Biology subjects

Mohorianu, I.

Publications and source records attributed to Mohorianu, I..

At least 19 recordsLinked to original sources

bioRxiv2026

Deep generative embeddings of gene expression and splicing reposition the interpretation of single-cell transcriptomic signatures

Single-cell transcriptomic analysis predominantly derives cell identity from gene expression analysis, while alternative splicing is processed separately despite its fundamental role for cell homeostasis. To overcome the limits of separate investigations, we developed a probabilistic deep learning framework, Crecerelle, enabling resolution of the contributions of gene expression and alternative splicing in each cell. Crecerelle learns cell embeddings from gene expressions and alternative splicing isoforms, to decipher their mutually dependent impact on the functional characterisation of cells in a data-driven manner, exemplified for the Tabula Muris dataset. This is enabled through a zero-and-N-inflated Dirichlet-Multinomial for a variational autoencoder that learns cell embeddings solely from splicing profiles, as well as a bi-modal variational autoencoder with a relevance-weighted mixture-of-experts variational posterior to consolidate the modality-specific contribution at single-cell level. Crecerelle reveals cell-type-specific isoform markers as well as subpopulations with unique isoforms and uncovers regulatory and disease-associated pathways not detected by gene expression analyses alone. This scalable and interpretable framework thus allows a more holistic study of transcriptomic regulation and will open a route to modality-relevance-weighted investigations across single-cell multiomics datasets and their influence on cellular homeostasis, tissue development and disease phenotypes.

cell biology↗

bioRxiv2026

SMEW: An interactive multi-scale toolkit for cross-condition and network-based analysis of spatial metabolomics data

Spatial metabolomics, measured through mass spectrometry imaging (MSI), provides high-throughput, spatially resolved information on metabolite distributions within tissues, including endogenous metabolites and exogenous compounds. This offers a direct readout of cellular biochemical activity and phenotypes, not fully captured by transcriptomics or proteomic profiling. However, inferring biologically meaningful patterns from noisy, high-dimensional MSI data, particularly across multiple samples and complex experimental designs, remains challenging, and often requires substantial programming expertise. Here we introduce SMEW (Spatial Metabolomics Enhanced Workflow), a flexible, interactive and shareable Shiny-based platform designed to enable code-free downstream analysis of spatial metabolomics MSI data. SMEW provides a unified environment for hierarchical analysis across bulk-, region- and pixel-level resolutions, allowing comparisons between experimental conditions like disease or treatment groups while highlighting coherent metabolic patterns and linking these patterns to biological pathways. The workflow leverages local spatial covariation to robustly summarise MSI data through dimensionality reduction, clustering and identification of spatially variable metabolites. In addition, metabolite co-localisation and covariation network analysis, together with spatially resolved pathway enrichment facilitate the biological interpretation of cross-condition datasets within a single integrated interface. SMEW is applicable across MSI technologies and mass resolutions, as illustrated through case studies on DESI and MALDI-ToF datasets from lung, liver, and kidney. By complementing existing MSI processing and visualisation tools with an accessible, multi-sample, and biologically interpretable analysis framework, SMEW enables functional, flexible, rigorous and intuitive exploration of spatial metabolomics datasets. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=84 SRC="FIGDIR/small/721059v1_ufig1.gif" ALT="Figure 1"> View larger version (29K): org.highwire.dtl.DTLVardef@1e2abaeorg.highwire.dtl.DTLVardef@753ee9org.highwire.dtl.DTLVardef@1756fc1org.highwire.dtl.DTLVardef@fbedc7_HPS_FORMAT_FIGEXP M_FIG C_FIG Key PointsO_LISMEW provides a flexible, interactive and shareable Shiny-based platform designed to enable code-free downstream analysis of spatial metabolomics MSI data C_LIO_LIThe SMEW framework enables hierarchical analysis at bulk-, region- and pixel levels within a unified framework without relying on extensive programming expertise C_LIO_LIThe pipeline integrates spatially aware clustering, pathway analysis and identification of metabolite co-localisation modules C_LIO_LIThe workflow facilitates flexible comparison of multi-sample experimental conditions through multivariate modelling, differential testing and covariation networks to study treatment- and disease-associated metabolite dynamics C_LIO_LISMEW has been applied to interrogate diverse biological questions, including characterising disease-associated remodelling in a mouse bleomycin model of pulmonary fibrosis, exploring the therapeutic index of antisense oligonucleotides in the liver and assessing metabolic heterogeneity in a small molecule-treated mouse renal tumour model C_LI

bioinformatics↗

bioRxiv2026

A comparative analysis of acellular versus red cell based subnormothermic machine perfusion in human kidney transplantation

