Biology subjects

Hackenberg, M.

Publications and source records attributed to Hackenberg, M..

At least 19 recordsLinked to original sources

bioRxiv2026

Benchmarking long-read RNA sequencing for de novo transcriptome assembly in non-model plant species: insights from Moricandia arvensis

De novo transcriptome assembly is the standard approach for constructing a reference transcriptome in non-model plants that lack a high-quality genome, yet short-read assemblies struggle to resolve full-length isoforms. Long-read Iso-Seq (PacBio) captures full-length transcripts directly, but its use as a primary reference and the choice of downstream assembly pipeline remains poorly benchmarked. Here, we construct genome-free Iso-Seq reference transcriptomes for two organs, flower and leaf, of the non-model species Moricandia arvensis (L.) DC. (Brassicaceae), and systematically compare pipeline strategies combining Iso-Seq clustering, CD-HIT redundancy reduction, and Cogent graph-based reconstruction, benchmarked by BUSCO completeness, RSEM short-read mapping, and TransDecoder ORF completeness. We find that the optimal pipeline is organ specific. For flower, CD-HIT pre-filtering followed by Cogent reconstruction produced a high-quality reference (95.3% BUSCO complete). For the leaf, the same Cogent step was detrimental, reducing BUSCO completeness from 90.1% to 78.0% by incorrectly merging distinct genes; therefore, CD-HIT at 95% identity without reconstruction was retained. We trace this divergence to organ-specific input-data characteristics: leaf transcripts show extreme full-length-read expression skew and predominantly single-isoform gene support, depriving Cogent's graph algorithm of the multi-isoform evidence it requires. We find that the concentration of full-length reads among the most highly expressed transcripts predicts pipeline suitability before reconstruction, with per-transcript read depth acting as a necessary but non-discriminating floor. Because the leaf reference lacked gene-level structure, we further recovered gene-isoform grouping using expression-aware read-clustering (Corset), which preserved completeness while restoring the paralog structure expected of a paleopolyploid genome and outperformed sequence-only clustering. Our results provide a robust, genome-free framework for constructing full-length reference transcriptomes in non-model plant species and demonstrate that pipeline choice must be evaluated per organ rather than assuming one size fits all.

bioinformatics↗

bioRxiv2026

Mapping spatial cell-cell communication programs by tailoring chains of cells for transformer neural networks

Recent advances in spatial transcriptomics and computational modeling enable the study of cellular interactions in situ. However, existing methods quantify ligand-receptor activity pairwise or between predefined cell groups, yielding overlapping signals and limited ability to summarize concurrent interactions into programs while localizing communication hotspots. We introduce scCChain, a transformer-based framework that integrates ligand-receptor activity into spatially resolved communication programs and localizes hotspots at spot and single-cell resolution. scCChain first derives candidate programs using structured dimensionality reduction. Subsequently, it samples programspecific communication chains by linking transcriptionally similar sender cells to candidate receivers via weighted random walks on a distance-informed cell graph, borrowing signal from similar neighbors. Transformer-based modeling then scores chains to prioritize communication programs and pinpoint hotspots across the tissue. Applied to human breast cancer spatial transcriptomics data at spot and single-cell resolution, scCChain supports both exploratory communication program discovery and targeted analysis of user-specified ligand-receptor pairs. In spot-level data, it prioritizes a tumor-associated program enriched for pro-angiogenic signaling that localizes to invasive regions. In imaging-based data, it highlights CXCL12-CXCR4 communication hotspots at cellular resolution. Here, we demonstrate that chain-based transformer modeling enables interpretable discovery and mapping of biological meaningful spatial communication programs within complex tissues.

bioinformatics↗

bioRxiv2025

mmContext: an open framework for multimodal contrastive learning of omics and text data

