A pan-tissue DNA-methylation epigenetic clock based on deep learning

Epigenetic ageing clocks: statistical methods and emerging computational challenges

Over the past decade, epigenetic clocks have emerged as powerful machine learning tools, not only to estimate chronological and biological age but also to assess the efficacy of anti-ageing, cellular rejuvenation and disease-preventive interventions. However, many computational and statistical challenges remain that limit our understanding, interpretation and application of epigenetic clocks. Here, we review these computational challenges, focusing on interpretation, cell-type heterogeneity and emerging single-cell methods, aiming to provide guidelines for the rigorous construction of interpretable epigenetic clocks at cell-type and single-cell resolution.

Weighted 2D-kernel density estimations provide a new probabilistic measure for epigenetic age

Epigenetic aging signatures provide insights into human aging, but traditional clocks rely on linear regression of DNA methylation levels, assuming linear trajectories. This study explores a non-parametric approach using 2D-kernel density estimation to determine epigenetic age. Our weighted model achieves similar predictive accuracy as conventional clocks and provides a variation score reflecting the inherent variability of age-related epigenetic changes within samples. This score is significantly increased in various diseases and associated with mortality risk in the Lothian Birth Cohort 1921. Thus, weighted 2D-kernel density estimation facilitates accurate epigenetic age predictions and offers an additional variable for biological age estimation.

Evaluating transcriptional alterations associated with ageing and developing age prediction models based on the human blood transcriptome

Ageing-related DNA methylome and proteome changes and machine-learned ageing clock models have been described previously; however, there is a dearth of ageing clock prediction models based on human blood transcript information. Applying various machine learning algorithms is expected to aid in the development of age prediction models. Using blood transcriptome data from healthy subjects ranging in age from 21 to 90 in the 10 K Immunomes repository, we evaluated differentially regulated transcripts, assessed enriched gene ontology, pathway and disease ontology analysis to characterize biological functions associated with the genes associated with age. Furthermore, we constructed and compared age prediction models developed by applying the Least Absolute Shrinkage and Selection Operator (LASSO), Elastic Net (EN), eXtreme Gradient Boosting (XGBoost) and Light Gradient-Boosting Machine (LightGBM) algorithms. Compared to LASSO (7 genes) and EN (9 genes) regularized regression, XGBoost (142 genes) and LightGBM (149 genes) Gradient Boosted Decision Tree methods performed better in this dataset (training set r = 0.836 (LASSO), 0.837 (EN), 1.000 (XGBoost) and 0.995 (LightGBM); test set: r = 0.883 (LASSO), 0.876 (EN), 0.931 (XGBoost) and 0.915 (LightGBM); external validation set: r = 0.535 (LASSO), 0.534 (EN), 0.591 (XGBoost) and 0.645 (LightGBM)). Blood transcriptome-based age prediction models may provide a simple method to monitor biological ageing, and provide additional molecular insight. Future studies to externally validate these models in various diverse large populations and molecular studies to elucidate the underlying mechanisms by which the gene expression levels may be related to ageing phenotypes would be advantageous.

Functionally enriched epigenetic clocks reveal tissue-specific discordant aging patterns in individuals with cancer

BACKGROUND

Aging is a key risk factor for many diseases, including cancer, and a better understanding of its underlying molecular mechanisms may help to prevent, delay, or treat age-related pathologies. Epigenetic alterations such as DNA methylation (DNAme) changes are a hallmark of aging and form the basis of so-called epigenetic clocks, yet their functional relevance and directionality in different organs during disease development is often unclear.

METHODS

Here, we link cell-specific age-related DNAme changes with three key hallmarks of aging and cancer (senescence, promoter methylation in genes associated with stem cell fate, and dysregulated proliferation) to comprehensively dissect their association with current and future cancer development, carcinogen exposure or preventive measures, and mortality using data in different organs from over 12,510 human and 105 mouse samples, benchmarking against existing epigenetic clocks.

RESULTS

Our findings offer insights into the association of functionally enriched groups of age-related DNAme changes with cancer, identify sites perturbed earliest during carcinogenesis, as well as those distinct between cancer and reprogramming that could inform strategies to prevent teratoma formation upon in vivo reprogramming. Surprisingly, both mouse and human data reveal accelerated aging in breast cancer tissue but decelerated epigenetic aging in some non-cancer surrogate samples from breast cancer patients, in particular cervical samples.

CONCLUSIONS

This work provides evidence for discordant systemic tissue aging in breast cancer.

The Multi-State Epigenetic Pacemaker enables the identification of combinations of factors that influence DNA methylation

Epigenetic clocks, DNA methylation-based predictive models of chronological age, are often utilized to study aging associated biology. Despite their widespread use, these methods do not account for other factors that also contribute to the variability of DNA methylation data. For example, many CpG sites show strong sex-specific or cell-type-specific patterns that likely impact the predictions of epigenetic age. To overcome these limitations, we developed a multidimensional extension of the Epigenetic Pacemaker, the Multi-state Epigenetic Pacemaker (MSEPM). We show that the MSEPM is capable of accurately modeling multiple methylation-associated factors simultaneously, while also providing site-specific models that describe the per site relationship between methylation and these factors. We utilized the MSEPM with a large aggregate cohort of blood methylation data to construct models of the effects of age-, sex-, and cell-type heterogeneity on DNA methylation. We found that these models capture a large faction of the variability at thousands of DNA methylation sites. Moreover, this approach allows us to identify sites that are primarily affected by aging and no other factors. An analysis of these sites reveals that those that lose methylation over time are enriched for CTCF transcription factor chip peaks, while those that gain methylation over time are associated with bivalent promoters of genes that are not expressed in blood. These observations suggest mechanisms that underlie age-associated methylation changes and suggest that age-associated increases in methylation may not have strong functional consequences on cell states. In conclusion, the MSEPM is capable of accurately modeling multiple methylation-associated factors, and the models produced can illuminate site-specific combinations of factors that affect methylation dynamics.

Quantification of Epigenetic Aging in Public Health

Estimators of biological age hold promise for use in preventive medicine, for early detection of chronic conditions, and for monitoring the effectiveness of interventions aimed at improving population health. Among the promising biomarkers in this field are DNA methylation-based biomarkers, commonly referred to as epigenetic clocks. This review provides a survey of these clocks, with an emphasis on second-generation clocks that predict human morbidity and mortality. It explores the validity of epigenetic clocks when considering factors such as race, sex differences, lifestyle, and environmental influences. Furthermore, the review addresses the current challenges and limitations in this research area.

An Update on Neuroaging on Earth and in Spaceflight

Over 400 articles on the pathophysiology of brain aging, neuroaging, and neurodegeneration were reviewed, with a focus on epigenetic mechanisms and numerous non-coding RNAs. In particular, this review the accent is on microRNAs, the discovery of whose pivotal role in gene regulation was recognized by the 2024 Nobel Prize in Physiology or Medicine. Aging is not a gradual process that can be easily modeled and described. Instead, multiple temporal processes occur during aging, and they can lead to mosaic changes that are not uniform in pace. The rate of change depends on a combination of external and internal factors and can be boosted in accelerated aging. The rate can decrease in decelerated aging due to individual structural and functional reserves created by cognitive, physical training, or pharmacological interventions. Neuroaging can be caused by genetic changes, epigenetic modifications, oxidative stress, inflammation, lifestyle, and environmental factors, which are especially noticeable in space environments where adaptive changes can trigger aging-like processes. Numerous candidate molecular biomarkers specific to neuroaging need to be validated to develop diagnostics and countermeasures.

