Biology subjects

Teschendorff, A.

Publications and source records attributed to Teschendorff, A..

At least 19 recordsLinked to original sources

bioRxiv2026

Quantifying epigenetic reprogramming and cellular plasticity in preneoplastic lesions reveals dedifferentiation as a key precancer hallmark

Epigenetic reprogramming and plasticity are two cancer hallmarks that increase the risk of carcinogenic transformation, yet quantifying these properties remains a fundamental challenge. Here we develop and validate a general pan-tissue computational framework called PLASTIC for quantifying cellular plasticity and epigenetic reprogramming from single-cell RNA-seq data. PLASTIC allows construction of Waddington-like landscapes allowing identification of high cancer risk states. When applied to a pan-preneoplastic tissue setting, PLASTIC reveals how the level of epigenetic reprogramming of precancerous cells correlates with cellular plasticity and cancer-risk, and that the high cancer risk state is characterized by increased dedifferentiation, even when compared to its fetal cell counterpart. Furthermore, PLASTIC predicts the emergence of metaplasia in normal lung cells from smokers, identifying a high cancer risk signature that we validate in independent data. In summary, PLASTIC can help quantify cellular plasticity and dedifferentiation of single cells in any tissue-type, opening up new strategies for cancer risk prediction and prevention.

cancer biology↗

bioRxiv2026

DNA methylation-based prediction of differentiation state at cell-type resolution reveals an age-associated epigenetic mesenchymal drift

Quantifying the differentiation state of cells is an important task in regenerative medicine, cellular reprogramming and in elucidating how differentiation hierarchies are altered in aging and disease. For DNA methylation data (DNAm) there is however an accute shortage of methods for quantifying differentiation state. Here we use machine learning to build EpiDiff, a predictor of differentiation that yields a continuous absolute differentiation index (DI) when applied to a DNAm sample. We extensively validate EpiDiff in single-cells, sorted cells and bulk tissue, as well as in in-vitro differentiation, organoid, reprogramming and cellular rejuvenation experiments. In mixed samples, EpiDiff measures the extent of differentiation across all cellular lineages. In sorted cell populations, we find that DI increases with age and, underpinning this increase, we discover that many cell-types undergo a developmental epigenetic drift that is most prominent towards the mesenchymal lineage. When applied to normal-tissues, EpiDiff is predictive of carcinogenic exposure and cancer risk. In summary, EpiDiff is a unique, general and versatile tool that can accurately quantify the global differentiation state of a DNAm sample with broad applicability to many disciplines.

bioinformatics↗

bioRxiv2025

Optimizing effect sizes and specificity trumps machine learning when building DNA methylation reference panels for cell-type deconvolution

Accurate cell-type deconvolution is critical for correct interpretation of Epigenome-Wide Association Studies. Such cell-type deconvolution involves estimating underlying cell-type fractions in a sample, which is accomplished using a DNA methylation reference panel built from sorted or single-cell DNAm data. Two competing approaches have emerged to build such reference panels, one which uses machine-learning, and another based on optimizing effect size and cell-type specificity. Here we demonstrate that the latter approach is preferable, because, owing to the relatively small number of sorted samples used in building panels, standard machine learning does not optimize effect size and cell-type specificity, causing the model to overfit and underperform when tested in independent data. Furthermore, adult blood panels built from cell-type specific hypomethylated markers improves inference when compared to panels built from hypermethylated ones. These insights provide important guidelines for optimizing the construction of future DNAm reference panels. To aid this task, we have added a function for building an optimized DNAm reference panel to our EpiDISH R-package.

bioinformatics↗

bioRxiv2025

Sex-specific non-linear DNA methylation trajectories across aging predict cancer risk and systemic inflammation

Aging is a multi-modal process, leaving distinct signatures across molecular layers, including the epigenome. DNA methylation changes are among the most robust markers of biological aging. Yet, most studies rely on models assuming linear relationships with age and often analyze mixed-sex cohorts, overlooking well-known sex differences in the timing and nature of aging phases. Such approaches risk obscuring critical, non-linear transitions and sex-specific trajectories that may better capture the biology of aging. We developed a computational approach to detect complex, non-linear trajectories and disentangle shared from sex-divergent patterns. Applied to whole-blood deconvoluted methylomes from 252 females and 246 males spanning ages 19-90 years, this analysis revealed convergent and divergent epigenetic aging pathways independent of immune cell composition. These non-linear trajectories were enriched for developmental transcription factor binding motifs, including NF1/CTF and REST, which are known for their oncogenic potential. Strikingly, a female-specific non-linear cluster was robustly associated with cancer onset and systemic inflammation. Our results uncover sex-specific, non-linear aging programs that better capture the dynamics of epigenetic change than linear models. These findings nominate candidate biomarkers for early disease risk and offer mechanistic insight into how aging trajectories diverge between the sexes.

