% CI = 1.17-2.12, p = 0.003), esophageal (non-significant), lung (HR = 1.31, 95% CI = 1.13-1.52, p < 0.001) and cervical (HR = 1.39, 95% CI = 1.10-1.75, p = 0.005) squamous cell carcinomas, we showed that the predicted risk score captures relevant information for survival beyond HNSCC. Correlation analysis indicated that the predicted risk score is strongly associated with various clinical factors, including human papillomavirus status, tumor volume and smoking history, although the specific factors vary across cohorts. These results emphasize the necessity for comprehensive validation and in-depth assessment of computational pathology-based models to better characterize the underlying patterns they learn during training.

pathology↗

bioRxiv2024

Molecular architecture of human dermal sleeping nociceptors

Human dermal sleeping nociceptors display ongoing activity in neuropathic pain, affecting 10% of the population. Despite advances in rodents, a molecular marker for these mechano-insensitive C-fibers (CMis) in human skin remains elusive, preventing targeted therapy. In this translational Patch-seq study, we combine single-cell transcriptomics following electrophysiological characterization with single-nucleus and spatial transcriptomics from pigs and humans. We functionally identified CMis in pig sensory neurons with patch-clamp using adapted protocols from human microneurography. We identified oncostatin-M-receptor (OSMR) and somatostatin (SST) as marker genes for CMis. Following dermal injection in healthy human volunteers, oncostatin-M, the ligand of OSMR, exclusively modulates CMis. We identified the entire molecular architecture of human dermal sleeping nociceptors, providing new therapeutic targets and the basis for a mechanistic understanding of neuropathic pain. One Sentence SummaryWe identify the molecular architecture and specifically OSMR and SST as molecular markers for human dermal sleeping nociceptors, key players in the generation of neuropathic pain. Short versionIn this Patch-seq study, we identify OSMR and SST as molecular markers for human dermal sleeping nociceptors, key players in the generation of neuropathic pain.

neuroscience↗

bioRxiv2024

Utilizing protein structure graph embeddings to predict the pathogenicity of missense variants

BackgroundGenetic variants can impact the structure of the corresponding protein, which can have detrimental effects on protein function. While the effect of protein truncating variants is often easier to evaluate, most genetic variants are missense variants. These variants are mostly single nucleotide variants which result in the exchange of a single amino acid. The effect on protein function of these variants can be challenging to deduce. To aid the interpretation of missense variants a variety of bioinformatic algorithms have been developed, yet current algorithms rarely directly use the protein structure as a feature to consider. ResultsWe developed a machine learning workflow that utilizes the protein-language-model ESMFold to predict the protein structure of missense variants, which is subsequently embedded using graph autoencoders. The generated embeddings are used in a classifier model which predicts pathogenicity. We provide evidence that the generated graph embeddings improve classification accuracy of a XGBoost pathogenicity predictor, which should lead to a wide applicability for human genetic diseases. Additionally, we explored different levels of abstraction of the graph embeddings and their influence on the classifier. Finally, we compare the utility of graph embeddings from different protein-folding models.

bioinformatics↗

bioRxiv2024

Mutational Constraint Analysis Workflow for Overlapping Short Open Reading Frames and Genomic Neighbours

Understanding the dark genome is a priority task following the complete sequencing of the human genome. Short open reading frames (sORFs) are a group of largely unexplored elements of the dark genome with the potential for being translated into microproteins. The definitive number of coding and regulatory sORFs is not known, however they could account for up to 1-2% of the human genome. This corresponds to an order of magnitude in the range of canonical coding genes. For a few sORFs a clinical relevance has already been demonstrated, but for the majority of potential sORFs the biological function remains unclear. A major limitation in predicting their disease relevance using large-scale genomic data is the fact that no population-level constraint metrics for genetic variants in sORFs are yet available. To overcome this, we used the recently released gno-mAD 4.0 dataset and analysed the constraint of a consensus set of sORFs and their genomic neighbours. We demonstrate that sORFs are mostly embedded into a moderately constraint genomic context, but within the gencode dataset we identified a subset of highly constrained sORFs comparable to highly constrained canonical genes.