Normothermic machine perfusion (NMP) typically uses red blood cell (RBC)-based perfusates, with associated challenges, including free haem-driven ferroptotic tubular injury. Acellular perfusates avoid these risks. Subnormothermic machine perfusion (SMP, 32{degrees}C) may reduce metabolic demands while preserving organ viability, but the effect of perfusate type under these conditions remains unclear. In a paired experimental design, kidneys from eight deceased donors were randomized 1:1 to 6 hours of SMP with RBC-based (RBC-SMP) or acellular perfusate (SNAP), followed by 4 hours of RBC-based NMP (37{degrees}C) to simulate reperfusion. Upregulation of transcriptomic ischaemia-reperfusion injury (IRI) related pathways and downregulation of oxidative phosphorylation at 6 hours SMP and 4 hours reperfusion were similar between RBC-SMP and SNAP kidneys. No differences were observed between groups in transcriptional profiles, metabolic pathway scores, urinary tubular injury biomarker concentrations, or histology. We integrated publicly available NMP transcriptomic data for comparison, demonstrating that compared to 6 hours RBC-NMP at 37{degrees}C, both SMP groups showed equivalent pro-inflammatory pathway activation and oxidative phosphorylation depletion. Thus, we provide evidence that SNAP maintains kidneys with equivalent transcriptional and urinary tubular injury biomarker profiles to RBC-SMP. These findings support the feasibility of SNAP for clinical translation in future studies.

physiology↗

bioRxiv2025

Spatio-Temporal Multi-Omics Profiling of Mechanisms and Biomarkers in Inhaled Drug-Induced Lung Toxicity

Comprehensive understanding and early detection of drug-induced lung toxicity remain critical challenges in respiratory drug development. In this study, we propose a multi-omics framework that integrates spatial and temporal tissue-specific transcriptomic signatures with proteomics from minimally invasive biofluids to understand mechanisms and identify safety biomarkers associated with lung toxicity. Using this framework, we identified a panel of candidate biomarkers in bronchoalveolar lavage fluid and plasma, including LCN2/NGAL, RETNLA, SP-D, SPP1/osteopontin, and MMP7, that correlate with histopathological features (e.g., inflammation and epithelial remodeling). We confirmed that these molecular biomarkers were consistently dysregulated across a range of inhaled lung toxicants, human disease (IPF), and environmental exposures (smoke, Alternaria), demonstrating broad applicability across different toxic exposures and translatability. Collectively, this study establishes a robust workflow for mechanism-guided biomarker discovery and proposes a panel of candidates for monitoring drug-induced lung injury in humans.

systems biology↗

bioRxiv2025

Spatially resolved integrative analysis of transcriptomic and metabolomic changes in tissue injury studies

Recent developments in spatially resolved -omics have enabled studies linking gene expression and metabolite levels to tissue morphology, offering new insights into biological pathways. By capturing multiple modalities on matched tissue sections, one can better probe how different biological entities interact in a spatially coordinated manner. However, such cross-modality integration presents experimental and computational challenges. To align multimodal datasets into a shared coordinate system and facilitate enhanced integration and analysis, we propose MAGPIE (Multi-modal Alignment of Genes and Peaks for Integrated Exploration), a framework for co-registering spatially resolved transcriptomics, metabolomics, and tissue morphology from the same or consecutive sections. We illustrate the generalisability and scalability of MAGPIE on spatial multi-omics data from multiple tissues, combining Visium with both MALDI and DESI mass spectrometry imaging. MAGPIE was also applied to newly generated multimodal datasets created using specialised experimental sampling strategy to characterise the metabolic and transcriptomic landscape in an in vivo model of drug-induced pulmonary fibrosis, to showcase the linking of small-molecule co-detection with endogenous responses in lung tissue. MAGPIE highlights the refined resolution and increased interpretability of spatial multimodal analyses in studying tissue injury, particularly in pharmacological contexts, and offers a modular, accessible computational workflow for data integration.

