Biology subjects

Heinig, M.

Publications and source records attributed to Heinig, M..

At least 19 recordsLinked to original sources

bioRxiv2026

Intraindividual longitudinal changes of the proteome reveal dynamic markers of biological aging

Proteomic aging clocks have emerged as biomarkers of aging, but most rely on cross-sectional data. Here, we present a proteomic aging clock based on longitudinal measurements of liquid chromatography-mass spectrometry plasma proteomics at up to three time points in the population-based KORA cohort, comprising 9,105 samples from 4,100 individuals in the final dataset aged 25-88 years. The model predicts chronological age and yields a measure of proteomic age acceleration associated with lifestyle and clinical risk factors. Risk factor-stratified analyses reveal shared proteomic signatures among cardiometabolic exposures and distinct signatures for smoking and physical activity. Within individuals, dyslipidemia and impaired kidney function track transitions to increases in age acceleration, while higher age acceleration is associated with increased all-cause mortality. Mediation analysis suggests that risk factors are linked to mortality indirectly through age acceleration. Affected proteins and pathways highlight extracellular matrix remodeling, complement and coagulation pathways, and liver-derived plasma proteins as key contributors to aging. Mendelian randomization identifies proteins with evidence consistent with a causal role in aging-related traits. Together, these findings are consistent with biological aging being dynamically responsive to changes in modifiable risk factors.

bioinformatics↗

bioRxiv2026

Spatially Integrated Multi-Omics reveals the Multicellular Landscape of Progenitor-Driven Glioblastoma Progression

Glioblastoma is the most lethal primary brain tumor, driven by complex interactions between plastic malignant cells and a diverse tumor microenvironment. Despite advances in single-cell profiling, how genomic drivers and the tumor microenvironment interact to determine tumor progression and patient survival remains poorly understood. While cellular states have been cataloged, the multi-cellular logic coordinating these into lethal phenotypes remains unresolved. Here, we integrate whole-exome sequencing, bulk and single-cell RNA sequencing, spatial transcriptomics, and histopathology from the MOSAIC cohort (n = 89) to deconstruct this inter- and intra-tumoral heterogeneity. We identify a robust latent multi-omic program in glioblastoma that delineates a transition from homeostatic neural precursors to an aggressive, immunosuppressive progenitor phenotype. This high-risk state, which predicts poor survival in both the MOSAIC discovery and independent validation cohorts (TCGA, CGGA; total n = 598), is sustained by dense intercellular communication networks linking malignant progenitors with myeloid and endothelial compartments. Spatially, this program maps to hypoxic, perinecrotic niches, directly linking molecular signaling with microvascular proliferation and tissue necrosis. Our findings provide a spatially resolved, multi-omic blueprint of the multicellular logic driving glioblastoma progression, offering a robust molecular framework for patient stratification and targeted intervention.

cancer biology↗

bioRxiv2026

Federated single-cell QTL meta-analysis reveals novel disease mechanisms

Genetic effects on gene expression are often cell type-specific and obscured in bulk analyses. To resolve this context-dependent regulation, we performed a federated cis-eQTL meta-analysis across 12 PBMC datasets (2,032 individuals, 2.5 million cells). Across six immune cell types, we identified cis-eQTLs for 6,592 genes and fine-mapped 14,985 independent loci. Notably, the 42% of eQTLs that were undetected in a bulk eQTL study on 43,301 whole blood samples also showed stronger enrichment for disease GWAS loci. We further identified three genome-wide significant and 65 suggestive loci affecting the abundance of (rare) immune cell types and validated these using previously reported hematological GWAS and bulk-derived trans-eQTLs. Integrating single-cell cis-eQTLs with bulk trans-eQTLs enabled us to anchor 6,382 trans-eGenes (37.2% novel) to upstream regulators and reconstruct directed gene regulatory relationships. For example, a hemorrhoidal disease-associated variant showed a CD4+ T cell-specific cis-eQTL on BACH1 that colocalized with 45 immune and metabolic trans-eGenes. These results demonstrate the power of single-cell QTL meta-analysis in interpreting complex trait genetics.