molecular biology↗

bioRxiv2025

DNA Methylation Ageing Atlas Across 17 Human Tissues

Aging involves widespread epigenetic remodeling across tissues, yet the nature and consistency of these changes remain unclear. We conducted a meta-analysis of more than 15,000 human methylomes spanning 17 tissues, identifying both conserved and tissue-specific aging signatures. We examined linear changes via differentially methylated positions, variability shifts via variably methylated positions, and Shannon-entropy to capture methylation disorder. Network analysis revealed fragile co-methylation modules largely resistant to beneficial perturbation. Key disruptors, including PCDHGA1, MEST, HDAC4, and HOX genes, exacerbated aging signals across tissues. Notably, a resilient module enriched for NAD{square} salvage metabolism supports therapeutic targeting of NAD{square} in aging. PCDHGA1 emerged as a conserved cross-tissue driver, suggesting protocadherin-mediated adhesion plays a broader role in maintaining structural and signaling stability in multiple organ systems. Our open-access atlas provides a foundational resource for dissecting the molecular architecture of human aging and identifying testable targets for intervention, biomarkers, and translational epigenetic therapies.

bioinformatics↗

bioRxiv2025

Unified high-resolution immune cell fraction estimation in blood tissue from birth to old age

BackgroundBlood is composed of many different immune cell-types, whose proportions vary throughout life. If not accounted for, these variations can seriously cause confounding or hamper interpretation of DNAm-based biomarkers. Although cell-type deconvolution methods can address this challenge for cord and adult blood, there is currently no method that can be applied to blood tissue from other age groups, including infants and children. ResultsHere we construct and extensively validate a DNAm reference panel, called UniLIFE, for 19 immune cell-types, applicable to blood tissue of any age. We use UniLIFE to delineate the dynamics of immune-cell fractions from birth to old age, and to infer disease associated immune cell fraction variations in newborns, infants, children and adults. In a prospective longitudinal study of type-1 diabetes in infants and children, UniLIFE identifies differentially methylated positions that precede type-1 diabetes diagnosis and that map to diabetes related signaling pathways. In contrast to previous studies, we are able to validate these biomarker associations in an independent DNAm dataset comprising purified monocytes from monozygotic twins discordant for type-1 diabetes, but not in lymphocytes, highlighting the importance of epigenetic changes in the innate immune system in the development of type-1 diabetes. ConclusionsIn summary, we present a life course immune-cell estimator for blood tissue of any age to help improve the identification and interpretation of blood-based DNAm biomarkers for any age groups and specially for longitudinal studies that include infants and children. The UniLIFE DNAm reference panel and algorithms to estimate cell-type fractions and perform cell-type deconvolution are available from our EpiDISH Bioconductor R-package: https://bioconductor.org/packages/release/bioc/html/EpiDISH.html

bioinformatics↗

bioRxiv2025

Epigenetic clocks and inflammaging: pitfalls caused by ignoring cell-type heterogeneity

Epigenetic clocks are machine learning predictors of chronological age that have consistently been shown to be also informative of biological age measures, such as all-cause mortality. A recent study has argued against the use of machine learning methods for building epigenetic clocks, on grounds that these predictors do not enrich for features that correlate with disease. In particular, it argues that standard epigenetic clocks cant measure inflammaging, an age-related trait, proposing an ad-hoc feature rectification strategy and an inflammation clock that predicts age-acceleration in inflammatory diseases. Here we demonstrate that this inflammation clock only captures an increase in the neutrophil to lymphocyte ratio associated with inflammatory diseases like rheumatoid arthritis, and that it fails to capture the shift from naive to mature lymphocytes, which is a key feature of inflammaging. As such, their clock does not measure inflammaging, and is subsumed and outperformed by an ordinary cell-type deconvolution algorithm. Our analysis underscores the critical importance for epigenetic clock and epigenome-wide association studies to always estimate cell-type fractions in the tissue of consideration, and to adjust for its variation, in order to disentangle effects due to changing cell-type composition from DNA methylation changes that happen in specific cell-types.