bioinformatics↗

bioRxiv2024

CNVizard: a lightweight streamlit application for an interactive analysis of copy number variants

Methods to call, analyze and visualize copy number variations (CNVs) from massive parallel sequencing data have been widely adopted in clinical practice and genetic research. To enable a streamlined analysis of CNV data, comprehensive annotation and good visualization are indispensable. The ability to detect single exon CNVs is another important feature for genetic testing. Nonetheless, most available open-source tools come with limitations in at least one of these areas. One drawback is that available tools deliver data in an unstructured and static format which requires subsequent visualization and formatting efforts. Here we present CNVizard, a lightweight streamlit app which requires minimal computational knowledge, and which is compatible with widely used CNV processing tools (CNVkit and AnnotSV). CNVizard can process short- and long-read sequencing data and provides an intuitive webapp-like experience enabling an interactive visualization of CNV data.

bioinformatics↗

bioRxiv2024

An organoid platform reveals MEK-PARP co-targeting to enhance radiation response in rectal cancer

Locally advanced rectal cancer is usually treated by neoadjuvant chemoradiotherapy. However, tumor response rates to this treatment vary greatly. Thus, most patients do not reach a complete remission and have to undergo tumor resection. In the present study, we introduce a patient-derived rectal cancer organoid platform that reflects clinical radiosensitivity and use this to screen 1596 drug-radiation combinations. We identify inhibitors of RAS-MAPK signaling, especially MEK inhibitors, strongly synergizing with radiation response. Mechanistically, MEK inhibitors suppressed radiation-induced activation of RAS-MAPK signaling, and selectively downregulated the homologous recombination DNA repair pathway component RAD51, thereby achieving radio-enhancement. Through testing drug-drug-radiation combinations in organoids and cell lines, we identified synergism between PARP and MEK inhibitors to further enhance the effect of radiation. Our data support clinical testing of combined MEK and PARP inhibition with radiotherapy in locally advanced rectal cancers. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=82 SRC="FIGDIR/small/597640v1_ufig1.gif" ALT="Figure 1"> View larger version (34K): org.highwire.dtl.DTLVardef@1d77d62org.highwire.dtl.DTLVardef@68b928org.highwire.dtl.DTLVardef@154142eorg.highwire.dtl.DTLVardef@f8f462_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2022

Single cell, whole embryo phenotyping of pleiotropic disorders of mammalian development

Mouse models are a critical tool for studying human diseases, particularly developmental disorders, as well as for advancing our general understanding of mammalian biology. However, it has long been suspected that conventional approaches for phenotyping are insufficiently sensitive to detect subtle defects throughout the developing mouse. Here we set out to establish single cell RNA sequencing (sc-RNA-seq) of the whole embryo as a scalable platform for the systematic molecular and cellular phenotyping of mouse genetic models. We applied combinatorial indexing-based sc-RNA-seq to profile 101 embryos of 26 genotypes at embryonic stage E13.5, altogether profiling gene expression in over 1.6M nuclei. The 26 genotypes include 22 mouse mutants representing a range of anticipated severities, from established multisystem disorders to deletions of individual enhancers, as well as the 4 wildtype backgrounds on which these mutants reside. We developed and applied several analytical frameworks for detecting differences in composition and/or gene expression across 52 cell types or trajectories. Some mutants exhibited changes in dozens of trajectories (e.g., the pleiotropic consequences of altering the Sox9 regulatory landscape) whereas others showed phenotypes affecting specific subsets of cells. We also identify differences between widely used wildtype strains, compare phenotyping of gain vs. loss of function mutants, and characterise deletions of topological associating domain (TAD) boundaries. Intriguingly, even among these 22 mutants, some changes are shared by heretofore unrelated models, suggesting that developmental pleiotropy might be "decomposable" through further scaling of this approach. Overall, our findings show how single cell profiling of whole embryos can enable the systematic molecular and cellular phenotypic characterization of mouse mutants with unprecedented breadth and resolution.