genetics↗

bioRxiv2025

Hidden sampling biases inflate performance in gene regulatory network inference

Accurate reconstruction of gene regulatory networks (GRNs) from single-cell transcriptomic data remains a major methodological challenge. Recent machine learning approaches, particularly graph neural networks and graph autoencoders, have reported improved performance, yet these gains do not consistently translate to realistic biological settings. Here, we show that a key reason for that is the way negative regulatory interactions are sampled for supervised training and evaluation. We find that widely used sampling strategies introduce node-degree biases that allow models to exploit trivial graph-structural cues rather than biological signals. Across multiple benchmarks, simple degree-based heuristics match or exceed state-of-the-art graph neural network models under these biased evaluation protocols. We further introduce a degree-aware sampling approach that eliminates these artifacts and provides more reliable assessments of GRN inference methods. Our results call for standardized, bias-aware benchmarking practices to ensure meaningful progress in supervised GRN inference from single-cell RNA-seq data.

bioinformatics↗

bioRxiv2025

Disease-associated variants are enriched for altering cell-type-specific gene co-expression relationships

Genes act within complex regulatory networks, and genetic variants can perturb these networks by altering gene co-expression. Here, we performed co-expression quantitative trait locus (co-eQTL) mapping using single-cell RNA-seq from the sc-eQTLGen consortium (1,330 donors, >2 million cells), enabling sensitive detection and prioritization of informative variant-gene-gene triplets. We identified co-eQTLs for 398 eGenes where a nearby genetic variant affected both the genes expression (cis-gene) and its co-expression with other genes, often implicating upstream regulators. For 181 genes, we inferred a likely upstream transcription factor, with motif disruption predicted for 41 genes. These upstream genes are more often loss-of-function intolerant and show more network connections, providing an explanation for why co-eQTL variants are 2.8x more strongly associated with immune diseases than classical eQTLs. These findings position co-eQTLs as mechanistic links between genetic variation and disease, revealing how variants can rewire cell-type-specific gene networks.

genetics↗

bioRxiv2024

Tissue-specific apparent mtDNA heteroplasmy and its relationship with ageing and mtDNA gene expression

Heteroplasmy in the mitochondrial DNA (mtDNA) accumulates with age, but how much it occurs in tissues and how it affects function and physiology is not well-studied. Here we present a comprehensive tissue-specific map of mtDNA heteroplasmy and mtRNA modifications, and their relationships with age and mitochondrial gene expression. We propose a robust variant calling pipeline for mtDNA heteroplasmy and mtRNA modifications using bulk RNAseq data from 49 tissues in GTEx, and a calibrated phenotype association framework for both types of variants. We identify 109 associations between them and donor age, and 784 associations with tissue-specific mtDNA gene expression. Of these, 7 and 18 show cell-type specificity within tissue. In addition, we find 9 instances where these variations mediate donor age effects on mtDNA gene expression. Finally, we confirm previously identified relationships between mt-tRNA modifications and gene expression on their 5, but read- and gene-level investigations reveal previously unmodeled complexities.

genomics↗

bioRxiv2024

Interplay between mitochondrial and nuclear DNA in gene expression regulation

Using data from 48 tissues and 684 European individuals in the GTEx dataset, we investigate the role of mitochondrial DNA (mtDNA) and its encoded genes in gene regulation. We perform a comprehensive multi-tissue eQTL analysis on mtDNA encoded protein-coding genes, identifying 111 mtDNA-cis-eQTLs (FDR<5%) and 260 nucDNA trans-eQTLs (P<5x10-8). We further identify 108 mtDNA associations with 68 nucDNA encoded genes (FDR<5%). Our results are not driven by nuclear mitochondrial sequences (NUMTs) or minority cell types within tissues. Incorporating mtDNA trans-eQTLs in gene expression networks improves prediction of nucDNA genes with mitochondrial function. nucDNA trans-eQTLs for mtDNA genes are enriched in genes involved in mitochondrial pathways and GWAS hits for complex traits and diseases, implicating dual regulation by both genomes in maintaining mitochondrial function and organismal health. Our multi-tissue map of dual-genome gene regulation is an essential step towards understanding mito-nuclear cross-talk in health and disease.

