Biology subjects

Rademacher, A.

Publications and source records attributed to Rademacher, A..

10 recordsLinked to original sources

bioRxiv2026

clinTALL: machine learning-driven multimodal subtypeclassification and treatment outcome prediction in pediatric T-ALL

BackgroundChildhood T-lineage acute lymphoblastic leukemia (T-ALL) is an aggressive hematologic malignancy with poor prognosis. Differently from B-cell precursor ALL, T-ALL lacks effective risk stratification strategies. A recent study has integrated whole genome and whole transcriptome data to define over 15 distinct molecular subtypes with prognostic significance. However, clinical translation of this knowledge remains challenging due to the complexity of interpreting high-dimensional multi-omics-based data. MethodsHere, we present clinTALL, a deep learning based multi-task pipeline for pediatric T-ALL subtype classification and treatment outcome estimation. The model integrates multimodal input data and uses a neural network architecture to generate a shared latent embedding for jointly learned multi-task prediction. The competing risk-based model was used to predict event-specific outcomes. The model was trained on a publicly available multimodal dataset comprising clinical, genomic and transcriptomic features of 1309 pediatric T-ALL samples. ResultsWe observed that the transcriptomic-only model achieved superior single modality results, with 92.2% accuracy for subtype prediction and a 65.9% concordance index (C-index) for event-free survival (EFS) in a cross-validation setup. Integrating all data modalities maintained high subtype classification accuracy (91.7%) and improved the overall concordance index for EFS estimation to 67.5%. The competing risk-based model enables accurate predictions of induction failure (C-index = 96.0%) and second malignant neoplasm (C-index = 62.1%). We validated molecular subtype predictions on an internal dataset of 120 pediatric T-ALL samples and obtained an accuracy of 81.8%. To facilitate the broad application of multi-omics based subtype prediction and treatment outcome inference, we provide clinTall as a Docker based application, allowing for user friendly access to the tool. The full source code of clinTALL is available on GitHub (https://github.com/UKWgenommedizin/clinTALL). ConclusionTogether, our machine learning-based framework allows for automated, accurate sub-type classification and treatment outcome inference using multimodal input data, advancing precision risk stratification for pediatric T-ALL.

bioinformatics↗

bioRxiv2025

Intra-subtype heterogeneity shapes treatment response in KMT2A-rearranged ALL across all age groups

BackgroundKMT2A-rearranged B-cell acute lymphoblastic leukemia (KMT2Ar B-ALL) exhibits significant heterogeneity in age of onset, developmental origins, and clinical outcomes. The interplay of individual factors influencing early treatment response within this high-risk molecular subtype remains poorly elucidated. We aimed to comprehensively assess how leukemic developmental state, fusion partner, and patient age jointly influence early treatment response and drug sensitivity. MethodsTo identify determinants of early treatment response to induction chemotherapy, we analysed 465 KMT2Ar B-ALL cases spanning a wide age range (1 month to 89 years) by integrating transcriptomic and genomic profiling with functional drug response and measurable residual disease (MRD) kinetics. Transcriptomic profiling was used to derive a developmental maturity score based on proximity to physiological B-cell differentiation from a normal B-lymphopoiesis reference. Using an ordinal regression model we identify gene expression programs associated with early MRD response. ResultsWe observed a strong inverse correlation between MRD clearance with advancing age (p=2.1E-04), proximity to early B-cell-precursor developmental state (low maturity score, p=1.3E-03) and AFF1 as fusion partner (p=7.0E-04). A combined model confirmed the predominate impact of both maturity and KMT2A fusion partner on MRD response, supporting the concept that the cells developmental state defines therapy response. Gene expression analysis identified cellular traits that relate to MRD response (e.g. chromatin organization, immune modulation and proliferation). This gene expression classifier grouped cases by MRD response but also by ex-vivo induction drug sensitivity. Notably, good responders to ex-vivo induction drugs were characterized by a higher maturity score (p=1.8E-03), whereas for less mature KMT2Ar B-ALL cases response profiles suggested higher Venetoclax sensitivity. ConclusionsOur study provides an integrative framework linking developmental phenotype, fusion partner, and MRD kinetics across the full age spectrum of KMT2Ar B-ALL. The maturity score, derived from bulk transcriptome data, offers a biologically relevant predictor of early treatment response and drug sensitivity. These insights may support future risk-adapted strategies and therapeutic targeting, particularly in immature KMT2Ar B-ALL.

