Biology subjects

Stegle, O.

Publications and source records attributed to Stegle, O..

At least 37 records · Page 2Linked to original sources

bioRxiv2024

Escape from X inactivation is directly modulated by levels of Xist non-coding RNA

In placental females, one copy of the two X chromosomes is largely silenced during a narrow developmental time window, in a process mediated by the non-coding RNA Xist 1. Here, we demonstrate that Xist can initiate X-chromosome inactivation (XCI) well beyond early embryogenesis. By modifying its endogenous level, we show that Xist has the capacity to actively silence genes that escape XCI both in neuronal progenitor cells (NPCs) and in vivo, in mouse embryos. We also show that Xist plays a direct role in eliminating TAD-like structures associated with clusters of escapee genes on the inactive X chromosome, and that this is dependent on Xists XCI initiation partner, SPEN 2. We further demonstrate that Xists function in suppressing gene expression of escapees and topological domain formation is reversible for up to seven days post-induction, but that sustained Xist up-regulation leads to progressively irreversible silencing and CpG island DNA methylation of facultative escapees. Thus, the distinctive transcriptional and regulatory topologies of the silenced X chromosome is actively, directly - and reversibly - controlled by Xist RNA throughout life.

developmental biology↗

bioRxiv2023

Four-Core Genotypes mice harbour a 3.2MB X-Y translocation that perturbs Tlr7 dosage

The Four Core Genotypes (FCG) is a mouse model system heavily used to disentangle the function of sex chromosomes and hormones. We report that a copy of a 3.2 MB region of the X chromosome has translocated to the YSry- chromosome and thus increased the expression of multiple genes including the auto-immune master regulator Tlr7. This previously-unreported X-Y translocation complicates the interpretation of studies reliant on FCG mice.

genetics↗

bioRxiv2023

HIPSD&R-seq enables scalable genomic copy number and transcriptome profiling

Single-cell DNA-sequencing (scDNA-seq) enables decoding somatic cancer variation. Existing methods are hampered by low throughput or cannot be combined with transcriptome sequencing in the same cell. We propose HIPSD&R-seq (HIgh-throughPut Single-cell Dna and Rna-seq), a scalable yet simple assay to profile low-coverage DNA and RNA in thousands of cells in parallel. Our approach builds on an accessible modification of the 10X Genomics platform for scATAC and multiome profiling. In applications to human cell models and primary tissue, we demonstrate the feasibility to detect rare clones and we combine the assay with combinatorial indexing to profile over 16,000 cells.

cancer biology↗

bioRxiv2023

Model-based inference of RNA velocity modules improves cell fate prediction

RNA velocity is a powerful paradigm that exploits the temporal information contained in spliced and unspliced RNA counts to infer transcriptional dynamics. Existing velocity models either rely on coarse biophysical simplifications or require extensive numerical approximations to solve the underlying differential equations. This results in loss of accuracy in challenging settings, such as complex or weak transcription rate changes across cellular trajectories. Here, we present cell2fate, a formulation of RNA velocity based on a linearization of the velocity ODE, which allows solving a biophysically accurate model in a fully Bayesian fashion. As a result, cell2fate decomposes the RNA velocity solutions into modules, which provides a new biophysical connection between RNA velocity and statistical dimensionality reduction. We comprehensively benchmark cell2fate in real-world settings, demonstrating enhanced interpretability and increased power to reconstruct complex dynamics and weak dynamical signals in rare and mature cell types. Finally, we apply cell2fate to a newly generated dataset from the developing human brain, where we spatially map RNA velocity modules onto the tissue architecture, thereby connecting the spatial organisation of tissues with temporal dynamics of transcription.

genomics↗

bioRxiv2023

Integration of variant annotations using deep set networks boosts rare variant association genetics

Rare genetic variants can strongly predispose to disease, yet accounting for rare variants in genetic analyses is statistically challenging. While rich variant annotations hold the promise to enable well-powered rare variant association tests, methods integrating variant annotations in a data-driven manner are lacking. Here, we propose DeepRVAT, a model based on set neural networks that learns burden scores from rare variants, annotations, and phenotypes. In contrast to existing methods, DeepRVAT yields a single, trait-agnostic, nonlinear gene impairment score, enabling both risk prediction and gene discovery in a unified framework. On 34 quantitative and 26 binary traits, using whole-exome-sequencing data from UK Biobank, we find that DeepRVAT offers substantial increases in gene discoveries and improved replication rates in held-out data. Moreover, we demonstrate that the integrative DeepRVAT gene impairment score greatly improves detection of individuals at high genetic risk. Finally, we show that pre-trained DeepRVAT scores generalize across traits, opening up the possibility to conduct highly computationally efficient rare variant tests.