genomics↗

bioRxiv2024

Integrated single-cell atlas of human atherosclerotic plaques

Atherosclerosis, a major cause of cardiovascular diseases, is characterized by the buildup of lipids and chronic inflammation in the arteries, leading to plaque formation and potential rupture. The underlying causal immune mechanisms and alterations in structural cell composition and plasticity driving plaque progression remain incompletely defined. Recent advances in single-cell transcriptomics (scRNA-seq) have provided deeper insights into the roles of immune and non-immune cells in atherosclerosis. However, existing public scRNA-seq datasets often lack comprehensive cell type coverage and consistent annotations, posing challenges for downstream analyses. In this study, we present an integrated single-cell atlas of human atherosclerotic plaques, encompassing 261,747 high-quality annotated cells from carotid, coronary, and femoral arteries. By benchmarking and applying the best-performing data integration method, scPoli, we achieved robust cell type annotations validated by expert consensus and surface protein measurements. This comprehensive atlas enables accurate automatic cell type annotation of new datasets, optimal experimental design, and deconvolution of existing as well as novel bulk RNA-seq data to comprehensively determine cell type proportions in human atherosclerotic lesions. It facilitates future studies by providing an interactive WebUI for easy data annotation and experimental design, while supporting various downstream applications, including integration of genetic association studies and experimental planning.

cell biology↗

bioRxiv2024

Investigating the performance of foundation models on human 3'UTR sequences

Foundation models, such as DNABERT and Nucleotide Transformer have recently shaped a new direction in DNA research. Trained in an unsupervised manner on a vast quantity of genomic data, they can be used for a variety of downstream tasks, such as promoter prediction, DNA methylation prediction, gene network prediction or functional variant prioritization. However, these models are often trained and evaluated on entire genomes, neglecting genome partitioning into different functional regions. In our study, we investigate the efficacy of various unsupervised approaches, including genome-wide and 3UTR-specific foundation models on human 3UTR regions. Our evaluation includes downstream tasks specific for RNA biology, such as recognition of binding motifs of RNA binding proteins, detection of functional genetic variants, prediction of expression levels in massively parallel reporter assays, and estimation of mRNA half-life. Remarkably, models specifically trained on 3UTR sequences demonstrate superior performance when compared to the established genome-wide foundation models in three out of four downstream tasks. Our results underscore the importance of considering genome partitioning into functional regions when training and evaluating foundation models. In addition, the proposed set of 3UTR-specific tasks can be used for benchmarking future models.

bioinformatics↗

bioRxiv2023

A gut meta-interactome map reveals modulation of human immunity by microbiome effectors

The molecular mechanisms by which the gut microbiome influences human health remain largely unknown. Pseudomonadota is the third most abundant phylum in normal gut microbiomes. Several pathogens in this phylum can inject so-called virulence effector proteins into host cells. We report the identification of intact type 3 secretion systems (T3SS) in 5 - 20% of commensal Pseudomonadota in normal human gut microbiomes. To understand their functions, we experimentally generated a high-quality protein-protein meta-interactome map consisting of 1,263 interactions between 289 bacterial effectors and 430 human proteins. Effector targets are enriched for metabolic and immune functions and for genetic variation of microbiome-influenced traits including autoimmune diseases. We demonstrate that effectors modulate NF-{kappa}B signaling, cytokine secretion, and adhesion molecule expression. Finally, effectors are enriched in metagenomes of Crohns disease, but not ulcerative colitis patients pointing toward complex contributions to the etiology of inflammatory bowel diseases. Our results suggest that effector-host protein interactions are an important regulatory layer by which the microbiome impacts human health.