bioinformatics↗

bioRxiv2025

A non-catalytic role for MLL2 in controlling chromatin organisation and mobility during the priming of pluripotent cells for differentiation

How the chromatin regulator MLL2 (KMT2B) influences cell differentiation remains poorly understood. MLL2 is the main histone 3 lysine 4 (H3K4) trimethyltransferase acting at bivalent promoters in embryonic stem cells (ESCs) and is required for ESCs to differentiate into neuroectoderm. We show here that this requirement occurs during exit from naive pluripotency, days before neuroectoderm differentiation is impaired. Although MLL2 knockout has only a subtle effect on transcription during exit, reducing the expression of a few important neuroectodermal transcription factors, it substantially remodels chromatin architecture, disrupting 3D chromatin loops associated with bivalent promoters. The enzymatic activity of MLL2 is not needed for stabilising these loops or for neuroectoderm differentiation. This non-catalytic function of MLL2 in stabilising 3D chromatin loops has implications for lineage specification. Because MLL2 shares features with all four MLLs, chromatin tethering, rather than H3K4 methylation, may represent the primary function of MLL proteins during lineage commitment.

cell biology↗

bioRxiv2024

Novel 3D Approach to Model Non-Alcoholic Fatty Liver Disease using human Pluripotent Stem Cells

Background and aimsNon-alcoholic fatty liver disease (NAFLD) is a major health care challenge and new therapies are urgently needed. However, the mechanisms underlying disease remain to be understood. Indeed, studying NAFLD remains challenging due to the lack of model systems recapitulating the different aspects of the human pathology. Human induced pluripotent stem cells (hiPSCs) offer a unique opportunity to address this limitation since they can be differentiated into large quantity of liver cells. Here, we took advantage of hiPSCs to develop a multi-cellular platform mimicking the complex interplays involved in NAFLD progression. MethodshiPSCs-derived hepatocyte like cells (HLCs), cholangiocytes, stellate cells, and macrophages were co-cultured in a collagen-based 3D system to reproduce the liver microenvironment. Fatty acid treatments led to a NAFLD phenotype involving cell-cell interactions which were investigated by transcriptomic and functional analyses. ResultsHepatic cells were grown up to 4weeks in 3D, retaining key functions and markers. Importantly, co-cultured cells spontaneously reorganised into physiologically relevant connections: HLCs arranged around biliary structures, which established contacts with stellate cells, while macrophages organised around HLCs. Fatty acid treatments induced steatosis and lipotoxicity in HLCs. Furthermore, fat-laden HLCs prompted a non-parenchymal cells response altering tissue architecture. ConclusionsOur multicellular platform provides a new approach to model interactions between human hepatic cells during NAFLD progression. Such approach has the potential to investigate the sequential events driving chronic liver diseases, including hepatocellular injury, inflammation and fibrosis. Furthermore, our system provides a unique and urgently needed tool to investigate the molecular mechanisms associated with NAFLD and ultimately to validate new targets for therapeutics development. List of abbreviationsCOs, cholangiocytes organoids; FFA, free fatty acids; hiPSCs, human induced pluripotent stem cells; HLCs, hepatocyte like cells; HSCs, hepatic stellate cells; M0, hiPSCs-derived macrophages; NAFLD, non-alcoholic fatty liver disease; NPCs, non-parenchymal cells; OA, oleic acid; PA, palmitic acid.

cell biology↗

bioRxiv2024

Increased Cholesterol Synthesis Drives Neurotoxicity in Patient Stem Cell-Derived Model of Multiple Sclerosis