bioinformatics↗

bioRxiv2023

SpatialData: an open and universal data framework for spatial omics

Spatially resolved omics technologies are transforming our understanding of biological tissues. However, handling uni- and multi-modal spatial omics datasets remains a challenge owing to large volumes of data, heterogeneous data types and the lack of unified spatially-aware data structures. Here, we introduce SpatialData, a framework that establishes a unified and extensible multi-platform file-format, lazy representation of larger-than-memory data, transformations, and alignment to common coordinate systems. SpatialData facilitates spatial annotations and cross-modal aggregation and analysis, the utility of which is illustrated via multiple vignettes, including integrative analysis on a multi-modal Xenium and Visium breast cancer study.

bioinformatics↗

bioRxiv2023

XClone: detection of allele-specific subclonal copy number variations from single-cell transcriptomic data

Somatic copy number alterations (CNAs) are major mutations that contribute to the development and progression of various cancers. Despite a few computational methods proposed to detect CNAs from single-cell transcriptomic data, the technical sparsity of such data makes it challenging to identify allele-specific CNAs, particularly in complex clonal structures. In this study, we present a statistical method, XClone, that strengthens the signals of read depth and allelic imbalance by effective smoothing on cell neighborhood and gene coordinate graphs to detect haplotype-aware CNAs from scRNA-seq data. By applying XClone to multiple datasets with challenging compositions, we demonstrated its ability to robustly detect different types of allele-specific CNAs and potentially indicate whole genome duplication, therefore enabling the discovery of corresponding subclones and the dissection of their phenotypic impacts.

bioinformatics↗

bioRxiv2023

Spatial and temporal transcriptomics of SHH-medulloblastoma with chromothripsis identifies multiple genetic clones that resist to treatment and lead to relapse

Paediatric medulloblastomas with chromothripsis are characterised by high genomic instability and are among the tumours with the worst prognosis. However, the determinants of their aggressiveness and the molecular makeup of chromothriptic medulloblastoma are not well understood. Here, we applied spatial transcriptomics to profile a cohort of 13 chromothriptic and non-chromothriptic medulloblastomas from the same molecular subgroup. Our data reveal a higher extent of spatial intra-tumour heterogeneity in chromothriptic medulloblastomas, which is associated with increased proliferation and stemness, but lower immune infiltration and differentiation. Spatial mapping of genetic subclones of the same tumour identify a regionally distinct architecture and clone-specific phenotypic features, with distinct degrees of differentiation, proliferation and immune infiltration between clones. We conducted temporal profiling of 11 patient-derived xenografts from chromothriptic medulloblastomas, covering the transition from the minimal residual disease stage to treatment-resistant regrown tumours. In chromothriptic medulloblastoma, an ecosystem of cells from multiple genetic clones resisting treatment and leading to relapse highlighted the importance of multi-clone interplay. Finally, we identified a potential role for tumour microtubes in treatment resistance in chromothriptic medulloblastoma, suggesting cell network communication as a putative target. HighlightsO_ST_ABSBiological insightsC_ST_ABSO_LIMedulloblastomas with chromothripsis are characterised by higher spatial intra-tumour heterogeneity, proliferation and stemness, but lower immune infiltration and differentiation, as compared with non-chromothriptic medulloblastomas C_LIO_LISpatially aware genetic clone assignment identifies phenotypic features such as degree of differentiation, proliferation and immune infiltration enriched in specific clones C_LIO_LICells from distinct genetic clones resist to treatment and give rise to relapse in patient-derived xenografts of medulloblastoma with chromothripsis C_LIO_LITumour microtubes point to putative molecular processes implicated in treatment resistance in chromothriptic medulloblastoma C_LI Technological advancesO_LIValidation of the spatial representation of patient-derived xenograft models, providing novel insights into the faithfulness of such models to study cancer C_LIO_LISpatial mapping of clones based on spatial transcriptomics and copy-number variant profiling C_LI