Interpretable deep learning of single-cell and epigenetic data reveals novel molecular insights in aging

Deep learning (DL) and explainable artificial intelligence (XAI) have emerged as powerful machine-learning tools to identify complex predictive data patterns in a spatial or temporal domain. Here, we consider the application of DL and XAI to large omic datasets, in order to study biological aging at the molecular level. We develop an advanced multi-view graph-level representation learning (MGRL) framework that integrates prior biological network information, to build molecular aging clocks at cell-type resolution, which we subsequently interpret using XAI. We apply this framework to one of the largest single-cell transcriptomic datasets encompassing over a million immune cells from 981 donors, revealing a ribosomal gene subnetwork, whose expression correlates with age independently of cell-type. Application of the same DL-XAI framework to DNA methylation data of sorted monocytes reveals an epigenetically deregulated inflammatory response pathway whose activity increases with age. We show that the ribosomal module and inflammatory pathways would not have been discovered had we used more standard machine-learning methods. In summary, the computational deep learning framework presented here illustrates how deep learning when combined with explainable AI tools, can reveal novel biological insights into the complex process of aging.

TimeFlies: an snRNA-seq aging clock for the fruit fly head sheds light on sex-biased aging

Although multiple high-performing epigenetic aging clocks exist, few are based directly on gene expression. Such transcriptomic aging clocks allow us to extract age-associated genes directly. However, most existing transcriptomic clocks model a subset of genes and are limited in their ability to predict novel biomarkers. With the growing popularity of single-cell sequencing, there is a need for robust single-cell transcriptomic aging clocks. Moreover, clocks have yet to be applied to investigate the elusive phenomenon of sex differences in aging. We introduce TimeFlies, a pan-cell-type scRNA-seq aging clock for the Drosophila melanogaster head. TimeFlies uses deep learning to classify the donor age of cells based on genome-wide gene expression profiles. Using explainability methods, we identified key marker genes contributing to the classification, with lncRNAs showing up as highly enriched among predicted biomarkers. The top biomarker gene across cell types is lncRNA:roX1, a regulator of X chromosome dosage compensation, a pathway previously identified as a top biomarker of aging in the mouse brain. We validated this finding experimentally, showing a decrease in survival probability in the absence of roX1 in vivo. Furthermore, we trained sex-specific TimeFlies clocks and noted significant differences in model predictions and explanations between male and female clocks, suggesting that different pathways drive aging in males and females.

Assessing Metabolic Ageing via DNA Methylation Surrogate Markers: A Multicohort Study in Britain, Ireland and the USA

Metabolomics and epigenomics have been used to develop 'ageing clocks' that assess biological age and identify 'accelerated ageing'. While metabolites are subject to short-term variation, DNA methylation (DNAm) may capture longer-term metabolic changes. We aimed to develop a hybrid DNAm-metabolic clock using DNAm as metabolite surrogates ('DNAm-metabolites') for age prediction. Within the UK Airwave cohort (n = 820), we developed DNAm metabolites by regressing 594 metabolites on DNAm and selected 177 DNAm metabolites and 193 metabolites to construct 'DNAm-metabolic' and 'metabolic' clocks. We evaluated clocks in their age prediction and association with noncommunicable disease risk factors. We additionally validated the DNAm-metabolic clock for the prediction of age and health outcomes in The Irish Longitudinal Study of Ageing (TILDA, n = 488) and the Health and Retirement Study (HRS, n = 4018). Around 70% of DNAm metabolites showed significant metabolite correlations (Pearson's r: > 0.30, p < 10-4) in the Airwave test set and overall stronger age associations than metabolites. The DNAm-metabolic clock was enriched for metabolic traits and was associated (p < 0.05) with male sex, heavy drinking, anxiety, depression and trauma. In TILDA and HRS, the DNAm-metabolic clock predicted age (r = 0.73 and 0.69), disability and gait speed (p < 0.05). In HRS, it additionally predicted time to death, diabetes, cardiovascular disease, frailty and grip strength. DNAm metabolite surrogates may facilitate metabolic studies using only DNAm data. Clocks built from DNAm metabolites provided a novel approach to assess metabolic ageing, potentially enabling early detection of metabolic-related diseases for personalised medicine.

Deep learning and generative artificial intelligence in aging research and healthy longevity medicine

With the global population aging at an unprecedented rate, there is a need to extend healthy productive life span. This review examines how Deep Learning (DL) and Generative Artificial Intelligence (GenAI) are used in biomarker discovery, deep aging clock development, geroprotector identification and generation of dual-purpose therapeutics targeting aging and disease. The paper explores the emergence of multimodal, multitasking research systems highlighting promising future directions for GenAI in human and animal aging research, as well as clinical application in healthy longevity medicine.

Histone mark age of human tissues and cell types

Aging is a complex and multifaceted process involving many epigenetic alterations. One key area of interest in aging research is the role of histone modifications, which can dynamically regulate gene expression. Here, we conducted a pan-tissue analysis of the dynamics of seven key histone modifications during human aging. Our histone-specific age prediction models showed surprisingly accurate performance, proving resilient to experimental and artificial noise. Simulation experiments for comparison with DNA methylation age predictors revealed competitive performance. Moreover, gene set enrichment analysis uncovered several critical developmental pathways for age prediction. Different from DNA methylation age predictors, genes known to be involved in aging biology are among the most important ones for the models. Last, we developed a pan-tissue pan-histone age predictor, suggesting that age-related epigenetic information is degenerated across the epigenome. This research highlights the power of histone marks as input for creating robust age predictors and opens avenues for understanding the role of epigenetic changes during aging.

Characterization of DNA methylation clock algorithms applied to diverse tissue types

BACKGROUND

DNA methylation (DNAm) data from human samples has been leveraged to develop "epigenetic clock" algorithms that predict age and other aging-related phenotypes. Some DNAm clocks were trained using DNAm obtained from blood cells, while other clocks were trained using data from diverse tissue/cell types. To assess how DNAm clocks perform across non-blood tissue types, we applied DNAm algorithms to DNAm data generated from 9 different human tissue types.

METHODS

We generated array-based DNAm measurements for 973 samples from deceased tissue donors from the GTEx (Genotype Tissue Expression) project representing nine distinct tissue types: lung, colon, prostate, ovary, breast, kidney, testis, skeletal muscle, and whole blood. For all samples, we generated DNAm clock estimates for 8 epigenetic clocks and characterized these tissue-specific clock estimates in terms of their distributions, correlations with chronological age, correlations of clock estimates between tissue types, and association with participant characteristics.

RESULTS

For each clock, the mean DNAm age estimate varied substantially across tissue types, and the mean values for the different clocks varied substantially within tissue types. For most clocks, the correlation with chronological age varied across tissue types, with blood often showing the strongest correlation. Each clock showed strong correlation across tissues, with some evidence of some residual correlation after adjusting for chronological age. In lung tissue, smoking generally had a positive association with epigenetic age.

CONCLUSIONS

This work demonstrates how differences in epigenetic aging among tissue types leads to clear differences in DNAm clock characteristics across tissue types. Tissue or cell-type specific epigenetic clocks are needed to optimize predictive performance of DNAm clocks in non-blood tissues and cell types.

Insights to aging prediction with AI based epigenetic clocks

Over the past century, human lifespan has increased remarkably, yet the inevitability of aging persists. The disparity between biological age, which reflects pathological deterioration and disease, and chronological age, indicative of normal aging, has driven prior research focused on identifying mechanisms that could inform interventions to reverse excessive age-related deterioration and reduce morbidity and mortality. DNA methylation has emerged as an important predictor of age, leading to the development of epigenetic clocks that quantify the extent of pathological deterioration beyond what is typically expected for a given age. Machine learning technologies offer promising avenues to enhance our understanding of the biological mechanisms governing aging by further elucidating the gap between biological and chronological ages. This perspective article examines current algorithmic approaches to epigenetic clocks, explores the use of machine learning for age estimation from DNA methylation, and discusses how refining the interpretation of ML methods and tailoring their inferences for specific patient populations and cell types can amplify the utility of these technologies in age prediction. By harnessing insights from machine learning, we are well-positioned to effectively adapt, customize and personalize interventions aimed at aging.

