Biological Age Prediction Research Articles

DeepAge: Harnessing Deep Neural Network for Epigenetic Age Estimation From DNA Methylation Data of Human Blood Samples

Accurate prediction of biological age from DNA methylation data is a critical endeavor in understanding the molecular mechanisms of aging and developing age-related disease interventions. Traditional epigenetic clocks rely on linear regression or basic machine learning models, which often fail to capture the complex, non-linear interactions within methylation data. This study introduces DeepAge, a novel deep learning framework utilizing Temporal Convolutional Networks (TCNs) to enhance the prediction of biological age from DNA methylation profiles using selected CpGs by a Dual-Correlation based apparoach. DeepAge leverages a sequence-based approach with dilated convolutions to effectively capture long-range dependencies between CpG sites, addressing the limitations of prior models by incorporating advanced network architectures including residual connections and dropout regularization. The dual correlation feature selection enhances our model's predictive capabilities by identifying the most age-relevant CpG sites. Our model outperforms existing epigenetic clocks across multiple datasets, offering significant improvements in accuracy and providing deeper insights into the epigenetic determinants of aging. The proposed method not only sets a new standard in age estimation but also highlights the potential of deep learning in biologically relevant feature extraction and interpretation, contributing to the broader field of computational biology and precision medicine.

Epistemic uncertainty challenges aging clock reliability in predicting rejuvenation effects.

Epigenetic aging clocks have been widely used to validate rejuvenation effects during cellular reprogramming. However, these predictions are unverifiable because the true biological age of reprogrammed cells remains unknown. We present an analytical framework to consider rejuvenation predictions from the uncertainty perspective. Our analysis reveals that the DNA methylation profiles across reprogramming are poorly represented in the aging data used to train clock models, thus introducing high epistemic uncertainty in age estimations. Moreover, predictions of different published clocks are inconsistent, with some even suggesting zero or negative rejuvenation. While not questioning the possibility of age reversal, we show that the high clock uncertainty challenges the reliability of rejuvenation effects observed during invitro reprogramming before pluripotency and throughout embryogenesis. Conversely, our method reveals a significant age increase after invivo reprogramming. We recommend including uncertainty estimation in future aging clock models to avoid the risk of misinterpreting the results of biological age prediction.

Exploring the correlation between DNA methylation and biological age using an interpretable machine learning framework

DNA methylation plays a significant role in regulating transcription and exhibits a systematic change with age. These changes can be used to predict an individual’s age. First, to identify methylation sites associated with biological age; second, to construct a biological age prediction model and preliminarily explore the biological significance of methylation-associated genes using machine learning. A biological age prediction model was constructed using human methylation data through data preprocessing, feature selection procedures, statistical analysis, and machine learning techniques. Subsequently, 15 methylation data sets were subjected to in-depth analysis using SHAP, GO enrichment, and KEGG analysis. XGBoost, LightGBM, and CatBoost identified 15 groups of methylation sites associated with biological age. The cg23995914 locus was identified as the most significant contributor to predicting biological age by calculating SHAP values. Furthermore, GO enrichment and KEGG analyses were employed to initially explore the methylated loci’s biological significance.

Diffusion deep learning for brain age prediction and longitudinal tracking in children through adulthood

Abstract Deep learning (DL)-based prediction of biological age in the developing human from a brain magnetic resonance imaging (MRI) (“brain age”) may have important diagnostic and therapeutic applications as a non-invasive biomarker of brain health, aging, and neurocognition. While previous deep learning tools for predicting brain age have shown promising capabilities using single-institution, cross-sectional datasets, our work aims to advance the field by leveraging multi-site, longitudinal data with externally validated and independently implementable code to facilitate clinical translation and utility. This builds on prior foundational efforts in brain age modeling to enable broader generalization and individual’s longitudinal brain development. Here, we leveraged 32,851 T1-weighted MRI scans from healthy children and adolescents aged 3 to 30 from 16 multisite datasets to develop and evaluate several DL brain age frameworks, including a novel regression diffusion DL network (AgeDiffuse). In a multisite external validation (5 datasets), we found that AgeDiffuse outperformed conventional DL frameworks, with a mean absolute error (MAE) of 2.78 years (interquartile range [IQR]: [1.2-3.9]). In a second, separate external validation (3 datasets), AgeDiffuse yielded an MAE of 1.97 years (IQR: [0.8-2.8]). We found that AgeDiffuse brain age predictions reflected age-related brain structure volume changes better than biological age (R2 = 0.48 vs. R2 = 0.37). Finally, we found that longitudinal predicted brain age tracked closely with chronological age at the individual level. To enable independent validation and application, we made AgeDiffuse publicly available and usable for the research community.