genomics↗

bioRxiv2021

Therapeutic targeting of SLC6A8 creatine transporter inhibits KRAS mutant and wildtype colon cancer and modulates human creatine levels

Colorectal cancer (CRC) is a leading cause of cancer mortality. Creatine metabolism was previously shown to critically regulate colon cancer progression. We report that RGX-202, an oral small-molecule SLC6A8 creatine transporter inhibitor, robustly inhibits creatine import in vitro and in vivo, reduces intracellular phosphocreatine and ATP levels and induces tumor cell apoptosis in CRC. RGX-202 suppressed tumor growth across KRAS wild-type and KRAS mutant xenograft, syngeneic and patient-derived xenograft colorectal cancers. Anti-tumor efficacy correlated with tumoral expression of creatine kinase B. Combining RGX-202 with 5- fluorouracil or the DHODH inhibitor leflunomide caused regressions of multiple colorectal xenograft and PDX tumors of distinct mutational backgrounds. RGX-202 also perturbed creatine metabolism in metastatic CRC patients enrolled in a Phase-1 trial, mirroring pharmacodynamic effects on creatine metabolism observed in mice. This is, to our knowledge, the first demonstration of pre-clinical and human pharmacodynamic activity for creatine metabolism targeting in oncology, revealing a critical target for CRC.

cancer biology↗

bioRxiv2021

Dissecting CD8+ T cell pathology of severe SARS-CoV-2 infection by single-cell epitope mapping

The current COVID-19 pandemic represents a global challenge. A better understanding of the immune response against SARS-CoV-2 is key to unveil the differences in disease severity and to develop future vaccines targeting novel SARS-CoV-2 variants. Feature barcode technology combined with CITE-seq antibodies and DNA-barcoded peptide-MHC I Dextramer reagents enabled us to identify relevant SARS-CoV-2-derived epitopes and compare epitope-specific CD8+ T cell populations between mild and severe COVID-19. We identified a strong CD8+ T cell response against an S protein-derived epitope. CD8+ effector cells in severe COVID-19 displayed hyperactivation, T cell exhaustion and were missing characteristics of long-lived memory T cells. We identify A*0101 WTAGAAAYY as an immunogenic CD8+ T cell epitope with the ability to drive clonal expansion. We provide an in-depth characterization of the CD8+ T cell-mediated response to SARS-CoV-2 infection which will be relevant for the development of molecular and targeted therapies and potential adjustments of vaccination strategies.

immunology↗

bioRxiv2020

High complexity cellular barcoding and clonal tracing reveals stochastic and deterministic parameters of radiation resistance

The impact of functional heterogeneity in response to radiation therapy is poorly understood to the present. It remains elusive whether clonal selection of tumor cells in response to ionizing radiation (IR) is a deterministic or stochastic process. We applied high-resolution lentiviral cellular barcoding for quantitative clonal tracking and deconvolution of clonal dynamics in response to IR. Clonal tracking of over 400.000 HNSCC patient-derived tumor cells and the analyses of over 1500 million sequencing reads in clonogenic survival assays reveals that fractionated IR induced a strong selective pressure for clonal reduction. This significantly exceeded uniform clonal survival probabilities indicative for a strong clone-to clone difference within tumor cell lines. IR induced clonal reduction affected the majority of tumor cells ranging between 96-75% and correlated to the degree of radiation sensitivity. Survival and clonogenicity is characterized by an intensive clonal distortion and dominance of individual tumor cells. Survival to IR is driven by a deterministic clonal selection of a smaller population which commonly survives radiation, while increased clonogenic capacity is a result of clonal competition of cells which have been selected stochastically. A 2-fold increase in radiation resistance results in a 4-fold (p<0.05) higher deterministic clonal selection showing that the ratio of these parameters is amenable to radiation sensitivity which correlates to prognostic biomarkers of HNSCC. Evidence for the existence of a rare subpopulation with an intrinsically radiation resistant phenotype commonly surviving IR was found at a frequency of 0.6-3.3% (p<0.001, FDR 3%). With cellular barcoding we introduce a novel functional heterogeneity associated qualitative readout for tracking dynamics of clonogenic survival in response to radiation. This enables the quantification of intrinsically radiation resistant tumor cells from patient samples and reveals the contribution of stochastic and deterministic clonal selection processes in response to IR.