Epigenetic Clock: DNA Methylation as a Marker of Biological Age and Age-Associated Diseases

Age is one of the key criteria of human health used in practical medicine to predict the risk of common chronic diseases. However, biological age, which reflects the state of an individual organism, functional capabilities, social well-being, and risk of premature death from various causes, often does not coincide with chronological age. To determine biological age of a particular individuals and the rate of their aging, specific panels of DNA methylation markers called "epigenetic clock" (EC) were proposed. This review summarizes the data about the main types of ECs developed to date and their key characteristics. We described the results of works studying individual aging rates in common age-associated diseases and outlined main directions, development of which could expand application of ECs in fundamental and practical medicine. There is no doubt that revealing complex mechanisms underlying interaction between the rate of epigenetic aging and the risk of age-associated diseases could play a key role for prediction and early diagnosis, as well as for the development of preventive measures that could delay onset of the disease.

A quantitative benchmark of neural network feature selection methods for detecting nonlinear signals

Classification and regression problems can be challenging when the relevant input features are diluted in noisy datasets, in particular when the sample size is limited. Traditional Feature Selection (FS) methods address this issue by relying on some assumptions such as the linear or additive relationship between features. Recently, a proliferation of Deep Learning (DL) models has emerged to tackle both FS and prediction at the same time, allowing non-linear modeling of the selected features. In this study, we systematically assess the performance of DL-based feature selection methods on synthetic datasets of varying complexity, and benchmark their efficacy in uncovering non-linear relationships between features. We also use the same settings to benchmark the reliability of gradient-based feature attribution techniques for Neural Networks (NNs), such as Saliency Maps (SM). A quantitative evaluation of the reliability of these approaches is currently missing. Our analysis indicates that even simple synthetic datasets can significantly challenge most of the DL-based FS and SM methods, while Random Forests, TreeShap, mRMR and LassoNet are the best performing FS methods. Our conclusion is that when quantifying the relevance of a few non linearly-entangled predictive features diluted in a large number of irrelevant noisy variables, DL-based FS and SM interpretation methods are still far from being reliable.

Epigenetic Clocks: Beyond Biological Age, Using the Past to Predict the Present and Future

Predicting health trajectories and accurately measuring aging processes across the human lifespan remain profound scientific challenges. Assessing the effectiveness and impact of interventions targeting aging is even more elusive, largely due to the intricate, multidimensional nature of aging-a process that defies simple quantification. Traditional biomarkers offer only partial perspectives, capturing limited aspects of the aging landscape. Yet, over the past decade, groundbreaking advancements have emerged. Epigenetic clocks, derived from DNA methylation patterns, have established themselves as powerful aging biomarkers, capable of estimating biological age and assessing aging rates across diverse tissues with remarkable precision. These clocks provide predictive insights into mortality and age-related disease risks, effectively distinguishing biological age from chronological age and illuminating enduring questions in gerontology. Despite significant progress in epigenetic clock development, substantial challenges remain, underscoring the need for continued investigation to fully unlock their potential in the science of aging.

Breaking new ground on human health and well-being with epigenetic clocks: A systematic review and meta-analysis of epigenetic age acceleration associations

Epigenetic clocks provide an accurate molecular readout of epigenetic age and epigenetic age acceleration (EAA) derived from DNA methylation data have shown promise as biomarkers of ageing. This systematic review synthesised research on associations between EAA measures and various physiological, cognitive, social, and environmental factors. A comprehensive search strategy identified 299 publications reporting 1050 unique EAA-factor associations based on 53 methylation clocks. Random-effects meta-analyses pooled results across studies for selected EAA-factor pairs. Significant pooled associations emerged, providing insights into relationships between specific factors and accelerated epigenetic ageing. We developed a novel four-level classification system to categorise this diverse range of factors and enable a structured synthesis. To aid further research planning in this rapidly evolving field, TEAPEE (Tracker of EAA Associations with Phenotype & Environmental Exposure) - an interactive, searchable web table detailing all EAA-factor associations - was developed, cataloguing the epigenetic clocks, associated factors, classification categories, and direct links to the original studies. This resource will empower future investigations into the multifaceted determinants of epigenetic ageing, contributing to a deeper understanding of the epigenome's sensitivity to various life experiences and exposures.

Potential Predictor of Epigenetic Age Acceleration in Men: 2D:4D Finger Pattern

OBJECTIVES

Second to fourth digit ratio is widely known indicator of prenatal sex hormones proportion. Higher prenatal androgenization results in longer fourth finger and lower 2D:4D index. The aim of this study was to determine whether the 2D:4D digit ratio is associated with DNA methylation (DNAm) age dependently on sex.

MATERIAL AND METHODS

The study included 182 adults (106 females and 76 males) with a mean age of 51.5 ± 13 years. The investigation consisted of three main parts: a survey, anthropometric dimensions measurements (fingers length) and methylome analysis using collected blood samples. Genome-wide methylation was analyzed using EPIC microarray technology. Epigenetic age and epigenetic age acceleration were calculated using several widely applied algorithms.

RESULTS

Males with the female left hand pattern had more accelerated epigenetic age than those with the male pattern as calculated with PhenoAge and DNAmTL clocks.

CONCLUSIONS

Finger female pattern 2D:4D above or equal to 1 in males is associated with epigenetic age acceleration, indicating that prenatal exposure to estrogens in males may be related to aging process in the later ontogenesis.

Profiling the transcriptomic age of single-cells in humans

Although aging clocks predicting the age of individual organisms have been extensively studied, the age of individual cells remained largely unexplored. Most recently single-cell omics clocks were developed for the mouse, however, extensive profiling the age of human cells is still lacking. To fill this gap, here we use available scRNA-seq data of 1,058,909 blood cells of 508 healthy, human donors (between 19 and 75 years), for developing single-cell transcriptomic clocks and predicting the age of human blood cells. By the application of the proposed cell-type-specific single-cell clocks, our main observations are that (i) transcriptomic age is associated with cellular senescence; (ii) the transcriptomic age of classical monocytes as well as naive B and T cells is decreased in moderate COVID-19 followed by an increase for some cell types in severe COVID-19; and (iii) the human embryo cells transcriptomically rejuvenated at the morulae and blastocyst stages. In summary, here we demonstrate that single-cell transcriptomic clocks are useful tools to investigate aging and rejuvenation at the single-cell level.

Probabilistic inference of epigenetic age acceleration from cellular dynamics

The emergence of epigenetic predictors was a pivotal moment in geroscience, propelling the measurement and concept of biological aging into a quantitative era; however, while current epigenetic clocks show strong predictive power, they are data-driven in nature and are not based on the underlying biological mechanisms driving methylation dynamics. We show that predictions of these clocks are susceptible to several confounding non-age-related phenomena that make interpretation of these estimates and associations difficult. To address these limitations, we developed a probabilistic model describing methylation transitions at the cellular level. Our approach reveals two measurable components, acceleration and bias, which directly reflect perturbations of the underlying cellular dynamics. Acceleration is the proportional increase in the speed of methylation transitions across CpG sites, whereas bias corresponds to global changes in methylation levels. Using data from 15,900 participants from the Generation Scotland study, we develop a robust inference framework and show that these are two distinct processes confounding current epigenetic predictors. Our results show improved associations of acceleration and bias with physiological traits known to impact healthy aging, such as smoking and alcohol consumption, respectively. Furthermore, a genome-wide association study of epigenetic age acceleration identified seven genomic loci.

Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping

Unsupervised learning, particularly clustering, plays a pivotal role in disease subtyping and patient stratification, especially with the abundance of large-scale multi-omics data. Deep learning models, such as variational autoencoders (VAEs), can enhance clustering algorithms by leveraging inter-individual heterogeneity. However, the impact of confounders-external factors unrelated to the condition, e.g. batch effect or age-on clustering is often overlooked, introducing bias and spurious biological conclusions. In this work, we introduce four novel VAE-based deconfounding frameworks tailored for clustering multi-omics data. These frameworks effectively mitigate confounding effects while preserving genuine biological patterns. The deconfounding strategies employed include (i) removal of latent features correlated with confounders, (ii) a conditional VAE, (iii) adversarial training, and (iv) adding a regularization term to the loss function. Using real-life multi-omics data from The Cancer Genome Atlas, we simulated various confounding effects (linear, nonlinear, categorical, mixed) and assessed model performance across 50 repetitions based on reconstruction error, clustering stability, and deconfounding efficacy. Our results demonstrate that our novel models, particularly the conditional multi-omics VAE (cXVAE), successfully handle simulated confounding effects and recover biologically driven clustering structures. cXVAE accurately identifies patient labels and unveils meaningful pathological associations among cancer types, validating deconfounded representations. Furthermore, our study suggests that some of the proposed strategies, such as adversarial training, prove insufficient in confounder removal. In summary, our study contributes by proposing innovative frameworks for simultaneous multi-omics data integration, dimensionality reduction, and deconfounding in clustering. Benchmarking on open-access data offers guidance to end-users, facilitating meaningful patient stratification for optimized precision medicine.

A systematic review of phenotypic and epigenetic clocks used for aging and mortality quantification in humans

Aging is the leading driver of disease in humans and has profound impacts on mortality. Biological clocks are used to measure the aging process in the hopes of identifying possible interventions. Biological clocks may be categorized as phenotypic or epigenetic, where phenotypic clocks use easily measurable clinical biomarkers and epigenetic clocks use cellular methylation data. In recent years, methylation clocks have attained phenomenal performance when predicting chronological age and have been linked to various age-related diseases. Additionally, phenotypic clocks have been proven to be able to predict mortality better than chronological age, providing intracellular insights into the aging process. This review aimed to systematically survey all proposed epigenetic and phenotypic clocks to date, excluding mitotic clocks (i.e., cancer risk clocks) and those that were modeled using non-human samples. We reported the predictive performance of 33 clocks and outlined the statistical or machine learning techniques used. We also reported the most influential clinical measurements used in the included phenotypic clocks. Our findings provide a systematic reporting of the last decade of biological clock research and indicate possible avenues for future research.

DNA methylation correlates of chronological age in diverse human tissue types

BACKGROUND

While the association of chronological age with DNA methylation (DNAm) in whole blood has been extensively studied, the tissue-specificity of age-related DNAm changes remains an active area of research. Studies investigating the association of age with DNAm in tissues such as brain, skin, immune cells, fat, and liver have identified tissue-specific and non-specific effects, thus, motivating additional studies of diverse human tissue and cell types.

RESULTS

Here, we performed an epigenome-wide association study, leveraging DNAm data (Illumina EPIC array) from 961 tissue samples representing 9 tissue types (breast, lung, colon, ovary, prostate, skeletal muscle, testis, whole blood, and kidney) from the Genotype-Tissue Expression (GTEx) project. We identified age-associated CpG sites (false discovery rate < 0.05) in 8 tissues (all except skeletal muscle, n = 47). This included 162,002 unique hypermethylated and 90,626 hypomethylated CpG sites across all tissue types, with 130,137 (80%) hypermethylated CpGs and 74,703 (82%) hypomethylated CpG sites observed in a single tissue type. While the majority of age-associated CpG sites appeared tissue-specific, the patterns of enrichment among genomic features, such as chromatin states and CpG islands, were similar across most tissues, suggesting common mechanisms underlying cellular aging. Consistent with previous findings, we observed that hypermethylated CpG sites are enriched in regions with repressed polycomb signatures and CpG islands, while hypomethylated CpG sites preferentially occurred in non-CpG islands and enhancers. To gain insights into the functional effects of age-related DNAm changes, we assessed the correlation between DNAm and local gene expression changes to identify age-related expression quantitative trait methylation (age-eQTMs). We identified several age-eQTMs present in multiple tissue-types, including in the CDKN2A, HENMT1, and VCWE regions.

CONCLUSION

Overall, our findings will aid future efforts to develop biomarkers of aging and understand mechanisms of aging in diverse human tissue types.

B Cells Isolated from Individuals Who Do Not Respond to the HBV Vaccine Are Characterized by Higher DNA Methylation-Estimated Aging Compared to Responders

Healthcare workers (HCWs) are a high-risk group for hepatitis B virus (HBV) infection. Notably, about 5-10% of the general population does not respond to the HBV vaccination. In this study, we aimed to investigate DNA methylation (DNAm) in order to estimate the biological age of B cells from HCW of both sexes, either responder (R) or non-responder (NR), to HBV vaccination. We used genome-wide DNA methylation data to calculate a set of biomarkers in B cells collected from 41 Rs and 30 NRs between 22 and 62 years old. Unresponsiveness to HBV vaccination was associated with accelerated epigenetic aging (DNAmAge, AltumAge, DunedinPoAm) and was accompanied by epigenetic drift. Female non-responders had higher estimates of telomere length and lower CRP inflammation risk score when compared to responders. Overall, epigenetic differences between responders and non-responders were more evident in females than males. In this study we demonstrated that several methylation DNAm-based clocks and biomarkers are associated with an increased risk of non-response to HBV vaccination, particularly in females. Based on these results, we propose that accelerated epigenetic age could contribute to vaccine unresponsiveness. These insights may help improve the evaluation of the effectiveness of vaccination strategies, especially among HCWs and vulnerable patients.

Determining the relationship of p16INK4a and additional molecular markers of aging with clinical frailty in hematologic malignancy

PURPOSE

Older adults with hematologic malignancies (HM) have unique challenges due to age and fitness. The primary aim of this pilot study was to benchmark the ability of multiple biomarkers of aging (p16, epigenetic clocks, T cell gene expression profiles, and T cell receptor excision circles (TREC) to identify frailty as measured by a clinical impairment index (I2) in patients with HM.

METHODS

70 patients newly diagnosed with HM had peripheral blood T lymphocytes (PBTL) analyzed for p16INK4a expression using the OSU_Senescence Nanostring CodeSet. PBTL epigenetic age was measured using 7 epigenetic clocks, and TREC were quantified by qRT-PCR. A composite clinical impairment index (I2) was generated by combining values from 11 geriatric metrics (Independent Activities of Daily Living (iADL), physical health score, Short Physical Performance Battery (SPPB), Body Mass Index (BMI), Eastern Cooperative Oncology Group (ECOG) performance status, self-reported KPS, Blessed Orientation Memory Concentration (BOMC), polypharmacy, Mental Health Inventory (MHI)-17, Medical Outcomes Study (MOS) subscales). Clinical frailty was defined as a score of 7 or greater on the I2.

RESULTS

Age-adjusted p16INK4a was similar in newly diagnosed patients and healthy controls (p > 0.1). PBTL p16INK4a levels correlated positively with the Hannum [r = 0.35, 95% CI (0.09-0.75); p adj. = 0.04] and PhenoAge [r = 0.37, 95% CI (0.11-0.59); p adj. = 0.04] epigenetic clocks. The discrimination ability of the I2 model was calculated using the area under the receiver operating characteristic curve (AUC). After adjusting for chronologic age and disease group, baseline p16INK4a [AUC = 0.76, 95% CI (0.56-0.98); p = 0.01], Hannum [AUC = 0.70, 95% CI (0.54-0.85); p = 0.01], PhenoAge [AUC = 0.71, 95% CI (0.55-0.86); p = 0.01], and DunedinPACE [AUC = 0.73, 95% CI (0.57-0.88); p =  < 0.01] measures showed the greatest potential to identify clinical frailty using the I2.

CONCLUSIONS

Our pilot data suggest that multiple blood-based aging biomarkers have potential to identify frailty in older adults with HM.

IMPLICATIONS FOR CANCER SURVIVORS

We developed the I2 index to quantify impairments across geriatric domains and discovered that PBTL p16, Hannum, PhenoAge, and DunedinPACE are promising indicators of frailty in HM.