cancer biology↗

bioRxiv2023

Single cell DNA methylation ageing in mouse blood

Ageing is the accumulation of changes and overall decline of the function of cells, organs and organisms over time. At the molecular and cellular level, the concept of biological age has been established and biomarkers of biological age have been identified, notably epigenetic DNA-methylation based clocks. With the emergence of single-cell DNA methylation profiling methods, the possibility to study biological age of individual cells has been proposed, and a first proof-of-concept study, based on limited single cell datasets mostly from early developmental origin, indicated the feasibility and relevance of this approach to better understand organismal changes and cellular ageing heterogeneity. Here we generated a large single-cell DNA methylation and matched transcriptome dataset from mouse peripheral blood samples, spanning a broad range of ages (10-101 weeks of age). We observed that the number of genes expressed increased at older ages, but gene specific changes were small. We next developed a robust single cell DNA methylation age predictor (scEpiAge), which can accurately predict age in a broad range of publicly available datasets, including very sparse data and it also predicts age in single cells. Interestingly, the DNA methylation age distribution is wider than technically expected in 19% of single cells, suggesting that epigenetic age heterogeneity is present in vivo and may relate to functional differences between cells. In addition, we observe differences in epigenetic ageing between the major blood cell types. Our work provides a foundation for better single-cell and sparse data epigenetic age predictors and highlights the significance of cellular heterogeneity during ageing. Highlights- Model to estimate DNA methylation age in single cells - Large multi-omics dataset of single cells from murine blood - Epigenetic age deviations from chronological age are greater than technical expected from technical variability - Number of genes expressed increases with chronological and epigenetic age

cell biology↗

bioRxiv2022

Spatially resolved single-cell multiomics map of human trophoblast differentiation in early pregnancy

The relationship between the human placenta, the extraembryonic organ built by the fetus, and the decidua, the mucosal layer of the uterus, is essential to nurture and protect the fetus during pregnancy. Extravillous trophoblast cells (EVTs) anchor the placenta and infiltrate the decidua, transforming the maternal arteries into high conductance vessels. Defects in trophoblast invasion and arterial transformation established during early pregnancy underlie common pregnancy disorders such as pre-eclampsia. Despite its importance, how EVT invasion is regulated in humans is still unclear due the inaccessibility of the entire pregnant uterus and, until recently, a lack of reliable in vitro models. Here, we have generated a spatially-resolved multiomics single-cell atlas of the entire maternal-fetal interface including the myometrium, allowing us to resolve the full trajectory of trophoblast differentiation. We have used this cellular map to elucidate the main regulatory programmes mediating EVT invasion and show that they are preserved in trophoblast organoids. We define the transcriptomes of the final cell states of trophoblast invasion: placental bed giant cells (fused multinucleated EVTs) and endovascular EVTs (which form plugs inside the maternal arteries). We reconstruct the cell-cell communication events contributing to trophoblast invasion and GC formation, and define the dual role of interstitial EVTs and endovascular EVTs in mediating arterial transformation during early pregnancy. Together, our data provides a comprehensive analysis of postimplantation trophoblast differentiation in humans that can be used as a blueprint to design accurate multilineage placental in vitro models.

genomics↗

bioRxiv2022

Mutation-Attention (MuAt): deep representation learning of somatic mutations for tumour typing and subtyping

Cancer genome sequencing enables accurate classification of tumours and tumour sub-types. However, prediction performance is still limited using exome-only sequencing and for tumor types with low somatic mutation burden such as many pediatric tumours. Moreover, the ability to leverage deep representation learning in discovery of tumour entities remains unknown. We introduce here Mutation-Attention (MuAt), a deep neural network to learn representations of simple and complex somatic alterations for prediction of tumour types and subtypes. MuAt achieved prediction accuracy of 89% for whole genomes (24 tumour types) and 64% for whole exomes (20 types), and a top-5 accuracy of 97% and 90%, respectively. Tumour representations learnt by MuAt included tumour entities such as acral melanoma, SHH-activated medulloblastoma, SPOP-associated prostate cancer, microsatellite instability, and MUTYH-associated pancreatic endocrine tumours although these tumour subtypes and subgroups were not used as training labels. Integrated representations of somatic alterations hold significant potential to drive discovery of novel tumour entities and clinical application.

genomics↗

bioRxiv2022

Long-read sequencing of diagnosis and post-therapy medulloblastoma reveals complex rearrangement patterns and epigenetic signatures