SummaryMultimodal approaches are increasingly leveraged for integrating omics data with textual biological knowledge. Yet there is still no accessible, standardized framework that enables systematic comparison of omics representations with different text encoders within a unified workflow. We present mmContext, a lightweight and extensible multimodal embedding framework built on top of the open-source Sentence Transformers library. The software allows researchers to train or apply models that jointly embed omics and text data using any numeric representation stored in an AnnData .obsm layer and any text encoder available in Hugging Face. mmContext supports integration of diverse biological text sources and provides pipelines for training, evaluation, and data preparation. We train and evaluate models for a RNA-Seq and text integration task, and demonstrate their utility through zero-shot classification of cell types and diseases across four independent datasets. By releasing all models, datasets, and tutorials openly, mmContext enables reproducible and accessible multimodal learning for omics-text integration. Availability and implementationPretrained checkpoints and full source code for our custom MMContextEncoder are available on Hugging Face huggingface.co/jo-mengr. The Python package github.com/mengerj/mmcontext provides the model implementation and training and evaluation scripts for custom training.

bioinformatics↗

bioRxiv2025

An integrative framework identifies cooperative targeting of host pathways by tick salivary miRNAs

Ticks are ectoparasites that modulate host responses to sustain prolonged blood feeding, and in Ixodes ricinus, salivary microRNAs (miRNAs) represent promising candidates for manipulating host gene expression. Using phylogenetic footprinting combined with cooperative targeting analysis, we identified deeply conserved miRNA-mRNA interactions that appear essential for the tick lifecycle and its ability to parasitize diverse vertebrate hosts. Our findings show that conserved tick miRNAs can mimic host miRNAs by exploiting shared target sites on host transcripts, allowing them to modulate regulatory circuits critical during tick feeding. We identified twelve core salivary miRNAs--highly expressed and enriched in conserved target sites--that cooperatively target twenty-two human genes predominantly expressed in skin and neural tissues. This cooperative suppression converges on host hub genes like PDGFRA and NRG1, linking tick miRNA activity to MAPK and PI3K-AKT pathways that regulate immune defense, tissue repair, and sensory responses. Overall, our results demonstrate that I. ricinus miRNAs can synergistically disrupt host homeostasis, and they introduce a broadly applicable framework that leverages evolutionary conservation to detect meaningful cross-species miRNA-mRNA interactions. HighlightsThe study shows that tick miRNAs with Bilaterian node of origin mimic host miRNAs to exploit conserved target sites of host mRNAs. It demonstrates that these conserved tick miRNAs cooperatively repress key host defense and repair pathways. It introduces a framework that leverages evolutionary conservation to detect and explain functional, conserved cross-species miRNA-mRNA regulation. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=71 SRC="FIGDIR/small/692500v1_ufig1.gif" ALT="Figure 1"> View larger version (23K): org.highwire.dtl.DTLVardef@6eea0aorg.highwire.dtl.DTLVardef@1c739d2org.highwire.dtl.DTLVardef@a53910org.highwire.dtl.DTLVardef@1458544_HPS_FORMAT_FIGEXP M_FIG C_FIG

systems biology↗

bioRxiv2025

Adding layers of information to scRNA-seq data using pre-trained language models

Pre-trained language models promise to enrich single-cell analyses with contextual information from large biomedical text corpora, but it remains unclear how to optimally align this knowledge with quantitative scRNA-seq data. To address this, we construct text-based training datasets from both scRNA-seq data and biomedical literature targeted to the experimental setting at hand. We then fine-tune lightweight encoder-only biomedical language models to learn a shared, literature-enriched representation. Controlled evaluations across immune and developmental datasets show that this representation preserves cell identity while adding robust and interpretable contextual layers of functional, disease-associated, and developmental information to single-cell analysis workflows.

bioinformatics↗

bioRxiv2025

Evaluating discrepancies in dimensionality reduction for time-series single-cell RNA-sequencing data

There are various dimensionality reduction techniques for visually inspecting dynamical patterns in time-series single-cell RNA-sequencing (scRNA-seq) data. However, the lack of one-to-one correspondence between cells across time points makes it difficult to uniquely uncover temporal structure in a low-dimensional manifold. The use of different techniques may thus lead to discrepancies in the representation of dynamical patterns. However, The extent of these discrepancies remains unclear. To investigate this, we propose an approach for reasoning about such discrepancies based on synthetic time-series scRNA-seq data generated by variational autoencoders. The synthetic dynamical patterns induced in a low-dimensional manifold reflect biologically plausible temporal patterns, such as dividing cell clusters during a differentiation process. We consider manifolds from different dimensionality reduction techniques, such as principal component analysis (PCA), t-distributed stochastic neighbor embedding (t-SNE), uniform manifold approximation and projection (UMAP) and single-cell variational inference (scVI). We illustrate how the proposed approach allows for reasoning about to what extent low-dimensional manifolds, obtained from different techniques, can capture different dynamical patterns. The results indicate that these techniques may not reliably represent dynamics when used in isolation. Thus, the proposed synthetic dynamical pattern approach provides a foundation for guiding future methods development to detect complex patterns in time-series scRNA-seq data.