Designing interpretable deep learning applications for functional genomics: a quantitative analysis

Deep learning applications have had a profound impact on many scientific fields, including functional genomics. Deep learning models can learn complex interactions between and within omics data; however, interpreting and explaining these models can be challenging. Interpretability is essential not only to help progress our understanding of the biological mechanisms underlying traits and diseases but also for establishing trust in these model's efficacy for healthcare applications. Recognizing this importance, recent years have seen the development of numerous diverse interpretability strategies, making it increasingly difficult to navigate the field. In this review, we present a quantitative analysis of the challenges arising when designing interpretable deep learning solutions in functional genomics. We explore design choices related to the characteristics of genomics data, the neural network architectures applied, and strategies for interpretation. By quantifying the current state of the field with a predefined set of criteria, we find the most frequent solutions, highlight exceptional examples, and identify unexplored opportunities for developing interpretable deep learning models in genomics.

Towards a Novel Frontier in the Use of Epigenetic Clocks in Epidemiology

Health problems associated with aging are a major public health concern for the future. Aging is a complex process with wide intervariability among individuals. Therefore, there is a need for innovative public health strategies that target factors associated with aging and the development of tools to assess the effectiveness of these strategies accurately. Novel approaches to measure biological age, such as epigenetic clocks, have become relevant. These clocks use non-sequential variable information from the genome and employ mathematical algorithms to estimate biological age based on DNA methylation levels. Therefore, in the present study, we comprehensively review the current status of the epigenetic clocks and their associations across the human phenome. We emphasize the potential utility of these tools in an epidemiological context, particularly in evaluating the impact of public health interventions focused on promoting healthy aging. Our review describes associations between epigenetic clocks and multiple traits across the life and health span. Additionally, we highlighted the evolution of studies beyond mere associations to establish causal mechanisms between epigenetic age and disease. We explored the application of epigenetic clocks to measure the efficacy of interventions focusing on rejuvenation.

A miRNA-based epigenetic molecular clock for biological skin-age prediction

Skin aging is one of the visible characteristics of the aging process in humans. In recent years, different biological clocks have been generated based on protein or epigenetic markers, but few have focused on biological age in the skin. Arrest the aging process or even being able to restore an organism from an older to a younger stage is one of the main challenges in the last 20 years in biomedical research. We have implemented several machine learning models, including regression and classification algorithms, in order to create an epigenetic molecular clock based on miRNA expression profiles of healthy subjects to predict biological age-related to skin. Our best models are capable of classifying skin samples according to age groups (18-28; 29-39; 40-50; 51-60 or 61-83 years old) with an accuracy of 80% or predict age with a mean absolute error of 10.89 years using the expression levels of 1856 unique miRNAs. Our results suggest that this kind of epigenetic clocks arises as a promising tool with several applications in the pharmaco-cosmetic industry.

A multi-organization epigenetic age prediction based on a channel attention perceptron networks

DNA methylation indicates the individual's aging, so-called Epigenetic clocks, which will improve the research and diagnosis of aging diseases by investigating the correlation between methylation loci and human aging. Although this discovery has inspired many researchers to develop traditional computational methods to quantify the correlation and predict the chronological age, the performance bottleneck delayed access to the practical application. Since artificial intelligence technology brought great opportunities in research, we proposed a perceptron model integrating a channel attention mechanism named PerSEClock. The model was trained on 24,516 CpG loci that can utilize the samples from all types of methylation identification platforms and tested on 15 independent datasets against seven methylation-based age prediction methods. PerSEClock demonstrated the ability to assign varying weights to different CpG loci. This feature allows the model to enhance the weight of age-related loci while reducing the weight of irrelevant loci. The method is free to use for academics at www.dnamclock.com/#/original.

Application of AI in biological age prediction

The development of anti-aging interventions requires quantitative measurement of biological age. Machine learning models, known as "aging clocks," are built by leveraging diverse aging biomarkers that vary across lifespan to predict biological age. In addition to traditional aging clocks harnessing epigenetic signatures derived from bulk samples, emerging technologies allow the biological age estimating at single-cell level to dissect cellular diversity in aging tissues. Moreover, imaging-based aging clocks are increasingly employed with the advantage of non-invasive measurement, making it suitable for large-scale human cohort studies. To fully capture the features in the ever-growing multi-modal and high-dimensional aging-related data and uncover disease associations, deep-learning based approaches, which are effective to learn complex and non-linear relationships without relying on pre-defined features, are increasingly applied. The use of big data and AI-based aging clocks has achieved high accuracy, interpretability and generalizability, guiding clinical applications to delay age-related diseases and extend healthy lifespans.

pyaging: a Python-based compendium of GPU-optimized aging clocks

MOTIVATION

Aging is intricately linked to diseases and mortality. It is reflected in molecular changes across various tissues which can be leveraged for the development of biomarkers of aging using machine learning models, known as aging clocks. Despite advancements in the field, a significant challenge remains: the lack of robust, Python-based software tools for integrating and comparing these diverse models. This gap highlights the need for comprehensive solutions that can handle the complexity and variety of data in aging research.

RESULTS

To address this gap, I introduce pyaging, a comprehensive open-source Python package designed to facilitate aging research. pyaging harmonizes dozens of aging clocks, covering a range of molecular data types such as DNA methylation, transcriptomics, histone mark ChIP-Seq, and ATAC-Seq. The package is not limited to traditional model types; it features a diverse array, from linear and principal component models to neural networks and automatic relevance determination models. Thanks to a PyTorch-based backend that enables GPU acceleration, pyaging is capable of rapid inference, even when dealing with large datasets and complex models. In addition, the package's support for multi-species analysis extends its utility across various organisms, including humans, various mammals, and Caenorhabditis elegans.

AVAILABILITY AND IMPLEMENTATION

pyaging is accessible on GitHub, at https://github.com/rsinghlab/pyaging, and the distribution is available on PyPi, at https://pypi.org/project/pyaging/. The software is also archived on Zenodo, at https://zenodo.org/doi/10.5281/zenodo.10335011.

Predicting human chronological age via AI analysis of dorsal hand versus facial images: A study in a cohort of Indian females

Predicting a person's chronological age (CA) from visible skin features using artificial intelligence (AI) is now commonplace. Often, convolutional neural network (CNN) models are built using images of the face as biometric data. However, hands hold telltale signs of a person's age. To determine the utility of using only hand images in predicting CA, we developed two deep CNNs based on 1) dorsal hand images (H) and 2) frontal face images (F). Subjects (n = 1454) were Indian women, 20-80 years, across three geographic cohorts (Mumbai, New Delhi and Bangalore) and having a broad variation in skin tones. Images were randomised: 70% of F and 70% of H were used to train CNNs. The remaining 30% of F and H were retained for validation. CNN validation showed mean absolute error for predicting CA using F and H of 4.1 and 4.7 years, respectively. In both cases correlations of predicted and actual age were statistically significant (r(F) = 0.93, r(H) = 0.90). The CNNs for F and H were validated for dark and light skin tones. Finally, by blurring or accentuating visible features on specific regions of the hand and face, we identified those features that contributed to the CNN models. For the face, areas of the inner eye corner and around the mouth were most important for age prediction. For the hands, knuckle texture was a key driver for age prediction. Collectively, for AI estimates of CA, CNNs based solely on hand images are a viable alternative and comparable to CNNs based on facial images.

Biologically informed deep learning for explainable epigenetic clocks

Ageing is often characterised by progressive accumulation of damage, and it is one of the most important risk factors for chronic disease development. Epigenetic mechanisms including DNA methylation could functionally contribute to organismal aging, however the key functions and biological processes may govern ageing are still not understood. Although age predictors called epigenetic clocks can accurately estimate the biological age of an individual based on cellular DNA methylation, their models have limited ability to explain the prediction algorithm behind and underlying key biological processes controlling ageing. Here we present XAI-AGE, a biologically informed, explainable deep neural network model for accurate biological age prediction across multiple tissue types. We show that XAI-AGE outperforms the first-generation age predictors and achieves similar results to deep learning-based models, while opening up the possibility to infer biologically meaningful insights of the activity of pathways and other abstract biological processes directly from the model.