Senescent neural progenitor cells have been identified in brain lesions of people with progressive multiple sclerosis (PMS). However, their role in disease pathobiology and contribution to the lesion environment remains unclear. By establishing directly induced neural stem/progenitor cell (iNSC) lines from PMS patient fibroblasts, we studied their senescent phenotype in vitro. Senescence was strongly associated with inflammatory signaling, hypermetabolism, and the senescence associated secretory phenotype (SASP). PMS-derived iNSCs displayed increased glucose-dependent fatty acid and cholesterol synthesis, which resulted in the accumulation of cholesteryl ester-enriched lipid droplets. An HMG-CoA reductase-mediated lipogenic state was found to induce secretion of the SASP in PMS iNSC conditioned media via transcriptional regulation by cholesterol-dependent transcription factors. SASP from PMS iNSCs induced neurotoxicity. Chemical targeting of HMG-CoA reductase using the cholesterol-lowering drug simvastatin (SV) prevented SASP release and resulting neurotoxicity. Our findings suggest a disease-associated, cholesterol-related, hypermetabolic phenotype of PMS iNSCs that leads to neurotoxic signaling and is rescuable pharmacologically.

neuroscience↗

bioRxiv2022

RBPMS promotes contractile smooth muscle splicing and alters phenotypic behaviour of human embryonic stem cell derived vascular smooth muscle cells.

Differentiated Vascular Smooth Muscle Cells (VSMCs) express a unique network of splice isoforms (smooth muscle specific alternative splicing - SM-AS) in functionally critical genes including those comprising the contractile machinery. We previously described RNA Binding Protein Multiple Splicing (RBPMS) as a potent driver of contractile, aortic tissue like SM-AS in VSMCs using rodent models. What is unknown is how RBPMS affects VSMC phenotype and behaviour. Here, we use human embryonic stem cell-derived VSMCs (hES-VSMCs) to dissect the role of RBPMS in SM-AS in human cells and determine the impact on VSMC phenotypic properties. hES-VSMCs are inherently immature and display only partially differentiated SM-AS patterns while RBPMS levels are undetectable endogenously. Hence, we used an over-expression system and found that RBPMS induces SM-AS patterns in hES-VSMCs akin to the contractile tissue VSMC splicing patterns in multiple events. We present in silico and experimental findings that support RBPMS splicing activity as mediated through direct binding and via functional cooperativity with splicing factor RBFOX2 on a significant subset of targets. Finally, we demonstrate that RBPMS is capable of altering the motility and the proliferative properties of hES-VSMCs to mimic a more differentiated state. Overall, this study emphasizes a critical splicing regulatory role for RBPMS in human VSMCs and provides evidence of phenotypic modulation by RBPMS.

molecular biology↗

bioRxiv2022

ClustAssess: tools for assessing the robustness of single-cell clustering

The transition from bulk to single-cell analyses refocused the computational challenges for high-throughput sequencing data-processing. The core of single-cell pipelines is partitioning cells and assigning cell-identities; extensive consequences derive from this step; generating robust and reproducible outputs is essential. From benchmarking established single-cell pipelines, we observed that clustering results critically depend on algorithmic choices (e.g. method, parameters) and technical details (e.g. random seeds). We present ClustAssess, a suite of tools for quantifying clustering robustness both within and across methods. The tools provide fine-grained information enabling (a) the detection of optimal number of clusters, (b) identification of regions of similarity (and divergence) across methods, (c) a data driven assessment of optimal parameter ranges. The aim is to assist practitioners in evaluating the robustness of cell-identity inference based on the partitioning, and provide information for choosing robust clustering methods and parameters. We illustrate its use on three case studies: a single-cell dataset of in-vivo hematopoietic stem and progenitors (10x Genomics scRNA-seq), in-vitro endoderm differentiation (SMART-seq), and multimodal in-vivo peripheral blood (10x RNA+ATAC). The additional checks offer novel viewpoints on clustering stability, and provide a framework for consistent decision-making on preprocessing, method choice, and parameters for clustering.

bioinformatics↗

bioRxiv2021

bulkAnalyseR: An accessible, interactive pipeline for analysing and sharing bulk sequencing results

Bulk sequencing experiments (single- and multi-omics) are essential for exploring wide-ranging biological questions. To facilitate interactive, exploratory tasks, coupled with the sharing of easily accessible information, we present bulkAnalyseR, a package integrating state-of-the-art approaches using an expression matrix as the starting point (pre-processing functions are available as part of the package). Static summary images are replaced with interactive panels illustrating quality-checking, differential expression analysis (with noise detection) and biological interpretation (enrichment analyses, identification expression patterns, followed by inference and comparison of regulatory interactions). bulkAnalyseR can handle different modalities, facilitating robust integration and comparison of cis-, trans- and customised regulatory networks. bulkAnalyseR is available on CRAN and GitHub, with extensive documentation and usage examples (https://github.com/Core-Bioinformatics/bulkAnalyseR, https://cran.r-project.org/web/packages/bulkAnalyseR/) ContactIrina Mohorianu iim22@cam.ac.uk

bioinformatics↗

bioRxiv2021

The sum of two halves may be different from the whole. Effects of splitting sequencing samples across lanes.