microbiology↗

bioRxiv2023

Multiplexing cortical brain organoids for the longitudinal dissection of developmental traits at single cell resolution

The combination of brain organoid and single cell omic technologies holds transformative potential to dissect human neurobiology at high resolution and with mechanistic precision. Delivering this promise in the context of human neurodiversity, physiological and pathological alike, requires however a major leap in scalability, given the need for experimental designs that include multiple individuals and, prospectively, population cohorts. To lay the foundation for this, we implemented and benchmarked complementary strategies to multiplex brain organoids. Following an extended longitudinal design with a uniquely informative set of timepoints, we pooled cells from different induced pluripotent stem cell lines either during organoids generation (upstream multiplexing in mosaic models) or before single cell-RNAseq library preparation (downstream multiplexing). We developed a new method, SCanSNP, and an aggregated call to deconvolve organoids cell identities, overcoming current criticalities in doublets prediction and low quality cells identification and improving accuracy over state of the art algorithms. Integrating single cell transcriptomes and analysing cell types across neurodevelopmental stages and multiplexing modalities, we validated the feasibility of both multiplexing methods in charting neurodevelopmental trajectories at high resolution, linking their specificity to genetic variation between individual lines. Together, this multiplexing suite of experimental and computational methods provides an enabling resource for disease modelling at scale and paves the way towards an in vitro epidemiology paradigm.

neuroscience↗

bioRxiv2023

Speos: An ensemble graph representation learning framework to predict core genes for complex diseases

Understanding phenotype-to-genotype relationships is a grand challenge of 21st century biology with translational implications. The recently proposed "omnigenic" model postulates that effects of genetic variation on traits are mediated by core-genes and -proteins whose activities mechanistically influence the phenotype, whereas peripheral genes encode a regulatory network that indirectly affects phenotypes via core gene products. We have developed a positive-unlabeled graph representation-learning ensemble-approach to predict core genes for diverse diseases using Mendelian disorder genes for training. Employing mouse knockout phenotypes for external validation, we demonstrate that our most confident predictions validate at rates on par with the Mendelian disorder genes, and all candidates exhibit core-gene properties like transcriptional deregulation in diseases and loss-of-function intolerance. Predicted candidates are enriched for drug targets and druggable proteins and, in contrast to Mendelian disorder genes, also for druggable but yet untargeted gene products. Model interpretation suggests key molecular mechanisms and physical interactions for core gene predictions. Our results demonstrate the potential of graph representation learning and pave the way for studying core gene properties and future drug development.

genetics↗

bioRxiv2022

Predictive model of transcriptional elongation control identifies trans regulatory factors from chromatin signatures

Promoter-proximal Polymerase II (Pol II) pausing is a key rate limiting step for gene expression. DNA and RNA-binding trans-acting factors regulating the extent of pausing have been identified. However, we lack a quantitative model of how interactions of these factors determine pausing, therefore the relative importance of implicated factors is unknown. Moreover, previously unknown regulators might exist. Here we address this gap with a machine learning model that accurately predicts the extent of promoter proximal Pol II pausing from large scale genome and transcriptome binding maps, as well as gene annotation and sequence composition features. We demonstrate high accuracy and generalizability of the model by validation on an independent cell line which reveals the models cell line agnostic character. Model interpretation in light of prior knowledge about molecular functions of regulatory factors confirms the interconnection of pausing with other RNA processing steps. Harnessing underlying feature contributions we assess the relative importance of each factor, quantify their predictive effects and systematically identify previously unknown regulators of pausing. We additionally identify 16 previously unknown 7SK ncRNA interacting RNA-binding proteins predictive of pausing. Our work provides a framework to further our understanding of the regulation of the critical early steps in transcriptional elongation. Key Points: Please provide 3 bullet points summarizing the manuscripts contribution to the field- ML model that accurately predicts promoter proximal Pol II pausing from ChIP and eClip-seq data - Quantification of the interconnection of pausing and other steps of gene regulation - Identification of novel putative trans regulators of pausing GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=77 SRC="FIGDIR/small/496993v1_ufig1.gif" ALT="Figure 1"> View larger version (21K): org.highwire.dtl.DTLVardef@1c7dfc3org.highwire.dtl.DTLVardef@133886eorg.highwire.dtl.DTLVardef@41f694org.highwire.dtl.DTLVardef@a396af_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