bioinformatics↗

bioRxiv2024

Cell-type specific epigenetic clocks to quantify biological age at cell-type resolution

The ability to accurately quantify biological age could help monitor and control healthy aging. Epigenetic clocks have emerged as promising tools for estimating biological age, yet they have been developed from heterogeneous bulk tissues, and are thus composites of two aging processes, one reflecting the change of cell-type composition with age and another reflecting the aging of individual cell-types. There is thus a need to dissect and quantify these two components of epigenetic clocks, and to develop epigenetic clocks that can yield biological age estimates at cell-type resolution. Here we demonstrate that in blood and brain, approximately 39% and 12% of an epigenetic clocks accuracy is driven by underlying shifts in lymphocyte and neuronal subsets, respectively. Using brain and liver tissue as prototypes, we build and validate neuron and hepatocyte specific DNA methylation clocks, and demonstrate that these cell-type specific clocks yield improved estimates of chronological age in the corresponding cell and tissue-types. We find that neuron and glia specific clocks display biological age acceleration in Alzheimers Disease with the effect being strongest for glia in the temporal lobe. Moreover, CpGs from these clocks display a small but significant overlap with the causal DamAge-clock, mapping to key genes implicated in neurodegeneration. The hepatocyte clock is found accelerated in liver under various pathological conditions. In contrast, non-cell-type specific clocks do not display biological age-acceleration, or only do so marginally. In summary, this work highlights the importance of dissecting epigenetic clocks and quantifying biological age at cell-type resolution.

bioinformatics↗

bioRxiv2023

A comprehensive map of the ageing blood methylome

During ageing, the human methylome exhibits both differential (i.e. change in mean) and variable (i.e. change in variance) shifts, along with a general rise in entropy. However, it remains unclear whether DNA methylation sites that increasingly diverge between people (i.e. variably methylated positions (VMPs)) are distinct from those undergoing changes in mean methylation levels (i.e. differentially methylated positions (DMPs)), which changes drive entropy, how they contribute to epigenetic age measured by epigenetic clocks, and whether cell type heterogeneity plays a role in these alterations. To address these questions, we conducted a comprehensive analysis using > 32,000 human blood methylomes from 56 datasets (age range = 6-101 years). Our findings revealed an unprecedented proportion of the blood methylome that is differentially methylated with age (48% DMPs; FDR< 0.005) and variably methylated with age (37% VMPs; FDR< 0.005), with many sites overlapping between the two groups (59% of DMPs are VMPs). We observed that bivalent and Polycomb regions become increasingly methylated and divergent between individuals, while quiescent regions lose methylation in a more homogeneous manner between individuals. Unexpectedly, both chronological and biological clocks, but not pace-of-aging clocks, show a strong enrichment for those CpGs that accrue both mean and variance changes during aging. Furthermore, we uncovered that it is the accumulation of DMPs shifting towards a methylation fraction of 50% that drive the increase in entropy, resulting in an overall smoothening of the epigenetic landscape. However, approximately a quarter of DMPs oppose this direction of change, exhibiting anti-entropic effects. While DMPs were mostly unaffected by changes in cell type composition, VMPs and entropy measurements showed moderate sensitivity to such alterations. This investigation represents the largest to date of genome-wide DNA methylation changes and ageing in a single tissue, offering valuable insights into primary molecular changes that hold meaning for chronological and biological ageing.

molecular biology↗

bioRxiv2023

Correlation of a pan-tissue epigenetic mitotic clock with tumor cell-of-origin

The cumulative number of stem cell divisions in a tissue, known as mitotic age, is thought to be a major determinant of cancer-risk. Somatic mutational and DNA methylation (DNAm) clocks are promising tools to molecularly track mitotic age, yet their relationship is underexplored and their potential for cancer risk prediction in normal tissues remains to be demonstrated. Here we build and validate an improved pan-tissue DNAm counter of total mitotic age called stemTOC. We demonstrate that stemTOCs mitotic age proxy increases with the tumor cell-of-origin fraction in each of 15 cancer-types, in precancerous lesions, and in normal tissues exposed to major cancer risk factors. Extensive benchmarking against 6 other mitotic counters shows that stemTOC compares favorably, specially in the preinvasive and normal-tissue contexts. By cross-correlating stemTOC to two clock-like somatic mutational signatures, we confirm the mitotic-like nature of only one of these. Our data points towards DNAm as a promising molecular substrate for detecting mitotic-age increases in normal tissues and precancerous lesions, and hence for developing cancer-risk prediction strategies.

cancer biology↗

bioRxiv2022

FORGEdb: systematic analysis of candidate causal variants to uncover target genes and mechanisms in complex traits.