The advances in high throughput sequencing (HTS) enabled the characterisation of biological processes at an unprecedented level of detail; the majority of hypotheses in molecular biology rely on analyses of HTS data. However, achieving increased robustness and reproducibility of results remains one of the main challenges. Although variability in results may be introduced at various stages, e.g. alignment, summarisation or detection of differences in expression, one source of variability was systematically omitted: the sequencing design which propagates through analyses and may introduce an additional layer of technical variation. We illustrate qualitative and quantitative differences arising from splitting samples across lanes, on bulk and single-cell sequencing. For bulk mRNAseq data, we focus on differential expression and enrichment analyses; for bulk ChIPseq data, we investigate the effect on peak calling, and peaks properties. At single-cell level, we concentrate on identifying cell subpopulations. We rely on markers used for assigning cell identities; both smartSeq and 10x data are presented. The observed reduction in the number of unique sequenced fragments reduces the level of detail on which the different prediction approaches depend. Further, the sequencing stochasticity adds in a weighting bias corroborated with variable sequencing depths and (yet unexplained) sequencing bias.

bioinformatics↗

bioRxiv2021

Mapping the biogenesis of forward programmed megakaryocytes from induced pluripotent stem cells

Platelet deficiency, known as thrombocytopenia, can cause haemorrhage and is treated with platelet transfusions. We developed a system for the production of platelet precursor cells, megakaryocytes, from pluripotent stem cells. These cultures can be maintained for >100 days, implying culture renewal by megakaryocyte progenitors (MKPs). However, it is unclear whether the MKP state in vitro mirrors the state in vivo, and MKPs cannot be purified using conventional surface markers. We performed single cell RNA sequencing throughout in vitro differentiation and mapped each state to its equivalent in vivo. This enabled the identification of 5 surface markers which reproducibly purify MKPs, allowing us an insight into their transcriptional and epigenetic profiles. Finally, we performed culture optimisation, increasing MKP production. Altogether, this study has mapped parallels between the MKP states in vivo and in vitro and allowed the purification of MKPs, accelerating the progress of in vitro-derived transfusion products towards the clinic.

developmental biology↗

bioRxiv2021

NOTCH1 drives immune-escape mechanisms in B cell malignancies

NOTCH1 is a recurrently mutated gene in Chronic Lymphocytic Leukemia (CLL) and Mantle Cell Lymphoma (MCL). Functional studies to investigate its role have been hampered by the inability to genetically manipulate primary human lymphoma cells, attributed to low transduction-efficacy and procedure-associated toxicity. To overcome these limitations, we have developed a novel method to retrovirally transfer genes into malignant human B cells. We generated isogenic human tumor cells from patients with CLL and MCL, differing only in their expression of NOTCH1. Our data demonstrate that NOTCH1 facilitates immune-escape of malignant B cells by up-regulating PD-L1, partly dependent on autocrine interferon-{gamma} signaling. In addition, NOTCH1 causes silencing of the entire HLA-class II locus via suppression of the transcriptional co-activator CIITA. These NOTCH1-mediated immune escape mechanisms are associated with the expansion of CD4+ T cells in vivo, further contributing to the poor clinical outcome of NOTCH1-mutated CLL and MCL.