bioRxiv2022

Identification of genetic variants that impact gene co-expression relationships using large-scale single-cell data

BackgroundExpression quantitative trait loci (eQTL) studies have shown how genetic variants affect downstream gene expression. To identify the upstream regulatory processes, single-cell data can be used. Single-cell data also offers the unique opportunity to reconstruct personalized co-expression networks--by exploiting the large number of cells per individual, we can identify SNPs that alter co-expression patterns (co-expression QTLs, co-eQTLs) using a limited number of individuals. ResultsTo tackle the large multiple testing burden associated with a genome-wide analysis (i.e. the need to assess all combinations of SNPs and gene pairs), we conducted a co-eQTL meta-analysis across four scRNA-seq peripheral blood mononuclear cell datasets from three studies (reflecting 173 unique participants and 1 million cells) using a novel filtering strategy followed by a permutation-based approach. Before analysis, we evaluated the co-expression patterns to be used for co-eQTL identification using different external resources. The subsequent analysis identified a robust set of cell-type-specific co-eQTLs for 72 independent SNPs that affect 946 gene pairs, which we then replicated in a large bulk cohort. These co-eQTLs provide novel insights into how disease-associated variants alter regulatory networks. For instance, one co-eQTL SNP, rs1131017, that is associated with several autoimmune diseases affects the co-expression of RPS26 with other ribosomal genes. Interestingly, specifically in T cells, the SNP additionally affects co-expression of RPS26 and a group of genes associated with T cell-activation and autoimmune disease. Among these genes, we identified enrichment for targets of five T-cell-activation-related transcriptional factors whose binding sites harbor rs1131017. This reveals a previously overlooked process and pinpoints potential regulators that could explain the association of rs1131017 with autoimmune diseases. ConclusionOur co-eQTL results highlight the importance of studying gene regulation at the context-specific level to understand the biological implications of genetic variation. With the expected growth of sc-eQTL datasets, our strategy--combined with our technical guidelines--will soon identify many more co-eQTLs, further helping to elucidate unknown disease mechanisms.

genetics↗

bioRxiv2022

Neural network approach to somatic SNP calling in WGS samples without a matched control.

Somatic variants are usually called by analysing the DNA sequences of a tumor sample in conjunction with a matched normal. However, a matched normal is not always available for instance in diagnostic settings. To unlock such data for basic research single-sample somatic variant calling is required. Previous approaches can not easily be applied in the case of typical whole genome sequencing (WGS) samples.We present a neural network-based approach for calling somatic single nucleotide polymorphism (SNP) variants in tumor WGS samples without a matched normal. The method does not require any manual tuning of filtering parameters and can be applied under the conditions of a typical WGS experiment. We demonstrate the effectiveness of the proposed approach by reporting its performance on 5 SNP datasets corresponding to 5 different cancer types. The proposed method is implemented in Python 3.6 and available as a GitHub repository at https://github.com/heiniglab/deepSNP.