Brain age of rhesus macaques over the lifespan

Through the application of machine learning algorithms to neuroimaging data the brain age methodology was shown to provide a useful individual-level biological age prediction and identify key brain regions responsible for the prediction. In this study, we present the methodology of constructing a rhesus macaque brain age model using a machine learning algorithm and discuss the key predictive brain regions in comparison to the human brain, to shed light on cross-species primate similarities and differences. Structural information of the brain (e.g., parcellated volumes) from brain magnetic resonance imaging of 43 rhesus macaques were used to develop brain atlas-based features to build a brain age model that predicts biological age. The best-performing model used 22 selected features and achieved an R2 of 0.72. We also identified interpretable predictive brain features including Right Fronto-orbital Cortex, Right Frontal Pole, Right Inferior Lateral Parietal Cortex, and Bilateral Posterior Central Operculum. Our findings provide converging evidence of the parallel and comparable brain regions responsible for both non-human primates and human biological age prediction.

A novel approach to quantifying individual's biological aging using Korea’s national health screening program toward precision public health

Accurate prediction of biological age can inform public health measures to extend healthy lifespans and reduce chronic conditions. Multiple theoretical models and methods have been developed; however, their applicability and accuracy are still not extensive. Here, we report Differential Aging and Health Index (DAnHI), a novel measure of age deviation, developed using physical and serum biomarkers from four million individuals in Korea's National Health Screening Program. Participants were grouped into aging statuses (< 26 vs. ≥ 26, < 27 vs. ≥ 27, …, < 75 vs. ≥ 75 years) as response variables in a binary logistic regression model with thirteen biomarkers as independent variables. DAnHI for each individual was calculated as the weighted mean of their relative probabilities of being classified into each older age status, based on model ages ranging from 26 to 75. DAnHI in our large study population showed a steady increase with the increase in age and was positively associated with death after adjusting for chronological age. However, the effect size of DAnHI on the risk of death varied according to the age group and sex. The hazard ratio was highest in the 50–59-year age group and then decreased as the individuals aged. This study demonstrates that routine health check-up biomarkers can be integrated into a quantitative measure for predicting aging-related health status and death via appropriate statistical models and methodology. Our DAnHI-based results suggest that the same level of aging-related health status does not indicate the same degree of risk for death.

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.

Application of a mathematical model to clarify the statistical characteristics of a pan-tissue DNA methylation clock.

DNA methylation clocks estimate biological age based on DNA methylation profiles. This study developed a mathematical model to describe DNA methylation aging and the establishment of a pan-tissue DNA methylation clock. The model simulates the aging dynamics of DNA methylation profiles based on passive demethylation as well as the process of cross-sectional bulk data acquisition. As a result, this study identified two conditions under which the pan-tissue DNA methylation clock can successfully predict biological age: one condition is that the target tissues are sufficiently well represented in the training dataset, and the other condition is that the target sample contains cell subsets that are common among different tissues. When either of these conditions is met, the clock performs well. It is also suggested that the epigenetic age of all samples in the target tissue tends to be either over or underestimated when biological age prediction fails. The model can reveal the statistical characteristics of DNA methylation clocks.

Development of a model for the prediction of biological age

Background and ObjectiveRates of aging vary markedly among individuals, and biological age serves as a more reliable predictor of current health status than does chronological age. As such, the ability to predict biological age can support appropriate and timely active interventions aimed at improving coping with the aging process. However, the aging process is highly complex and multifactorial. Therefore, it is more scientific to construct a prediction model for biological age from multiple dimensions systematically. MethodsPhysiological and biochemical parameters were evaluated to gage individual health status. Then, age-related indices were screened for inclusion in a model capable of predicting biological age. For subsequent modeling analyses, samples were divided into training and validation sets for subsequent deep learning model-based analyses (e.g. linear regression, lasso model, ridge regression, bayesian ridge regression, elasticity network, k-nearest neighbor, linear support vector machine, support vector machine, and decision tree models, and so on), with the model exhibiting the best ability to predict biological age thereby being identified. ResultsFirst, we defined the individual biological age according to the individual health status. Then, after 22 candidate indices (DNA methylation, leukocyte telomere length, and specific physiological and biochemical indicators) were screened for inclusion in a model capable of predicting biological age, 14 age-related indices and gender were used to construct a model via the Bagged Trees method, which was found to be the most reliable qualitative prediction model for biological age (accuracy=75.6%, AUC=0.84) by comparing 30 different classification algorithm models. The most reliable quantitative predictive model for biological age was found to be the model developed using the Rational Quadratic method (R2=0.85, RMSE=8.731 years) by comparing 24 regression algorithm models. ConclusionsBoth qualitative model and quantitative model of biological age were successfully constructed from a multi-dimensional and systematic perspective. The predictive performance of our models was similar in both smaller and larger datasets, making it well-suited to predicting a given individual's biological age.