bioinformatics↗

bioRxiv2024

Sparse dimensionality reduction for analyzing single-cell-resolved interactions

SummarySeveral approaches have been proposed to reconstruct interactions between groups of cells or individual cells from single-cell transcriptomics data, leveraging prior information about known ligand-receptor interactions. To enhance downstream analyses, we present an end-to-end dimensionality reduction workflow, specifically tailored for single-cell cell-cell interaction data. In particular, we demonstrate that sparse dimensionality reduction can pinpoint specific ligand-receptor interactions in relation to clusters of cell pairs. For sparse dimensionality reduction, we focus on the Boosting Autoencoder approach (BAE). Overall, we provide a comprehensive workflow, including result visualization, that simplifies the analysis of interaction patterns in cell pairs. This is supported by a Jupyter notebook that can readily be adapted to different datasets. Availability and implementationhttps://github.com/NiklasBrunn/Sparse-dimension-reduction Contactniklas.brunn@uniklinik-freiburg.de Supplementary material...

bioinformatics↗

bioRxiv2024

MirGeneDB 3.0: Improved taxonomic sampling, uniform nomenclature of novel conserved microRNA families, and updated covariance models.

We present a major update of MirGeneDB (3.0), the manually curated animal microRNA gene database. Beyond moving to a new server and the creation of a computational mirror, we have expanded the database with the addition of 33 invertebrate species, including representatives of 5 previously unsampled phyla, and 6 mammal species. MirGeneDB now contains entries for 21, 822 microRNA genes (5, 160 of these from the new species) belonging to 1743 microRNA families. The inclusion of these new species allowed us to refine both the evolutionary node of appearance of a number of microRNA genes/families, as well as MirGeneDBs phylogenetically informed nomenclature system. Updated covariance models of all microRNA families, along with all smallRNA read data are now downloadable. These enhanced annotations will allow researchers to analyze microRNA properties such as secondary structure and features of their biogenesis within a robust phylogenetic context and without the database plagued with numerous false positives and false negatives. In light of these improvements, MirGeneDB 3.0 will assume the responsibility for naming conserved novel metazoan microRNAs. MirGeneDB is part of RNAcentral and Elixir Norway, and is publicly and freely available at master.cloud.mirgenedb.org. Key PointsO_LIMajor update to the manually curated and uniformly named microRNA gene database MirGeneDB C_LIO_LI114 animal species, >1700 microRNA families and [~]20 000 genes searchable, browsable and downloadable C_LIO_LINew features to download all smallRNA read data and updated covariance models for each family C_LI

evolutionary biology↗

bioRxiv2024

Enhancement of network architecture alignment in comparative single-cell studies

1Animal data can provide meaningful context for human gene expression at the single-cell level. This context can improve cell-type or cell-state detection and clarify how well the animal models human biological processes. To achieve this, we propose a deep learning approach that identifies a unified latent space to map complex patterns between datasets. Specifically, we combine variational autoencoders with a data-level nearest neighbor search to align neural network architectures across species. We visualize commonalities by mapping cell samples into the latent space. The aligned latent representation facilitates information transfer in applications of liver, white adipose tissue, and glioblastoma cells from various animal models. We also identify genes that exhibit systematic differences and commonalities between species. The results are robust for small datasets and with large differences in the observed gene sets. Thus, we reliably uncover and exploit similarities between species to provide context for human single-cell data.

bioinformatics↗

bioRxiv2024

Full-length tRNAs lacking a functional CCA tail are selectively sorted into the lumen of extracellular vesicles