Epigenetic regulation of drug metabolism in aging: utilizing epigenetics to optimize geriatric pharmacotherapy

We explore the relationship between epigenetic aging and drug metabolism. We review current evidence for changes in drug metabolism in normal aging, followed by a description of how epigenetic modifications associated with age can regulate the expression and functionality of genes. In particular, we focus on the role of epigenome-wide studies of human and mouse liver in understanding these age-related processes with respect to xenobiotic processing. We highlight genes encoding drug metabolizing enzymes and transporters revealed to be affected by epigenetic aging in these studies. We conclude that substantial evidence exists for epigenetic aging impacting drug metabolism and transport genes, but more work is needed. We further highlight the promise of pharmacoepigenetics applied to enhancing drug safety in older adults.

eXplainable Artificial Intelligence (XAI) in aging clock models

XAI is a rapidly progressing field of machine learning, aiming to unravel the predictions of complex models. XAI is especially required in sensitive applications, e.g. in health care, when diagnosis, recommendations and treatment choices might rely on the decisions made by artificial intelligence systems. AI approaches have become widely used in aging research as well, in particular, in developing biological clock models and identifying biomarkers of aging and age-related diseases. However, the potential of XAI here awaits to be fully appreciated. We discuss the application of XAI for developing the "aging clocks" and present a comprehensive analysis of the literature categorized by the focus on particular physiological systems.

ResnetAge: A Resnet-Based DNA Methylation Age Prediction Method

Aging is a significant contributing factor to degenerative diseases such as cancer. The extent of DNA methylation in human cells indicates the aging process and screening for age-related methylation sites can be used to construct epigenetic clocks. Thereby, it can be a new aging-detecting marker for clinical diagnosis and treatments. Predicting the biological age of human individuals is conducive to the study of physical aging problems. Although many researchers have developed epigenetic clock prediction methods based on traditional machine learning and even deep learning, higher prediction accuracy is still required to match the clinical applications. Here, we proposed an epigenetic clock prediction method based on a Resnet neuro networks model named ResnetAge. The model accepts 22,278 CpG sites as a sample input, supporting both the Illumina 27K and 450K identification frameworks. It was trained using 32 public datasets containing multiple tissues such as whole blood, saliva, and mouth. The Mean Absolute Error (MAE) of the training set is 1.29 years, and the Median Absolute Deviation (MAD) is 0.98 years. The Mean Absolute Error (MAE) of the validation set is 3.24 years, and the Median Absolute Deviation (MAD) is 2.3 years. Our method has higher accuracy in age prediction in comparison with other methylation-based age prediction methods.

CellBiAge: Improved single-cell age classification using data binarization

Aging is a major risk factor for many diseases. Accurate methods for predicting age in specific cell types are essential to understand the heterogeneity of aging and to assess rejuvenation strategies. However, classifying organismal age at single-cell resolution using transcriptomics is challenging due to sparsity and noise. Here, we developed CellBiAge, a robust and easy-to-implement machine learning pipeline, to classify the age of single cells in the mouse brain using single-cell transcriptomics. We show that binarization of gene expression values for the top highly variable genes significantly improved test performance across different models, techniques, sexes, and brain regions, with potential age-related genes identified for model prediction. Additionally, we demonstrate CellBiAge's ability to capture exercise-induced rejuvenation in neural stem cells. This study provides a broadly applicable approach for robust classification of organismal age of single cells in the mouse brain, which may aid in understanding the aging process and evaluating rejuvenation methods.

Endometrial receptivity in women of advanced age: an underrated factor in infertility

BACKGROUND

Modern lifestyle has led to an increase in the age at conception. Advanced age is one of the critical risk factors for female-related infertility. It is well known that maternal age positively correlates with the deterioration of oocyte quality and chromosomal abnormalities in oocytes and embryos. The effect of age on endometrial function may be an equally important factor influencing implantation rate, pregnancy rate, and overall female fertility. However, there are only a few published studies on this topic, suggesting that this area has been under-explored. Improving our knowledge of endometrial aging from the biological (cellular, molecular, histological) and clinical perspectives would broaden our understanding of the risks of age-related female infertility.

OBJECTIVE AND RATIONALE

The objective of this narrative review is to critically evaluate the existing literature on endometrial aging with a focus on synthesizing the evidence for the impact of endometrial aging on conception and pregnancy success. This would provide insights into existing gaps in the clinical application of research findings and promote the development of treatment options in this field.

SEARCH METHODS

The review was prepared using PubMed (Medline) until February 2023 with the keywords such as 'endometrial aging', 'receptivity', 'decidualization', 'hormone', 'senescence', 'cellular', 'molecular', 'methylation', 'biological age', 'epigenetic', 'oocyte recipient', 'oocyte donation', 'embryo transfer', and 'pregnancy rate'. Articles in a language other than English were excluded.

OUTCOMES

In the aging endometrium, alterations occur at the molecular, cellular, and histological levels suggesting that aging has a negative effect on endometrial biology and may impair endometrial receptivity. Additionally, advanced age influences cellular senescence, which plays an important role during the initial phase of implantation and is a major obstacle in the development of suitable senolytic agents for endometrial aging. Aging is also accountable for chronic conditions associated with inflammaging, which eventually can lead to increased pro-inflammation and tissue fibrosis. Furthermore, advanced age influences epigenetic regulation in the endometrium, thus altering the relation between its epigenetic and chronological age. The studies in oocyte donation cycles to determine the effect of age on endometrial receptivity with respect to the rates of implantation, clinical pregnancy, miscarriage, and live birth have revealed contradictory inferences indicating the need for future research on the mechanisms and corresponding causal effects of women's age on endometrial receptivity.

WIDER IMPLICATIONS

Increasing age can be accountable for female infertility and IVF failures. Based on the complied observations and synthesized conclusions in this review, advanced age has been shown to have a negative impact on endometrial functioning. This information can provide recommendations for future research focusing on molecular mechanisms of age-related cellular senescence, cellular composition, and transcriptomic changes in relation to endometrial aging. Additionally, further prospective research is needed to explore newly emerging therapeutic options, such as the senolytic agents that can target endometrial aging without affecting decidualization. Moreover, clinical trial protocols, focusing on oocyte donation cycles, would be beneficial in understanding the direct clinical implications of endometrial aging on pregnancy outcomes.

Smoking affects epigenetic ageing of lung bronchoalveolar lavage cells in Multiple Sclerosis

BACKGROUND

A compelling body of evidence implicates cigarette smoking and lung inflammation in Multiple Sclerosis (MS) susceptibility and progression. Previous studies have reported epigenetic age (DNAm age) acceleration in blood immune cells and in glial cells of people with MS (pwMS) compared to healthy controls (HC).

OBJECTIVES

We aimed to examine biological ageing in lung immune cells in the context of MS and smoking.

METHODS

We analyzed age acceleration residuals in lung bronchoalveolar lavage (BAL) cells, constituted of mainly alveolar macrophages, from 17 pwMS and 22 HC in relation to smoking using eight DNA methylation-based clocks, namely AltumAge, Horvath, GrimAge, PhenoAge, Zhang, SkinBlood, Hannum, Monocyte clock as well as two RNA-based clocks, which capture different aspects of biological ageing.