The majority of disease-associated variants identified through genome-wide association studies (GWAS) are located outside of protein-coding regions and are overrepresented in sequences that regulate gene expression. Prioritizing candidate regulatory variants and potential biological mechanisms for further functional experiments, such as genome editing, can be challenging, especially in regions with a high number of variants in strong linkage disequilibrium or multiple proximal gene targets. Improved annotation of the regulatory genome can help identify promising variants and target genes for functional genomics experiments. To advance this area, we developed FORGEdb (https://forge2.altiusinstitute.org/files/forgedb.html), a web-based tool that can rapidly integrate data for individual genetic variants, providing information on associated regulatory elements, transcription factor (TF) binding sites and target genes for over 37 million variants. FORGEdb uses annotations derived from data across a wide range of biological samples to delineate the regulatory context for each variant at the cell type level. Multiple data types, such as Combined Annotation Dependent Depletion (CADD) scores, expression quantitative trait loci (eQTLs), activity-by-contact (ABC) interactions, Contextual Analysis of TF Occupancy (CATO) scores, transcription factor (TF) motifs, DNase I hotspots, histone mark ChIP-seq peaks and chromatin states, are included in FORGEdb and these annotations are integrated into a FORGEdb score to guide assessment of functional importance. In summary, FORGEdb provides an expansive and unique resource of genomic annotations and an integrated score that can be used to accelerate the translation of identified genetic loci into biological insight.

genomics↗

bioRxiv2021

Calling differential DNA methylation at cell-type resolution: avoiding misconceptions and promoting best practices

The accurate detection of cell-type specific DNA methylation alterations in the context of general epigenome studies is an important task to improve our understanding of epigenomics in disease development. Although a number of statistical algorithms designed to address this problem have emerged, the task remains challenging. Here we show that a recent commentary by Rahmani et al, that aims to address misconceptions and best practices in the field, continues to suffer from critical misconceptions in how statistical algorithms should be compared and evaluated. In addition, we report contradictory results on real EWAS datasets.

bioinformatics↗

bioRxiv2019

Calling differential DNA methylation at cell-type resolution: an objective status-quo

Due to cost and logistical reasons, Epigenome-Wide-Association Studies (EWAS) are normally performed in complex tissues, resulting in average DNA methylation profiles over potentially many different cell-types, which can obscure important cell-type specific associations with disease. Identifying the specific cell-types that are altered is a key hurdle for elucidating causal pathways to disease, and consequently statistical algorithms have recently emerged that aim to address this challenge. Comparisons between these algorithms are of great interest, yet here we find that the main comparative study so far was substantially biased and potentially misleading. By using this study as an example, we highlight some of the key issues that need to be considered to ensure that future assessments between methods are more objective.

bioinformatics↗

bioRxiv2018

ebayGSEA: An improved Gene Set Enrichment Analysis method for Epigenome-Wide-Association Studies

MotivationThe biological interpretation of differentially methylated sites derived from Epigenome-Wide-Association Studies remains a significant challenge. Gene Set Enrichment Analysis (GSEA) is a general tool to help aid biological interpretation, yet its correct and unbiased implementation in the EWAS context is difficult due to the differential probe representation of Illumina Infinium DNA methylation beadchips.\n\nResultsWe present a novel GSEA method, called ebayGSEA, which ranks genes, not CpGs, according to the overall level of differential methylation, as assessed using all the probes mapping to the given gene. Applied on simulated and real EWAS data, we show how ebayGSEA may exhibit higher sensitivity and specificity than the current state-of-the-art, whilst also avoiding differential probe representation bias. Thus, ebayGSEA will be a useful additional tool to aid the interpretation of EWAS data.\n\nAvailability and implementationebayGSEA is available from https://github.com/aet21/ebayGSEA, and has been incorporated into the ChAMP Bioconductor package (https://www.bioconductor.org).