Cancer genomes harbor a broad spectrum of structural variants (SV) driving tumorigenesis, a relevant subset of which are likely to escape discovery in short reads. We employed Oxford Nanopore Technologies (ONT) sequencing in a paired diagnostic and post-therapy medulloblastoma to unravel the haplotype-resolved somatic genetic and epigenetic landscape. We assemble complex rearrangements and such associated with telomeric sequences, including a 1.55 Megabasepair chromothripsis event. We uncover a complex SV pattern termed templated insertion thread, characterized by short (mostly <1kb) insertions showing prevalent self-concatenation into highly amplified structures of up to 50kbp in size. Templated insertion threads occur in 3% of cancers, with a prevalence ranging to 74% in liposarcoma, and frequent colocalization with chromothripsis. We also perform long-read based methylome profiling and discover allele-specific methylation (ASM) effects, complex rearrangements exhibiting differential methylation, and differential promoter methylation in seven cancer-driver genes. Our study shows the potential of long-read sequencing in cancer. Graphical abstract O_FIG_DISPLAY_L [Figure 1] M_FIG_DISPLAY I) We investigate a single patient with chromothriptic sonic hedgehog medulloblastoma (Li-Fraumeni syndrome), with tissue samples taken from blood, the primary tumor at diagnosis, and a post-treatment (relapse) tumor. II) Data on the three samples has been collected from four sources, 1) Illumina whole-genome, 2) Illumina transcriptome sequencing, 3) Illumina Infinium HumanMethylation450k, as well as 4) long-read whole-genome sequencing using Oxford Nanopore Technologies (ONT) sequencing. III) An integrative analysis combines genomic, epigenomic as well as transcriptomic data to provide a comprehensive analysis of this heavily rearranged tumor sample. Long and short read sequencing data is used to inform the analysis of complex structural genomic variants and methylation called from haplotyped ONT reads and validated through the methylation array data allows for a haplotype-resolved study of genomic and epigenomic variation, which can then be examined for transcriptional effect. IV) This integrative analysis allows us to identify a large number of inter- and intra-chromosomal genomic rearrangements (A) including a complex rearrangement pattern we term templated insertion threads (B), as well as sample-specific and haplotype specific methylation patterns of known cancer genes (C). C_FIG_DISPLAY

cancer biology↗

bioRxiv2022

PycoMeth: A toolbox for differential methylation testing from Nanopore methylation calls

AO_SCPLOWBSTRACTC_SCPLOWAdvances in base and methylation calling of Oxford Nanopore Technologies (ONT) sequencing data have opened up the possibility for joint profiling of genomic and epigenetic variation on the same long reads. Existing data storage and analysis frameworks that were developed for CpG-methylation arrays or short-read bisulfite sequencing data have severe shortcomings for handling of ONT data, failing to fully exploit methylation profiles obtained from long read technologies. To address these issues, we present pycoMeth, a toolbox to store, manage and analyse DNA methylation data obtained from long-read ONT sequencing data. Our toolbox centers around a new storage format called MetH5, which allows simultaneously for efficient storage of and rapid data access for read-level and reference-anchored methylation call data. Building on this storage format, we propose efficient algorithms for the segmentation and differential methylation testing of methylation calls from ONT data. Our methods draw from read-group and read-level information, as well as methylation call uncertainties, and allow for de novo discovery of methylation patterns and differentially methylated regions in a haplotyped multi-sample setting. We show that MetH5 is more efficient than existing solutions for storing ONT methylation calls, and carry out benchmarking for segmentation and differential methylation analysis, demonstrating increased performance and sensitivity of pycoMeth compared to existing solutions.

bioinformatics↗

bioRxiv2021

Single-Cell Atlas of Common Variable Immunodeficiency reveals germinal center-associated epigenetic dysregulation in B cell responses

Common variable immunodeficiency (CVID), the most prevalent symptomatic primary immunodeficiency, is characterized by impaired terminal B-cell differentiation and defective antibody responses. Incomplete genetic penetrance and a wide range of phenotypic expressivity in CVID suggest the participation of additional pathogenic mechanisms. Monozygotic (MZ) twins discordant for CVID are uniquely valuable for studying the contribution of epigenetics to the disease. We used single-cell epigenomics and transcriptomics to create a cell census of naive-to-memory B cell differentiation in a pair of CVID-discordant MZ twins. Our analysis identifies DNA methylation, chromatin accessibility and transcriptional defects in memory B cells that mirror defective cell-cell communication defects following activation. These findings were validated in a cohort of CVID patients and healthy donors. Our findings provide a comprehensive multi-omics map of alterations in naive-to-memory B-cell transition in CVID and reveal links between the epigenome and immune cell cross-talk. Our resource, publicly available at the Human Cell Atlas, paves the way for future diagnosis and treatments of CVID patients.

immunology↗

bioRxiv2021

Epithelial cells of the intestine acquire cell-intrinsic inflammation signatures during ageing