RESULTS

After adjustment for covariates, five epigenetic clocks showed significant differences between the groups. Four of them, Horvath (Padj = 0.028), GrimAge (Padj = 4.28 × 10-7), SkinBlood (Padj = 0.001) and Zhang (Padj = 0.02), uncovered the sole effect of smoking on ageing estimates, irrespective of the clinical group. The Horvath, SkinBlood and Zhang clocks showed a negative impact of smoking while GrimAge detected smoking-associated age acceleration in BAL cells. On the contrary, the AltumAge clock revealed differences between pwMS and HC and indicated that, in the absence of smoking, BAL cells of pwMS were epigenetically 5.4 years older compared to HC (Padj = 0.028). Smoking further affected epigenetic ageing in BAL cells of pwMS specifically as non-smoking pwMS exhibited a 10.2-year AltumAge acceleration compared to pwMS smokers (Padj = 0.0049). Of note, blood-derived monocytes did not show any MS-specific or smoking-related AltumAge differences. The difference between BAL cells of pwMS smokers and non-smokers was attributable to the differential methylation of 114 AltumAge-CpGs (Padj < 0.05) affecting genes involved in innate immune processes such as cytokine production, defense response and cell motility. These changes functionally translated into transcriptional differences in BAL cells between pwMS smokers and non-smokers.

CONCLUSIONS

BAL cells of pwMS display inflammation-related and smoking-dependent changes associated to epigenetic ageing captured by the AltumAge clock. Future studies examining potential confounders, such as the distribution of distinct BAL myeloid cell types in pwMS compared to control individuals in relation to smoking may clarify the varying performance and DNAm age estimations among epigenetic clocks.

Reappraisal of the Concept of Accelerated Aging in Neurodegeneration and Beyond

BACKGROUND

Genetic and epigenetic changes, oxidative stress and inflammation influence the rate of aging, which diseases, lifestyle and environmental factors can further accelerate. In accelerated aging (AA), the biological age exceeds the chronological age.

OBJECTIVE

The objective of this study is to reappraise the AA concept critically, considering its weaknesses and limitations.

METHODS

We reviewed more than 300 recent articles dealing with the physiology of brain aging and neurodegeneration pathophysiology.

RESULTS

(1) Application of the AA concept to individual organs outside the brain is challenging as organs of different systems age at different rates. (2) There is a need to consider the deceleration of aging due to the potential use of the individual structure-functional reserves. The latter can be restored by pharmacological and/or cognitive therapy, environment, etc. (3) The AA concept lacks both standardised terminology and methodology. (4) Changes in specific molecular biomarkers (MBM) reflect aging-related processes; however, numerous MBM candidates should be validated to consolidate the AA theory. (5) The exact nature of many potential causal factors, biological outcomes and interactions between the former and the latter remain largely unclear.

CONCLUSIONS

Although AA is commonly recognised as a perspective theory, it still suffers from a number of gaps and limitations that assume the necessity for an updated AA concept.

Yearning for machine learning: applications for the classification and characterisation of senescence

Senescence is a widely appreciated tumour suppressive mechanism, which acts as a barrier to cancer development by arresting cell cycle progression in response to harmful stimuli. However, senescent cell accumulation becomes deleterious in aging and contributes to a wide range of age-related pathologies. Furthermore, senescence has beneficial roles and is associated with a growing list of normal physiological processes including wound healing and embryonic development. Therefore, the biological role of senescent cells has become increasingly nuanced and complex. The emergence of sophisticated, next-generation profiling technologies, such as single-cell RNA sequencing, has accelerated our understanding of the heterogeneity of senescence, with distinct final cell states emerging within models as well as between cell types and tissues. In order to explore data sets of increasing size and complexity, the senescence field has begun to employ machine learning (ML) methodologies to probe these intricacies. Most notably, ML has been used to aid the classification of cells as senescent, as well as to characterise the final senescence phenotypes. Here, we provide a background to the principles of ML tasks, as well as some of the most commonly used methodologies from both traditional and deep ML. We focus on the application of these within the context of senescence research, by addressing the utility of ML for the analysis of data from different laboratory technologies (microscopy, transcriptomics, proteomics, methylomics), as well as the potential within senolytic drug discovery. Together, we aim to highlight both the progress and potential for the application of ML within senescence research.

Accurate age prediction from blood using a small set of DNA methylation sites and a cohort-based machine learning algorithm

Chronological age prediction from DNA methylation sheds light on human aging, health, and lifespan. Current clocks are mostly based on linear models and rely upon hundreds of sites across the genome. Here, we present GP-age, an epigenetic non-linear cohort-based clock for blood, based upon 11,910 methylomes. Using 30 CpG sites alone, GP-age outperforms state-of-the-art models, with a median accuracy of ∼2 years on held-out blood samples, for both array and sequencing-based data. We show that aging-related changes occur at multiple neighboring CpGs, with implications for using fragment-level analysis of sequencing data in aging research. By training three independent clocks, we show enrichment of donors with consistent deviation between predicted and actual age, suggesting individual rates of biological aging. Overall, we provide a compact yet accurate alternative to array-based clocks for blood, with applications in longitudinal aging research, forensic profiling, and monitoring epigenetic processes in transplantation medicine and cancer.

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery

Aging is a complex and multifactorial process that increases the risk of various age-related diseases and there are many aging clocks that can accurately predict chronological age, mortality, and health status. These clocks are disconnected and are rarely fit for therapeutic target discovery. In this study, we propose a novel approach to multimodal aging clock we call Precious1GPT utilizing methylation and transcriptomic data for interpretable age prediction and target discovery developed using a transformer-based model and transfer learning for case-control classification. While the accuracy of the multimodal transformer is lower within each individual data type compared to the state of art specialized aging clocks based on methylation or transcriptomic data separately it may have higher practical utility for target discovery. This method provides the ability to discover novel therapeutic targets that hypothetically may be able to reverse or accelerate biological age providing a pathway for therapeutic drug discovery and validation using the aging clock. In addition, we provide a list of promising targets annotated using the PandaOmics industrial target discovery platform.

A comparison of feature selection methodologies and learning algorithms in the development of a DNA methylation-based telomere length estimator

BACKGROUND

The field of epigenomics holds great promise in understanding and treating disease with advances in machine learning (ML) and artificial intelligence being vitally important in this pursuit. Increasingly, research now utilises DNA methylation measures at cytosine-guanine dinucleotides (CpG) to detect disease and estimate biological traits such as aging. Given the challenge of high dimensionality of DNA methylation data, feature-selection techniques are commonly employed to reduce dimensionality and identify the most important subset of features. In this study, our aim was to test and compare a range of feature-selection methods and ML algorithms in the development of a novel DNA methylation-based telomere length (TL) estimator. We utilised both nested cross-validation and two independent test sets for the comparisons.

RESULTS

We found that principal component analysis in advance of elastic net regression led to the overall best performing estimator when evaluated using a nested cross-validation analysis and two independent test cohorts. This approach achieved a correlation between estimated and actual TL of 0.295 (83.4% CI [0.201, 0.384]) on the EXTEND test data set. Contrastingly, the baseline model of elastic net regression with no prior feature reduction stage performed less well in general-suggesting a prior feature-selection stage may have important utility. A previously developed TL estimator, DNAmTL, achieved a correlation of 0.216 (83.4% CI [0.118, 0.310]) on the EXTEND data. Additionally, we observed that different DNA methylation-based TL estimators, which have few common CpGs, are associated with many of the same biological entities.

CONCLUSIONS

The variance in performance across tested approaches shows that estimators are sensitive to data set heterogeneity and the development of an optimal DNA methylation-based estimator should benefit from the robust methodological approach used in this study. Moreover, our methodology which utilises a range of feature-selection approaches and ML algorithms could be applied to other biological markers and disease phenotypes, to examine their relationship with DNA methylation and predictive value.