A Targeted Epigenetic Clock for the Prediction of Biological Age.

Epigenetic clocks were initially developed to track chronological age, but accumulating evidence indicates that they can also predict biological age. They are usually based on the analysis of DNA methylation by genome-wide methods, but targeted approaches, based on the assessment of a small number of CpG sites, are advisable in several settings. In this study, we developed a targeted epigenetic clock purposely optimized for the measurement of biological age. The clock includes six genomic regions mapping in ELOVL2, NHLRC1, AIM2, EDARADD, SIRT7 and TFAP2E genes, selected from a re-analysis of existing microarray data, whose DNA methylation is measured by EpiTYPER assay. In healthy subjects (n = 278), epigenetic age calculated using the targeted clock was highly correlated with chronological age (Spearman correlation = 0.89). Most importantly, and in agreement with previous results from genome-wide clocks, epigenetic age was significantly higher and lower than expected in models of increased (persons with Down syndrome, n = 62) and decreased (centenarians, n = 106; centenarians' offspring, n = 143; nutritional intervention in elderly, n = 233) biological age, respectively. These results support the potential of our targeted epigenetic clock as a new marker of biological age and open its evaluation in large cohorts to further promote the assessment of biological age in healthcare practice.

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

BackgroundBiological age (BA) has been recognized as a more accurate indicator of aging than chronological age (CA). However, the current limitations include: insufficient attention to the incompleteness of medical data for constructing BA; Lack of machine learning-based BA (ML-BA) on the Chinese population; Neglect of the influence of model overfitting degree on the stability of the association results.Methods and resultsBased on the medical examination data of the Chinese population (45–90 years), we first evaluated the most suitable missing interpolation method, then constructed 14 ML-BAs based on biomarkers, and finally explored the associations between ML-BAs and health statuses (healthy risk indicators and disease). We found that round-robin linear regression interpolation performed best, while AutoEncoder showed the highest interpolation stability. We further illustrated the potential overfitting problem in ML-BAs, which affected the stability of ML-Bas’ associations with health statuses. We then proposed a composite ML-BA based on the Stacking method with a simple meta-model (STK-BA), which overcame the overfitting problem, and associated more strongly with CA (r = 0.66, P < 0.001), healthy risk indicators, disease counts, and six types of disease.ConclusionWe provided an improved aging measurement method for middle-aged and elderly groups in China, which can more stably capture aging characteristics other than CA, supporting the emerging application potential of machine learning in aging research.

A computational solution for bolstering reliability of epigenetic clocks: Implications for clinical trials and longitudinal tracking.

Epigenetic clocks are widely used aging biomarkers calculated from DNA methylation data, but this data can be surprisingly unreliable. Here we show technical noise produces deviations up to 9 years between replicates for six prominent epigenetic clocks, limiting their utility. We present a computational solution to bolster reliability, calculating principal components from CpG-level data as input for biological age prediction. Our retrained principal-component versions of six clocks show agreement between most replicates within 1.5 years, improved detection of clock associations and intervention effects, and reliable longitudinal trajectories in vivo and in vitro. This method entails only one additional step compared to traditional clocks, requires no replicates or prior knowledge of CpG reliabilities for training, and can be applied to any existing or future epigenetic biomarker. The high reliability of principal component-based clocks is critical for applications to personalized medicine, longitudinal tracking, in vitro studies, and clinical trials of aging interventions.

Epigenetic aging: Biological age prediction and informing a mechanistic theory of aging.

Numerous studies have shown that epigenetic age-an individual's degree of aging based on patterns of DNA methylation-can be computed and is associated with an array of factors including diet, lifestyle, genetics, and disease. One can expect that still further associations will emerge with additional aging research, but to what end? Prediction of age was an important first step, but-in our view-the focus must shift from chasing increasingly accurate age computations to understanding the links between the epigenome and the mechanisms and physiological changes of aging. Here, we outline emerging areas of epigenetic aging research that prioritize biological understanding and clinical application. First, we survey recent progress in epigenetic clocks, which are beginning to predict not only chronological age but aging outcomes such as all-cause mortality and onset of disease, or which integrate aging signals across multiple biological processes. Second, we discuss research that exemplifies how investigation of the epigenome is building a mechanistic theory of aging and informing clinical practice. Such examples include identifying methylation sites and the genes most strongly predictive of aging-a subset of which have shown strong potential as biomarkers of neurodegenerative disease and cancer; relating epigenetic clock predictions to hallmarks of aging; and using longitudinal studies of DNA methylation to characterize human disease, resulting in the discovery of epigenetic indications of type 1 diabetes and the propensity for psychotic experiences.