cancer biology↗

bioRxiv2024

Two distinct chromatin modules regulate proinflammatory gene expression

Various mechanisms have been proposed to explain gene activation and co-regulation, including enhancer-promoter interactions via chromatin looping and the enrichment of transcription factors into hubs or condensates. However, these conclusions often stem from analyses of individual loci, and genome-wide studies exploring mechanistic differences with coupled gene expression are lacking. In this study, we dissected the proinflammatory gene expression program induced by TNF in primary human endothelial cells using NGS- and imaging-based techniques. Our findings, enabled by our novel RWireX approach for single-cell ATAC-seq analysis, revealed two distinct regulatory chromatin modules: autonomous links of co-accessibility (ACs) between separated sites, and domains of contiguous co-accessibility (DCs) with increased local transcription factor binding. Genes in ACs and DCs exhibited different transcriptional bursting kinetics, highlighting the existence of two structurally and functionally distinct regulatory chromatin modules in the proinflammatory response. These findings provide a novel mechanistic framework for understanding how cells achieve rapid and precise gene expression control. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=71 SRC="FIGDIR/small/606159v1_ufig1.gif" ALT="Figure 1"> View larger version (30K): org.highwire.dtl.DTLVardef@1f5f7e8org.highwire.dtl.DTLVardef@166b4d7org.highwire.dtl.DTLVardef@1ea9206org.highwire.dtl.DTLVardef@118d534_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LITwo distinct, non-mutually exclusive chromatin modules, ACs and DCs, that regulate proinflammatory gene expression were identified based on deep scATAC-seq. C_LIO_LIACs represent long-range genomic interactions with regulation occurring more by transcription burst frequency. C_LIO_LIDCs are regions of increased local transcription factor binding that can modulate transcription burst size. C_LIO_LIThe AC/DC model integrates sequencing-based evidence for chromatin looping with microscopy observations of transcription factor hubs/condensates into a unified model. C_LIO_LIOur findings provide a novel framework for understanding how cells achieve rapid and precise gene expression control. C_LI

genomics↗

bioRxiv2024

Comparison of spatial transcriptomics technologies used for tumor cryosections

BackgroundSpatial transcriptomics (ST) technologies are revolutionizing our understanding of intra-tumor heterogeneity and the tumor microenvironment by revealing single-cell molecular profiles within their spatial tissue context. The rapid evolution of ST methods, each with unique features, presents a challenge in selecting the most appropriate technology for specific research objectives. Here, we compare four imaging-based ST methods - RNAscope HiPlex, Molecular Cartography, MERFISH/Merscope, and Xenium - together with sequencing-based ST (Visium). These technologies were used to study cryosections of medulloblastoma with extensive nodularity (MBEN), a tumor chosen for its distinct microanatomical features. ResultsOur analysis reveals that automated imaging-based ST methods are well suited to delineating the intricate MBEN microanatomy, capturing cell-type-specific transcriptome profiles. We devise approaches to compare the sensitivity and specificity of the different methods together with their unique attributes to guide method selection based on the research aim. Furthermore, we demonstrate how reimaging of slides after the ST analysis can markedly improve cell segmentation accuracy and integrate additional transcript and protein readouts to expand the analytical possibilities and depth of insights. ConclusionsThis study highlights key distinctions between various ST technologies and provides a set of parameters for evaluating their performance. Our findings aid in the informed choice of ST methods and delineate approaches for enhancing the resolution and breadth of spatial transcriptomic analyses, thereby contributing to advancing ST applications in solid tumor research.

genomics↗

bioRxiv2024

Medulloblastoma oncogene aberrations are not involved in tumor initiation, but essential for disease progression and therapy resistance

Despite recent advances in understanding disease biology, treatment of Group 3/4 medulloblastoma remains a therapeutic challenge in pediatric neuro-oncology. Bulk-omics approaches have identified considerable intertumoral heterogeneity in Group 3/4 medulloblastoma, including the presence of clear single-gene oncogenic drivers in only a subset of cases, whereas in the majority of cases, large-scale copy-number aberrations prevail. However, intratumoral heterogeneity, the role of oncogene aberrations, and broad CNVs in tumor evolution and treatment resistance remain poorly understood. To dissect this interplay, we used single-cell technologies (snRNA-seq, snATAC-seq, spatial transcriptomics) on a cohort of Group 3/4 medulloblastoma with known alterations in the oncogenes MYC, MYCN, and PRDM6. We show that large-scale chromosomal aberrations are early tumor initiating events, while the single-gene oncogenic events arise late and are typically sub-clonal, but MYC can become clonal upon disease progression to drive further tumor development and therapy resistance. We identify that the subclones are mostly interspersed across tumor tissue using spatial transcriptomics, but clear segregation is also present. Using a population genetics model, we estimate medulloblastoma initiation in the cerebellar unipolar brush cell-lineage starting from the first gestational trimester. Our findings demonstrate how single-cell technologies can be applied for early detection and diagnosis of this fatal disease.