Small extracellular vesicles (sEVs) are heterogenous lipid membrane particles typically less than 200 nm in size and secreted by most cell types either constitutively or upon activation signals. sEVs isolated from biofluids contain RNAs, including small non-coding RNAs (ncRNAs), that can be either encapsulated within the EV lumen or bound to the EV surface. EV-associated microRNAs (miRNAs) are, despite a relatively low abundance, extensively investigated for their selective incorporation and their role in cell-cell communication. In contrast, the sorting of highly-structured ncRNA species is understudied, mainly due to technical limitations of traditional small RNA sequencing protocols. Here, we adapted ALL-tRNAseq to profile the relative abundance of highly structured and potentially methylated small ncRNA species, including transfer RNAs (tRNAs), small nucleolar RNAs (snoRNAs), and Y RNAs in bulk EV preparations. We determined that full-length tRNAs, typically 75 to 90 nucleotides in length, were the dominant small ncRNA species (>60% of all reads in the 18-120 nucleotides size-range) in all cell culture-derived EVs, as well as in human plasma-derived EV samples, vastly outnumbering 21 nucleotides-long miRNAs. Nearly all EV-associated tRNAs were protected from external RNAse treatment, indicating a location within the EV lumen. Strikingly, the vast majority of luminal-sorted, full-length, nucleobase modification-containing EV-tRNA sequences, harbored a dysfunctional 3 CCA tail, 1 to 3 nucleotides truncated, rendering them incompetent for amino acid loading. In contrast, in non-EV associated extracellular particle fractions (NVEPs), tRNAs appeared almost exclusively fragmented or nicked into tRNA-derived small RNAs (tsRNAs) with lengths between 18 to 35 nucleotides. We propose that in mammalian cells, tRNAs that lack a functional 3 CCA tail are selectively sorted into EVs and shuttled out of the producing cell, offering a new perspective into the physiological role of secreted EVs and luminal cargo-selection.

molecular biology↗

bioRxiv2024

Whole genome sequencing of four Ixodes species expands understanding of tick evolution

Ticks, hematophagous acari, pose a significant threat by transmitting various pathogens to their vertebrate hosts during feeding. Despite advances in tick genomics, high-quality genomes were lacking until recently, particularly in the genus Ixodes, which includes the main vectors of Lyme disease. Here, we present the complete genome sequences of four tick species, derived from a single female individual, with a particular focus on the European species Ixodes ricinus, achieving a chromosome-level assembly. Additionally, draft assemblies were generated for the three other Ixodes species, I. persulcatus, I. pacificus and I. hexagonus. The quality of the four genomes and extensive annotation of several important gene families have allowed us to study the evolution of gene repertoires at the level of the genus Ixodes and of the tick group. We have determined gene families that have undergone major amplifications during the evolution of ticks, while an expression atlas obtained for I. ricinus reveals striking patterns of specialization both between and within gene families. Notably, several gene family amplifications are associated with a proliferation of single-exon genes. The integration of our data with existing genomes establishes a solid framework for the study of gene evolution, improving our understanding of tick biology. In addition, our work lays the foundations for applied research and innovative control targeting these organisms.

genomics↗

bioRxiv2024

Infusing structural assumptions into dimension reduction for single-cell RNA sequencing data to identify small gene sets

Dimensionality reduction greatly facilitates the exploration of cellular heterogeneity in single-cell RNA sequencing data. While most of such approaches are data-driven, it can be useful to incorporate biologically plausible assumptions about the underlying structure or the experimental design. We propose the boosting autoencoder (BAE) approach, which combines the advantages of unsupervised deep learning for dimensionality reduction and boosting for formalizing assumptions. Specifically, our approach selects small sets of genes that explain latent dimensions. As illustrative applications, we explore the diversity of neural cell identities and temporal patterns of embryonic development.