An epigenetic clock model for assessing the human biological age of healthy aging

Objective: To construct an epigenetic clock model for assessing and calibrating human biological age. Methods: Convenience sampling was used to select 186 subjects from the longevity cohort of Guangxi Zhuang Antonornous Region from July 1 to November 30, 2019, and 124 subjects from the physical examination population of the Seventh Medical Center of the PLA General Hospital from October 1 to December 31, 2020. Self-designed questionnaire was applied to collect demographic characteristics and family history of disease. Physical examination was applied to determine heart rate and blood pressure. Fasting peripheral venous blood was drawn for determination of fasting plasma glucose, plasma total cholesterol, triglyceride, plasma high-density lipoprotein cholesterol, plasma low-density lipoprotein cholesterol and telomere length. Methylation levels of EDARADD cg09809672, IPO8 cg19722847, NHLRC1 cg22736354, P2RX6 cg05442902 and SCGN cg06493994 were detected by targeted methylation site sequencing. A total of 54 subjects with unqualified quality control of DNA methylation and telomere length were excluded, and 256 subjects' data were finally analyzed. Trend test was used for the change of methylation level among different ages groups, multiple linear regression method was used to build prediction models of biological age. Kendal rank correlation analysis was used to evaluate the correlation of age gap (Gregorian calendar age minus biological age) with telomere length. Independent sample t-test was used to compare the health-related indicators between subjects with different age gap within different age groups. Results: The M(Q1, Q3)of age of subjects were 67 (51, 91) years old, including 166 females (64.84%). With increase of age, the methylation levels of gene loci were decreased (EDARADD cg09809672, IPO8 cg19722847 and P2RX6 cg05442902) and increased (NHLRC1 cg22736354 and SCGN cg06493994) (all P values<0.05). The established biological age prediction model was as follows: Y=-53.121×EDARADD cg09809672-137.564×IPO8 cg19722847+141.040×NHLRC1 cg22736354-67.893×P2RX6 cg05442902+149.547×SCGNcg06493994+4.592×sex+64.185 (R2=0.86, P<0.001), where Y was the biological age, and the items in the equation were methylation level, sex (male =1, female =2) and intercept in sequence. The Kendall rank correlation coefficient between age gap and telomere length was 0.731 (P<0.001). Compared with the subjects whose age gaP<0, the subjects with age gaP≥0 had higher systolic blood pressure in adolescence [(88.50±8.89) and (109.83±9.48) mmHg, respectively, 1 mmHg=0.133 kPa]; lower TC [(5.48±0.23) and (3.98±0.54) mmol/L, respectively, ] and TG [(3.51±0.32) and (3.41±0.20) mmol/L] in young adults; lower fasting blood glucose in middle age [(6.17±0.67) and (5.37±0.79) mmol/L, respectively, ] and higher diastolic blood pressure in nonagenarian age [(76.99±6.78) and (83.97±9.36) mmHg, respectively, ] (all P values<0.05). Conclusion: The constructed epigenetic clock model can be used to evaluate and calibrate human biological age.

Application of Correlation Pre-Filtering Neural Network to DNA Methylation Data: Biological Aging Prediction.

We introduce the CPFNN (Correlation Pre-Filtering Neural Network) for biological age prediction based on blood DNA methylation data. The model is built on 20,000 top correlated DNA methylation features and trained by 1810 healthy samples from GEO database. The input data format and the instructions for parser and CPFNN model are detailed in this chapter. Followed by two potential uses, age acceleration detection and unknown age prediction are discussed.

A Computational Solution to Bolster Epigenetic Clock Reliability for Clinical Trials and Longitudinal Tracking

Epigenetic clocks are widely used aging biomarkers, but they are calculated from methylation data for individual CpGs that can be surprisingly unreliable. We report that technical noise causes six major epigenetic clocks to deviate by 3 to 9 years between replicates. We present a novel computational solution: we perform principal component analysis followed by biological age prediction using principal components, extracting shared age-related changes across CpGs while ignoring noise from individual CpGs. Our novel principal-component versions of six clocks show agreement between most technical replicates within 1 year, and increased stability in short- and long-term longitudinal studies. This requires only one additional step compared to traditional clocks, does not require prior knowledge of CpG reliabilities, and can improve the reliability of any existing or future epigenetic biomarker. The extremely high reliability of principal component epigenetic clocks makes them particularly useful for personalized medicine and clinical trials evaluating novel aging interventions.