systems biology↗

bioRxiv2019

Homozygous NMNAT2 mutation in sisters with polyneuropathy and erythromelalgia

We identified a homozygous missense mutation in the gene encoding NAD synthesizing enzyme NMNAT2 in two siblings with childhood onset polyneuropathy with erythromelalgia. No additional homozygotes for this rare allele, which leads to amino acid substitution T94M, were present among the unaffected relatives tested or in the 60,000 exomes of the ExAC database. For axons to survive, axonal NMNAT2 activity has to be maintained above a threshold level but the T94M mutation confers a partial loss of function both in the ability of NMNAT2 to support axon survival and in its enzymatic properties. Electrophysiological tests and histological analysis of sural nerve biopsies in the patients were consistent with loss of distal sensory and motor axons. Thus, it is likely that NMNAT2 mutation causes this pain and axon loss phenotype making this the first disorder associated with mutation of a key regulator of Wallerian-like axon degeneration in humans. This supports indications from numerous animal studies that the Wallerian degeneration pathway is important in human disease and raises important questions about which other human phenotypes could be linked to this gene.

neuroscience↗

bioRxiv2018

Aberrant DEGS1 sphingolipid metabolism impairs central and peripheral nervous system function in humans

Sphingolipids including ceramides are important components of cellular membranes and functionally associated with fundamental processes such as cell differentiation, neuronal signaling and myelin sheath formation. Defects in the synthesis or degradation of sphingolipids are associated with various neurological pathologies, however, the entire spectrum of disorders affecting sphingolipid metabolism remains elusive. By whole-exome sequencing in a patient with a multisystem neurological disorder of both the central and peripheral nervous system, we identified a homozygous variant p.(Ala280Val) in DEGS1, encoding an enzyme of the ceramide synthesis pathway. The blood sphingolipid profile and patient-derived fibroblasts both showed a significant shift from the unsaturated to the dihydro-forms of sphingolipids. Moreover, an atypical and potentially toxic sphingolipid metabolite is formed as consequence of the altered synthesis pathway. The changes in the sphingolipid profile were recapitulated in a CRISPR/Cas-based DEGS1 knockout HAP1-cell model and by chemical inhibition of DEGS1, suggesting a loss of DEGS1 function in the disease. DEGS1 insufficiency is thus a novel cause for a multisystem neurological disorder. A sphingolipid-rich diet may correct the metabolic profile and improve the clinical outcome of affected individuals and suggests that this heritable condition might be treatable.\n\nAbbreviations

genetics↗

bioRxiv2017

Sensory neuropathy-causing mutations in ATL3 cause aberrant ER membrane tethering

The ER is a complex network of sheets and tubules that is continuously being remodeled. The relevance of this membrane dynamics is underscored by the fact that mutations in Atlastins (ATL), the ER fusion proteins in mammals, cause neurodegeneration. How defects in this process disrupt neuronal homeostasis is largely unknown. Here we show by EM volume reconstruction of transfected cells, neurons and patient fibroblasts that the HSAN-causing ATL3 mutants promote aberrant ER tethering hallmarked by bundles of laterally attached ER tubules. In vitro, these mutants cause excessive liposome tethering, recapitulating the results in cells. Moreover, ATL3 variants retain their dimerization-dependent GTPase activity, but are unable to promote membrane fusion, suggesting a defect on an intermediate step of the ATL3 functional cycle. Our data therefore show that the effects of ATL3 mutations on ER network organization stretch beyond a loss of fusion, shedding a new light on neuropathies caused by atlastin defects.

cell biology↗