cancer biology↗

bioRxiv2024

Gene regulatory network landscape of Group 3/4 medulloblastoma

Cellular heterogeneity in Group 3 and Group 4 medulloblastomas is a major driver of therapeutic intractability. Here, we describe transcription factor-driven mechanisms underlying cellular diversity in these tumors using a comprehensive single-nucleus multi-omics atlas. Our analysis reveals that rather than fixed entities, these tumors exist along a continuum of cell states defined by four molecular identity axes. We show that perturbing transcription factor activity drives cellular plasticity, a key contributor to tumor heterogeneity. Strikingly, modulation of the lineage determinant PAX6 redirects tumor cells along both Group 3- and Group 4-like differentiation trajectories, modeling a bi-lineage medulloblastoma, and reduces aggressiveness of MYC-driven tumor models. Together, our findings reveal how oncogenic signals co-opt cerebellar unipolar brush cell programs to rewire tumor cell states and identify cellular plasticity as a potential targetable vulnerability in medulloblastoma.

cancer biology↗

bioRxiv2022

Compartments in medulloblastoma with extensive nodularity are connected through differentiation along the granular precursor lineage

Medulloblastoma with extensive nodularity (MBEN) are cerebellar tumors with two histologically distinct compartments and varying disease course. In some children MBEN progresses, while others show spontaneous differentiation into more benign tumors. However, the mechanisms that control the tug-of-war between proliferation and differentiation are not well understood. Here, we dissected this process with a multi-modal single cell transcriptome analysis. We found that the internodular MBEN compartment comprised proliferating early cerebellar granular neuronal precursors (CGNP)-like tumor cells as well as stromal, vascular, and immune cells. In contrast, the nodular compartment consisted of postmitotic, neuronally differentiated MBEN cells. Both compartments were connected through an intermediate cell stage of actively migrating CGNPs. Furthermore, astrocyte-like tumor cells were identified that had branched off the main CGNP developmental trajectory. Cells with an astroglial phenotype were found in close proximity to migrating, late CGNP-like and postmitotic neuronally differentiated cells. Our study reveals how the spatial tissue organization is linked to the developmental trajectory of proliferating tumor cells through a migrating precursor stage into differentiated tumor cells with a more benign phenotype. We anticipate that our framework for integrating single nucleus RNA-sequencing and spatial transcriptomics will help to uncover intercompartmental interactions also in other cancers with varying histology.

cancer biology↗

bioRxiv2021

Transcription activation is enhanced by multivalent interactions independent of liquid-liquid phase separation

Transcription factors (TFs) consist of a DNA binding and an activation domain (AD) that are considered to be independent and exchangeable modules. However, recent studies conclude that also the physico-chemical properties of the AD can control TF assembly at chromatin by driving a phase separation into transcriptional condensates. Here, we dissected transcription activation by comparing different synthetic TFs at a reporter gene array with real-time single-cell fluorescence microscopy readouts. In these experiments, binding site occupancy, residence time and co-activator recruitment in relation to multivalent TF interactions were compared. While phase separation propensity and activation strength of the AD were correlated, the actual formation of liquid-like TF droplets had a neutral or inhibitory effect on transcription activation. Rather, we conclude that multivalent AD mediated interactions increase the transcription activation capacity of a TF by stabilizing chromatin binding and mediating the recruitment of co-activators independent of phase separation.

biophysics↗

bioRxiv2020

DNA sequence-dependent formation of heterochromatin nanodomains

The mammalian epigenome contains thousands of heterochromatin nanodomains (HNDs) marked by di- and trimethylation of histone H3 at lysine 9, which have a typical size of 3-10 nucleosomes. However, the (epi)genetic determinants of their location and boundaries are only partly understood. Here, we compare four HND types in mouse embryonic stem cells, that are defined by histone methylases SUV39H1/2 or GLP, transcription factor ADNP or chromatin remodeller ATRX. Based on a novel chromatin hierarchical lattice framework termed ChromHL, we are able to predict HND maps with singe-nucleotide resolution. We find that HND nucleation can be rationalized by DNA sequence specific protein binding to PAX3/9, ADNP and LINE1 repeats. Depending on type of microdomains, boundaries are determined either by CTCF binding sites or by nucleosome-nucleosome and nucleosome-HP1 interactions. Our new framework allows predicting how patterns of H3K9me2/3 and other chromatin nanodomains are established and changed in processes such as cell differentiation.

genomics↗

bioRxiv2017

Real-Time Observation Of Light-Controlled Transcription In Living Cells

Gene expression is a tightly controlled process that is coordinated in space and time. To dissect its dynamic regulation with high temporal resolution, we introduce an optogenetic tool termed BLInCR (Blue Light-Induced Chromatin Recruitment) that combines rapid and reversible light-dependent recruitment of effector proteins with a real-time readout for transcription. We used BLInCR to control the activity of a reporter gene cluster in the human osteosarcoma cell line U2OS by reversibly recruiting the viral transactivator VP16. RNA production was detectable [~]2 minutes after VP16 recruitment and readily decreased when VP16 dissociated from the cluster in the absence of light. Quantitative assessment of the activation process revealed biphasic activation kinetics with a pronounced early phase in cells treated with the histone deacetylase inhibitor SAHA. Comparison with kinetic models for transcription activation suggests that the gene cluster undergoes a maturation process when activated. BLInCR will facilitate the study of transcription dynamics in living cells.

cell biology↗