bioinformatics↗

bioRxiv2020

Identification and characterization of distinct murine brown adipocyte lineages

Brown adipose tissue (BAT) plays an important role in the regulation of body weight and glucose homeostasis. While increasing evidence supports white adipose tissue heterogeneity, little is known about heterogeneity within murine BAT. Using single cell RNA sequencing of the stromal vascular fraction of murine BAT and analysis of 67 brown preadipocyte and adipocyte clones we unravel heterogeneity within brown preadipocytes. Statistical analysis of gene expression profiles from these clones identifies markers distinguishing brown adipocyte lineages. We confirm the presence of distinct brown adipocyte populations in vivo using three identified markers; Eif5, Tcf25, and Bin1. Functionally, we demonstrate that loss of Bin1 enhances UCP1 expression and mitochondrial respiration, suggesting that Bin1 marks a dormant brown adipocyte type. The existence of multiple brown adipocyte lineages suggests distinct functional properties of BAT depending on its cellular composition, with potentially distinct function in thermogenesis and the regulation of whole body energy homeostasis.

cell biology↗

bioRxiv2020

Trans control of cardiac mRNA translation in a protein length-dependent fashion

Little is known about the impact of naturally occurring genetic variation on the rates with which proteins are synthesized by ribosomes. Here, we investigate how genetic influences on mRNA translational efficiency are associated with complex disease phenotypes using a panel of rat recombinant inbred lines. We identify a locus for cardiac hypertrophy that is associated with a translatome-wide and protein length-dependent shift in translational efficiency. This master regulator primarily affects the translation of very short and very long protein-coding sequences, altering the physiological stoichiometric translation rates of sarcomere proteins. Mechanistic dissection of this locus points to altered ribosome assembly, characterized by accumulation of polysome half-mers, changed ribosomal configurations and misregulation of the small nucleolar RNA SNORA48. We postulate that this locus enhances a pre-existing negative correlation between protein length and translation initiation in diseased hearts. Our work shows that a single genomic locus can trigger a complex, translation-driven molecular mechanism that contributes to phenotypic variability between individuals. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/133298v1_ufig1.gif" ALT="Figure 1"> View larger version (62K): org.highwire.dtl.DTLVardef@e46986org.highwire.dtl.DTLVardef@1cc0b08org.highwire.dtl.DTLVardef@e86011org.highwire.dtl.DTLVardef@1fb3706_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIGenetic variability impacts protein synthesis rates in a rat model for cardiac hypertrophy C_LIO_LIA trans locus affects stoichiometric translation rates of cardiac sarcomeric proteins C_LIO_LIThis master regulator locus induces a global, protein length-dependent shift in translation C_LIO_LIDysregulated ribosome assembly induces half-mer formation and affects translation initiation rate C_LI

genetics↗

bioRxiv2020

Network reconstruction for trans acting genetic loci using multi-omics data and prior information

BackgroundMolecular multi-omics data provide an in-depth view on biological systems, and their integration is crucial to gain insights in complex regulatory processes. These data can be used to explain disease related genetic variants by linking them to intermediate molecular traits (quantitative trait loci, QTL). Molecular networks regulating cellular processes leave footprints in QTL results as so-called trans -QTL hotspots. Reconstructing these networks is a complex endeavor and use of biological prior information has been proposed to alleviate network inference. However, previous efforts were limited in the types of priors used or have only been applied to model systems. In this study, we reconstruct the regulatory networks underlying trans -QTL hotspots using human cohort data and data-driven prior information. ResultsWe devised a strategy to integrate QTL with human population scale multi-omics data and comprehensively curated prior information from large-scale biological databases. State-of-the art network inference methods applied to these data and priors were used to recover the regulatory networks underlying trans -QTL hotspots. We benchmarked inference methods and showed, that Bayesian strategies using biologically-informed priors outperform methods without prior data in simulated data and show better replication across datasets. Application of our approach to human cohort data highlighted two novel regulatory networks related to schizophrenia and lean body mass for which we generated novel functional hypotheses. ConclusionWe demonstrate, that existing biological knowledge can be leveraged for the integrative analysis of networks underlying trans associations to deduce novel hypotheses on cell regulatory mechanisms.

bioinformatics↗