During ageing, cell-intrinsic and extrinsic factors lead to the decline of tissue function and organismal health. Disentangling these factors is important for developing effective strategies to prolong organismal healthspan. Here, we addressed this question in the mouse intestinal epithelium, which forms a dynamic interface with its microenvironment and receives extrinsic signals affecting its homeostasis and tissue ageing. We systematically compared transcriptional profiles of young and aged epithelial cells in vivo and ex vivo in cultured intestinal organoids. We found that all cell types of the aged epithelium exhibit an inflammation phenotype, which is marked by MHC class II upregulation and most pronounced in enterocytes. This was accompanied by elevated levels of the immune tolerance markers PD-1 and PD-L1 in the aged tissue microenvironment, indicating dysregulation of immunological homeostasis. Intestinal organoids from aged mice still showed an inflammation signature after weeks in culture, which was concurrent with increased chromatin accessibility of inflammation-associated loci. Our results reveal a cell-intrinsic, persistent inflammation phenotype in aged epithelial cells, which might contribute to systemic inflammation observed during ageing.

genomics↗

bioRxiv2021

A spatial multi-omics atlas of the human lung reveals a novel immune cell survival niche

Multiple distinct cell types of the human lung and airways have been defined by single cell RNA sequencing (scRNAseq). Here we present a multi-omics spatial lung atlas to define novel cell types which we map back into the macro- and micro-anatomical tissue context to define functional tissue microenvironments. Firstly, we have generated single cell and nuclei RNA sequencing, VDJ-sequencing and Visium Spatial Transcriptomics data sets from 5 different locations of the human lung and airways. Secondly, we define additional cell types/states, as well as spatially map novel and known human airway cell types, such as adult lung chondrocytes, submucosal gland (SMG) duct cells, distinct pericyte and smooth muscle subtypes, immune-recruiting fibroblasts, peribronchial and perichondrial fibroblasts, peripheral nerve associated fibroblasts and Schwann cells. Finally, we define a survival niche for IgA-secreting plasma cells at the SMG, comprising the newly defined epithelial SMG-Duct cells, and B and T lineage immune cells. Using our transcriptomic data for cell-cell interaction analysis, we propose a signalling circuit that establishes and supports this niche. Overall, we provide a transcriptional and spatial lung atlas with multiple novel cell types that allows for the study of specific tissue microenvironments such as the newly defined gland-associated lymphoid niche (GALN).

cell biology↗

bioRxiv2021

FISHFactor: A Probabilistic Factor Model for Spatial Transcriptomics Data with Subcellular Resolution

MotivationFactor analysis is a widely used tool for unsupervised dimensionality reduction of high-throughput data sets in molecular biology, with recently proposed extensions designed specifically for spatial transcriptomics data. However, these methods expect (count) matrices as data input and are therefore not directly applicable to single molecule resolution data, which are in the form of coordinate lists annotated with genes and provide insight into subcellular spatial expression patterns. To address this, we here propose FISHFactor, a probabilistic factor model that combines the benefits of spatial, non-negative factor analysis with a Poisson point process likelihood to explicitly model and account for the nature of single molecule resolution data. In addition, FISHFactor shares information across a potentially large number of cells in a common weight matrix, allowing consistent interpretation of factors across cells and yielding improved latent variable estimates. ResultsWe compare FISHFactor to existing methods that rely on aggregating information through spatial binning and cannot combine information from multiple cells, and show that our method leads to more accurate results on simulated data. We demonstrate on a real data set that FISHFactor is able to identify major subcellular expression patterns and spatial gene clusters in a data-driven manner. Availability and ImplementationThe model implementation, data simulation and experiment scripts are available under https://www.github.com/bioFAM/FISHFactor. Contactb.velten@dkfz.de

bioinformatics↗

bioRxiv2021

SpatialDE2: Fast and localized variance component analysis of spatial transcriptomics

Spatial transcriptomics is now a mature technology, allowing to assay gene expression changes in the histological context of complex tissues. A canonical analysis workflow starts with the identification of tissue zones that share similar expression profiles, followed by the detection of highly variable or spatially variable genes. Rapid increases in the scale and complexity of spatial transcriptomic datasets demand that these analysis steps are conducted in a consistent and integrated manner, a requirement that is not met by current methods. To address this, we here present SpatialDE2, which unifies the mapping of tissue zones and spatial variable gene detection as integrated software framework, while at the same time advancing current algorithms for both of these steps. Formulated in a Bayesian framework, the model accounts for the Poisson count noise, while simultaneously offering superior computational speed compared to previous methods. We validate SpatialDE2 using simulated data and illustrate its utility in the context of two real-world applications to the spatial transcriptomics profiles of the mouse brain and human endometrium.

bioinformatics↗