cancer biology↗

bioRxiv2021

Defining the transcriptional signature of esophageal-to-skin lineage conversion

The ability of epithelial cells to rewire their cell fate program beyond their physiological repertoire has become a new paradigm in stem cell biology. This plasticity leaves behind the concept of strict stem cell hierarchies, opening up new exciting questions about its limits and underlying regulation. Here we developed a heterotypic 3D culture system to study the mechanisms modulating changes in the identity of adult esophageal epithelial cells. We demonstrate that, when exposed to the foreign stroma of adult skin, esophageal cells transition towards hair follicle identity and architecture. Heterotypic transplantation experiments recapitulated this cell fate conversion process in vivo. Single-cell RNA sequencing and histological analysis, capturing the temporality of this process, reveal that most esophageal cells switching towards skin identity remain in an intermediate state marked by a transient regenerative profile and a particularly strong hypoxic signature. Inhibition of HIF1a establishes the central role of this pathway in regulating epithelial cell plasticity, driving cells away from their transition state in favor of cell fate conversion.

cell biology↗

bioRxiv2020

feamiR: Feature selection based on Genetic Algorithms for predicting miRNA-mRNA interactions

microRNAs play a key role in RNA interference, the sequence-driven targeting of mRNAs that regulates their translation to proteins, through translation inhibition or the degradation of the mRNA. Around ~ 30% of animal genes may be tuned by microRNAs. The prediction of miRNA/mRNA interactions is hindered by the short length of the interaction (seed) region (~7- 8nt). We collate several large datasets overviewing validated interactions and propose feamiR, a novel pipeline comprising optimised classification approaches (Decision Trees/Random Forests and an efficient feature selection based on embryonic Genetic Algorithms used in conjunction with Support Vector Machines) aimed at identifying discriminative nucleotide features, on the seed, compensatory and flanking regions, that increase the prediction accuracy for interactions. Common and specific combinations of features illustrate differences between reference organisms, validation techniques or tissue/cell localisation. feamiR revealed new key positions that drive the miRNA/mRNA interactions, leading to novel questions on the mode-of-action of miRNAs.

bioinformatics↗

bioRxiv2020

Integrative Single-cell RNA-Seq and ATAC-Seq Analysis of Human Foetal Liver and Bone Marrow Haematopoiesis

Regulation of haematopoiesis during human development remains poorly defined. Here, we applied single-cell (sc)RNA-Seq and scATAC-Seq analysis to over 8,000 human immunophenotypic blood cells from foetal liver and bone marrow. We inferred their differentiation trajectory and identified three highly proliferative oligopotent progenitor populations downstream from haematopoietic stem cell/multipotent progenitors (HSC/MPPs). Along this trajectory, we observed opposing patterns of chromatin accessibility and differentiation that coincided with dynamic changes in the activity of distinct lineage-specific transcription factors. Integrative analysis of chromatin accessibility and gene expression revealed extensive epigenetic but not transcriptional priming of HSC/MPPs prior to their lineage commitment. Finally, we refined and functionally validated the sorting strategy for the HSC/MPPs and achieved around 90% enrichment. Our study provides a useful framework for future investigation of human developmental haematopoiesis in the context of blood pathologies and regenerative medicine.

cell biology↗

bioRxiv2020

Live-cell 3D single-molecule tracking reveals how NuRD modulates enhancer dynamics

Enhancer-promoter dynamics are crucial for the spatiotemporal control of gene expression, but it remains unclear whether these dynamics are controlled by chromatin regulators, such as the nucleosome remodelling and deacetylase (NuRD) complex. The NuRD complex binds to all active enhancers to modulate transcription and here we use Hi-C experiments to show that it blurs TAD boundaries and increases the proximity of intermediate-range (~1 Mb) genomic sequences and enhancer-promoter interactions. To understand whether NuRD alters the dynamics of 3D genome structure, we developed an approach to segment and extract key biophysical parameters from 3D single-molecule trajectories of the NuRD complex determined using live-cell imaging. Unexpectedly, this revealed that the intact NuRD complex decompacts chromatin structure and makes NuRD-bound enhancers move faster, increasing the overall volume of the nucleus that these key regulatory regions explore. Interestingly, we also uncovered a rare fast-diffusing state of NuRD bound enhancers that exhibits directed motion. The NuRD complex reduces the amount of time that enhancers remain in this fast-diffusing state, which we propose might otherwise re-organise enhancer-promoter proximity. Thus, we uncover an intimate connection between a chromatin remodeller and the spatial dynamics of the local regions of the genome to which it binds.

cell biology↗