bioinformatics↗

bioRxiv2023

HSV-1 miRNAs are posttranscriptionaly edited in latently infected human ganglia

Viruses use miRNAs to enable efficient replication, control host defense mechanisms, and regulate latent infection. Herpes simplex virus 1 (HSV-1) expresses multiple miRNAs, whose functions are largely unknown. The evolutionary conservation of many HSV-1 miRNAs in the closely related HSV-2 suggests their functional importance. miRNAs, similar to other transcripts, can undergo various posttranscriptional modifications that may affect their biogenesis, stability and targeting. To investigate whether editing occurs in HSV-1 miRNAs, we sequenced samples from latently infected human ganglia. We show that one of the six HSV-1 miRNAs (miR-H2 to -H8) that define HSV-1 latency, miR-H2, exhibits A-to-I hyperediting within the miRNA seed sequence. We observed the same specific miR-H2 hyperediting phenomenon in miRNAs isolated from the ganglia of latently infected mice and, to a lesser extent, during productive infection in cultured cells. Curiously, we found no evidence of editing of the encoded HSV-2 homolog in latently infected mice or in cultured cells. The efficient loading of the edited miRNAs onto the RISC complex, indicates their ability to function as miRNAs. To investigate the potential of the edited miRNA to alter mRNA targeting, we predicted the host and viral targets for the modified miRNAs. Nucleotide substitution in the seed region significantly increased the number of potential host and viral targets. Most notably, ICP4, an essential viral protein, was predicted to be an additional target. Using transfection assays, we demonstrated that edited miRNAs have the potential to regulate ICP4 in addition to the previously identified target ICP0. Our study identifies a specific hyperedited HSV-1 mRNA, miR-H2, and highlights how the virus can use a single miRNA to target multiple transcripts during persistent, latent infection. ImportanceHerpes simplex virus 1 is an important human pathogen and intensively studied for many decades. Nevertheless, the molecular mechanisms regulating its establishment, maintenance, and reactivation from latency are poorly understood. Here, we show that HSV- 1 encoded miR-H2 is post-transcriptionally edited in latently infected human tissues. Hyperediting of viral miRNAs increases the targeting potential of these miRNAs and may play an important role in regulating latency. We show that the edited miR-H2 (miR-H2-e) can target ICP4, an essential viral protein. Interestingly, we found no evidence of hyperdating of its homolog, miR-H2, which is expressed by the closely related virus HSV-2. The discovery of posttranslational modifications of viral miRNA in the latency phase suggests that these processes may also be important for other non-coding viral RNA in the latency phase, including the intron LAT, which in turn may be crucial for understanding the biology of this virus.

microbiology↗

bioRxiv2022

The performance of deep generative models for learning joint embeddings of single-cell multi-omics data

Recent extensions of single-cell studies to multiple data modalities raise new questions regarding experimental design. For example, the challenge of sparsity in single-omics data might be partly resolved by compensating for missing information across modalities. In particular, deep learning approaches, such as deep generative models (DGMs), can potentially uncover complex patterns via a joint embedding. Yet, this also raises the question of sample size requirements for identifying such patterns from single-cell multi-omics data. Here, we empirically examine the quality of DGM-based integrations for varying sample sizes. We first review the existing literature and give a short overview of deep learning methods for multi-omics integration. Next, we consider eight popular tools in more detail and examine their robustness to different cell numbers, covering two of the most common multi-omics types currently favored. Specifically, we use data featuring simultaneous gene expression measurements at the RNA level and protein abundance measurements for cell surface proteins (CITE-seq), as well as data where chromatin accessibility and RNA expression are measured in thousands of cells (10x Multiome). We examine the ability of the methods to learn joint embeddings based on biological and technical metrics. Finally, we provide recommendations for the design of multi-omics experiments and discuss potential future developments.

bioinformatics↗

bioRxiv2021

Incorporating structural knowledge into unsupervised deep learning for two-photon imaging data

Live imaging techniques, such as two-photon imaging, promise novel insights into cellular activity patterns at a high spatio-temporal resolution. While current deep learning approaches typically focus on specific supervised tasks in the analysis of such data, we investigate how structural knowledge can be incorporated into an unsupervised generative deep learning model directly at the level of the video frames. We exemplify the proposed approach with two-photon imaging data from hippocampal CA1 neurons in mice, where we account for spatial structure with convolutional neural network components, disentangle the neural activity of interest from the neuropil background signal with separate foreground and background encoders and model gradual temporal changes by imposing smoothness constraints. Taken together, our results illustrate how such architecture choices facilitate a modeling approach that combines the flexibility of deep learning with the benefits of domain knowledge, providing an interpretable, purely image-based model of activity signals from live imaging data. Teaser sentenceUsing a neural network architecture that reflects domain knowledge provides an interpretable model of live cell imaging data.