bioinformatics↗

bioRxiv2018

Identification of differentially methylated cell-types in Epigenome-Wide Association Studies

An outstanding challenge of Epigenome-Wide Association Studies (EWAS) performed in complex tissues is the identification of the specific cell-type(s) responsible for the observed differential DNA methylation. Here, we present a novel statistical algorithm, called CellDMC, which is able to identify not only differentially methylated positions, but also the specific cell-type(s) driving the differential methylation. We provide extensive validation of CellDMC on in-silico mixtures of DNA methylation data generated with different technologies, as well as on real mixtures from epigenome-wide-association and cancer epigenome studies. We demonstrate how CellDMC can achieve over 90% sensitivity and specificity in scenarios where current state-of-the-art methods fail to identify differential methylation. By applying CellDMC to a smoking EWAS performed in buccal swabs, we identify differentially methylated positions occurring in the epithelial compartment, which we validate in smoking-related lung cancer. CellDMC may help towards the identification of causal DNA methylation alterations in disease.

bioinformatics↗

bioRxiv2018

Tensorial Blind Source Separation for Improved Analysis of Multi-Omic Data

There is an increased need for integrative analyses of multi-omic data. Although several algorithms for analysing multi-omic data exist, no study has yet performed a detailed comparison of these methods in biologically relevant contexts. Here we benchmark a novel tensorial independent component analysis (tICA) algorithm against current state-of-the-art methods. Using simulated and real multi-omic data, we find that tICA outperforms established methods in identifying biological sources of data variation at a significantly reduced computational cost. Using two independent multi cell-type EWAS, we further demonstrate how tICA can identify, in the absence of genotype information, mQTLs at a higher sensitivity than competing multi-way algorithms. We validate mQTLs found with tICA in an independent set, and demonstrate that approximately 75% of mQTLs are independent of blood cell subtype. In an application to multi-omic cancer data, tICA identifies many gene modules whose expression variation across tumors is driven by copy number or DNA methylation changes, but whose deregulation relative to the normal state is independent such alterations, an important finding that we confirm by direct analysis of individual data types. In summary, tICA is a powerful novel algorithm for decomposing multi-omic data, which will be of great value to the research community.

bioinformatics↗

bioRxiv2018

Appraising the causal relevance of DNA methylation for risk of lung cancer

DNA methylation changes in peripheral blood have been identified in relation to lung cancer risk. However, the causal nature of these associations remains to be fully elucidated. Meta-analysis of four epigenome-wide association studies (918 cases, 918 controls) revealed differential methylation at 16 CpG sites (FDR < 0.05) in relation to lung cancer risk. A two-sample Mendelian randomization analysis, using genetic instruments for methylation at 14 of the 16 CpG sites, and 29,863 cases and 55,586 controls from the TRICL-ILCCO lung cancer consortium, was performed to appraise the causal role of methylation at these sites on lung cancer. This approach provided little evidence that DNA methylation in peripheral blood at the 14 CpG sites play a causal role in lung cancer development, including for cg05575921 AHRR, where methylation is strongly associated with lung cancer risk. Further studies are needed to investigate the causal role played by DNA methylation in lung tissue.

epidemiology↗

bioRxiv2018

Quantifying Waddington’s epigenetic landscape: a comparison of single-cell potency measures

Over 60 years ago Waddington proposed an epigenetic landscape model of cellular differentiation, whereby cell-fate transitions are modelled as canalization events, with stable cell states occupying the basins or attractor states1, 2. A key ingredient of this landscape is the energy potential, or height3, which correlates with cell-potency. To date, very few explicit biophysical models for estimating single-cell potency have been proposed. Using 9 independent experiments, encompassing over 6,600 high-quality single-cell RNA-Seq profiles, we here demonstrate that single-cell potency can be approximated as the graph entropy of a Markov Chain process on a model signaling network. Our analysis highlights that other proposed single-cell potency measures are not robust, whilst also revealing that integration with orthogonal systems-level information improves potency estimates. Thus, this study provides a foundation for an improved systems-level understanding of single-cell potency, which may have profound implications for the discovery of novel stem-and progenitor cell phenotypes.

bioinformatics↗