Decoding the role of transcriptomic clocks in the human prefrontal cortex

Aging is a complex process with interindividual variability, which can be measured by aging biological clocks. Aging clocks are machine-learning algorithms guided by biological information and associated with mortality risk and a wide range of health outcomes. One of these aging clocks are transcriptomic clocks, which uses gene expression data to predict biological age; however, their functional role is unknown. Here, we profiled two transcriptomic clocks (RNAAgeCalc and knowledge-based deep neural network clock) in a large dataset of human postmortem prefrontal cortex (PFC) samples. We identified that deep-learning transcriptomic clock outperforms RNAAgeCalc to predict transcriptomic age in the human PFC. We identified associations of transcriptomic clocks with psychiatric-related traits. Further, we applied system biology algorithms to identify common gene networks among both clocks and performed pathways enrichment analyses to assess its functionality and prioritize genes involved in the aging processes. Identified gene networks showed enrichment for diseases of signal transduction by growth factor receptors and second messenger pathways. We also observed enrichment of genome-wide signals of mental and physical health outcomes and identified genes previously associated with human brain aging. Our findings suggest a link between transcriptomic aging and health disorders, including psychiatric traits. Further, it reveals functional genes within the human PFC that may play an important role in aging and health risk.

Examining the biological mechanisms of human mental disorders resulting from gene-environment interdependence using novel functional genomic approaches

We explore how functional genomics approaches that integrate datasets from human and non-human model systems can improve our understanding of the effect of gene-environment interplay on the risk for mental disorders. We start by briefly defining the G-E paradigm and its challenges and then discuss the different levels of regulation of gene expression and the corresponding data existing in humans (genome wide genotyping, transcriptomics, DNA methylation, chromatin modifications, chromosome conformational changes, non-coding RNAs, proteomics and metabolomics), discussing novel approaches to the application of these data in the study of the origins of mental health. Finally, we discuss the multilevel integration of diverse types of data. Advance in the use of functional genomics in the context of a G-E perspective improves the detection of vulnerabilities, informing the development of preventive and therapeutic interventions.

Refining epigenetic prediction of chronological and biological age

BACKGROUND

Epigenetic clocks can track both chronological age (cAge) and biological age (bAge). The latter is typically defined by physiological biomarkers and risk of adverse health outcomes, including all-cause mortality. As cohort sample sizes increase, estimates of cAge and bAge become more precise. Here, we aim to develop accurate epigenetic predictors of cAge and bAge, whilst improving our understanding of their epigenomic architecture.

METHODS

First, we perform large-scale (N = 18,413) epigenome-wide association studies (EWAS) of chronological age and all-cause mortality. Next, to create a cAge predictor, we use methylation data from 24,674 participants from the Generation Scotland study, the Lothian Birth Cohorts (LBC) of 1921 and 1936, and 8 other cohorts with publicly available data. In addition, we train a predictor of time to all-cause mortality as a proxy for bAge using the Generation Scotland cohort (1214 observed deaths). For this purpose, we use epigenetic surrogates (EpiScores) for 109 plasma proteins and the 8 component parts of GrimAge, one of the current best epigenetic predictors of survival. We test this bAge predictor in four external cohorts (LBC1921, LBC1936, the Framingham Heart Study and the Women's Health Initiative study).

RESULTS

Through the inclusion of linear and non-linear age-CpG associations from the EWAS, feature pre-selection in advance of elastic net regression, and a leave-one-cohort-out (LOCO) cross-validation framework, we obtain cAge prediction with a median absolute error equal to 2.3 years. Our bAge predictor was found to slightly outperform GrimAge in terms of the strength of its association to survival (HRGrimAge = 1.47 [1.40, 1.54] with p = 1.08 × 10-52, and HRbAge = 1.52 [1.44, 1.59] with p = 2.20 × 10-60). Finally, we introduce MethylBrowsR, an online tool to visualise epigenome-wide CpG-age associations.

CONCLUSIONS

The integration of multiple large datasets, EpiScores, non-linear DNAm effects, and new approaches to feature selection has facilitated improvements to the blood-based epigenetic prediction of biological and chronological age.

Comparative epigenomics by machine learning approach for neuroblastoma

BACKGROUND

Neuroblastoma (NB) is the second most common pediatric solid tumor. Because the number of genetic mutations found in tumors are small, even in some patients with unfavorable NB, epigenetic variation is expected to play an important role in NB progression. DNA methylation is a major epigenetic mechanism, and its relationship with NB prognosis has been a concern. One limitation with the analysis of variation in DNA methylation is the lack of a suitable analytical model. Therefore, in this study, we performed a random forest (RF) analysis of the DNA methylome data of NB from multiple databases.

RESULTS

RF is a popular machine learning model owing to its simplicity, intuitiveness, and computational cost. RF analysis identified novel intermediate-risk patient groups with characteristic DNA methylation patterns within the low-risk group. Feature selection analysis based on probe annotation revealed that enhancer-annotated regions had strong predictive power, particularly for MYCN-amplified NBs. We developed a gene-based analytical model to identify candidate genes related to disease progression, such as PRDM8 and FAM13A-AS1. RF analysis revealed sufficient predictive power compared to other machine learning models.

CONCLUSIONS

RF is a useful tool for DNA methylome analysis in cancer epigenetic studies, and has potential to identify a novel cancer-related genes.

Epigenetic clocks provide clues to the mystery of uterine ageing

BACKGROUND

Rising maternal ages and age-related fertility decline are a global challenge for modern reproductive medicine. Clinicians and researchers pay specific attention to ovarian ageing and hormonal insufficiency in this regard. However, uterine ageing is often left out of the picture, with the majority of reproductive clinicians being close to unanimous on the absence of age-related functional decline in the uterine tissues. Therefore, most existing techniques to treat an age-related decline in implantation rates are based primarily on hormonal supplementation and oocyte donation. Solving the issue of uterine ageing might lead to an adjustment to these methods.

OBJECTIVE AND RATIONALE

A focus on uterine ageing and the possibility of slowing it emerged with the development of the information theory of ageing, which identifies genomic instability and erosion of the epigenetic landscape as important drivers of age-related decline in the functionality of most cells and tissues. Age-related smoothing of this landscape and a decline in tissue function can be assessed by measuring the ticking of epigenetic clocks. Within this review, we explore whether the uterus experiences age-related alterations using this elegant approach. We analyse existing data on epigenetic clocks in the endometrium, highlight approaches to improve the accuracy of the clocks in this cycling tissue, speculate on the endometrial pathologies whose progression might be predicted by the altered speed of epigenetic clocks and discuss the possibilities of slowing down the ticking of these clocks.

SEARCH METHODS

Data for this review were identified by searches of Medline, PubMed and Google Scholar. References from relevant articles using the search terms 'ageing', 'maternal age', 'female reproduction', 'uterus', 'endometrium', 'implantation', 'decidualization', 'epigenetic clock', 'biological age', 'DNA methylation', 'fertility' and 'infertility' were selected. A total of 95 articles published in English between 1985 and 2022 were included, six of which describe the use of the epigenetic clock to evaluate uterine/endometrium ageing.

OUTCOMES

Application of the Horvath and DNAm PhenoAge epigenetic clocks demonstrated a poor correlation with chronological age in the endometrium. Several approaches were suggested to enhance the predictive power of epigenetic clocks for the endometrium. The first was to increase the number of samples in the training dataset, as for the Zang clock, or to use more sophisticated clock-building algorithms, as for the AltumAge clock. The second method is to adjust the clocks according to the dynamic nature of the endometrium. Using either approach revealed a strong correlation with chronological age in the endometrium, providing solid evidence for age-related functional decline in this tissue. Furthermore, age acceleration/deceleration, as estimated by epigenetic clocks, might be a promising tool to predict or to gain insights into the origin of various endometrial pathologies, including recurrent implantation failure, cancer and endometriosis. Finally, there are several strategies to slow down or even reverse epigenetic clocks that might be applied to reduce the risk of age-related uterine impairments.

WIDER IMPLICATIONS

The uterine factor should be considered, along with ovarian issues, to correct for the decline in female fertility with age. Epigenetic clocks can be tested to gain a deeper understanding of various endometrial disorders.