bioinformatics↗

bioRxiv2021

Unbiased and UMI-informed sequencing of cell-free miRNAs at single-nucleotide resolution

Terminal nucleotidyl transferases are enzymes that add non-templated nucleotides to RNA molecules. In the case of microRNAs, this process was shown to be functionally relevant for their maturation process and generation of isomiRs with non-canonical mRNA targets. Deconvolution of these posttranscriptional modifications is challenging in particular for extracellular miRNAs that are considered as a target for minimally-invasive diagnostics. Massively parallel RNA sequencing is the only method that can truthfully reveal isomiR diversity in biological samples and determine relative quantities. Improvements aside, current small RNA sequencing strategies remain imprecise. We developed IsoSeek that diverges from these methods by making use of randomized 5- and 3-adapters combined with a 10N unique molecular identifier (UMI). Using synthetic miRNA and isomiR spike-in sets and testing depletion and RNA competition strategies in 7 sequencing rounds of >100 samples, we rigorously optimized and validated the technical accuracy of the IsoSeek method. In genetically-altered HEK293, we characterized the terminal uridylase (TUT4/TUT7) dependent miRNA uridylome and discovered extensive uridylation of disease-associated miRNAs. Notably, 3-uridylated isomiR profiles of plasma extracellular vesicles (EVs) rely on UMI-correction. Thus, IsoSeek advances our knowledge of cell-free miRNAs and supports development into non-invasive biomarkers.

genomics↗

bioRxiv2021

GeneCodis 4: Expanding the modular enrichment analysis to regulatory elements

GeneCodis is a web-based tool for functional enrichment analysis that allows researchers to integrate different sources of annotations. It extracts sets of significant concurrent annotations and assigns a statistical score to evaluate those that are significantly enriched in the input list. Since its first release in 2007, it has been widely employed to analyze lists of genes in order to interpret the underlying biological mechanisms. Here we present GeneCodis 4, a new release that expands the functional analysis provided by the application, to accept regulatory elements, including lists of, transcription factors, CpG sites and miRNAs. It also incorporates new annotation databases and improved interactive visualizations functionalities to explore results. GeneCodis 4 is freely available at https://genecodis.genyo.es.

bioinformatics↗

bioRxiv2018

Unification of miRNA and isomiR research: the mirGFF3 format and the mirtop API

BackgroundMicroRNAs (miRNAs) are small RNA molecules ([~]22 nucleotide long) involved in post-transcriptional gene regulation. Advances in high-throughput sequencing technologies led to the discovery of isomiRs, which are miRNA sequence variants. While many miRNA-seq analysis tools exist, a lack of consensus on miRNA/isomiR analyses exists, and the resulting diversity of output formats hinders accurate comparisons between tools and precludes data sharing and the development of common downstream analysis methods.\n\nFindingsTo overcome this situation, we present here a community-based project, miRTOP (miRNA Transcriptomic Open Project) working towards the optimization of miRNA analyses. The aim of miRTOP is to promote the development of downstream analysis tools that are compatible with any existing detection and quantification tool. Based on the existing GFF3 format, we first created a new standard format, mirGFF3, for the output of miRNA/isomiR detection and quantification results from small RNA-seq data. Additionally, we developed a command line Python tool, mirtop, to manage the mirGFF3 format. Currently, mirtop can convert into mirGFF3 the outputs of commonly used pipelines, such as seqbuster, miRge2.0, isomiR-SEA, sRNAbench, and Prost!, as well as BAM files. Its open architecture enables any tool or pipeline to output results in mirGFF3.\n\nConclusionsCollectively a comprehensive isomiR categorization system, along with the accompanying mirGFF3 and mirtop API provide a complete solution for the standardization of miRNA and isomiR analysis, enabling data sharing, reporting, comparative analyses, and benchmarking, while promoting the development of common miRNA methods focusing on downstream steps to miRNA detection, annotation, and quantification.

bioinformatics↗