Comparison of Biological Age Prediction Models Using Clinical Biomarkers Commonly Measured in Clinical Practice Settings: AI Techniques Vs. Traditional Statistical Methods

In this work, we used the health check-up data of more than 111,000 subjects for analysis, using only the data with all 35 variables entered. For the prediction of biological age, traditional statistical methods and four AI techniques (RF, XGB, SVR, and DNN), which are widely used recently, were simultaneously used to compare the predictive power. This study showed that AI models produced about 1.6 times stronger linear relationship on average than statistical models. In addition, the regression analysis on the predicted BA and CA revealed similar differences in terms of both the correlation coefficients (linear model: 0.831, polynomial model: 0.996, XGB model: 0.66, RF model: 0.927, SVR model: 0.787, DNN model: 0.998) and R2 values. Through this work, we confirmed that AI techniques such as the DNN model outperformed traditional statistical methods in predicting biological age.

Omega 3/6 ratios: Usage and epigenetic changes in health and Alzheimer's disease

AbstractBackgroundFrom our previous research, we identified circulating metabolites that are associated with cognitive performance and dementia including polyunsaturated fatty acids (PUFA), in particular plasma levels of Omega‐3. To follow up on this finding, we sought to explore potential mechanism mediating the effect of Omega‐3 and Omega‐6 on these neurodegenerative phenotypes. Herein, we hypothesized that modulation of immune response and inflammation may mediate the effect of PUFA on neurodegenerative diseases.Methods1.Validate our PUFA ratios hypothesis on the MAPT intervention study. 2.Evaluate the epigenetic mechanisms modulation derived from genetics using genome‐wide association studies. 3.Corroborate previous results on real methylation data from the DELCODE study using the EPIC array. 4.Evaluate the biological age derived by epigenetics and its implications in healthy and cognitively affected subjects.ResultsThe effect of the PUFA ratios follows the same trend on the MAPT study as previously described during our first period. We observed that the methylation of the genes coding for key enzymes of PUFA metabolism is modulated by levels of Omega‐3. Likewise, these methylation changes have been described in susceptibility genes for Alzheimer’s disease (AD) that modulate immune response. Finally, methylated signals also enabled the prediction of biological age. We observed that biological age in DELCODE participants differs from chronological age depending on the status of the participants.ConclusionsThe present analysis suggests that PUFA ratios are associated with cognitive decline as we previously described. Moreover, we suggest that genetic signals located in the immune‐related gene cluster MS4A are modulated by changes in DNA methylation. The same epigenetic modulation seems to mediate the genetic effect discovered in the sTREM2 GWAS. Interestingly, methylation modulation in MS4A gene has been previously reported using an independent approach. Likewise, TREM2 locus seems to be affected by methylation changes. Finally, our methylation data allowed us to analyze the biological age of cognitively affected individuals and find some clear differences with their chronological ages. Using our results, we will explore whether PUFA levels, and the ratio built on Omega‐6 and Omega‐3, modulate levels of CSF sTREM2 offering thereby a potential connection between PUFA levels and neuroinflammation in the context of AD.

Prediction of biological age by morphological staging of sarcopenia in Caenorhabditis elegans.

ABSTRACTSarcopenia encompasses a progressive decline in muscle quantity and quality. Given its close association with ageing, it may represent a valuable healthspan marker. The commonalities with human muscle structure and facile visualization possibilities make Caenorhabditis elegans an attractive model for studying the relationship between sarcopenia and healthspan. However, classical visual assessment of muscle architecture is subjective and has low throughput. To resolve this, we have developed an image analysis pipeline for the quantification of muscle integrity in confocal microscopy images from a cohort of ageing myosin::GFP reporter worms. We extracted a variety of morphological descriptors and found a subset to scale linearly with age. This allowed establishing a linear model that predicts biological age from a morphological muscle signature. To validate the model, we evaluated muscle architecture in long-lived worms that are known to experience delayed sarcopenia by targeted knockdown of the daf-2 gene. We conclude that quantitative microscopy allows for staging sarcopenia in C. elegans and may foster the development of image-based screens in this model organism to identify modulators that mitigate age-related muscle frailty and thus improve healthspan.

Artificial Intelligence Confirming Treatment Success: The Role of Gender- and Age-Specific Scales in Performance Evaluation.

Summary:In plastic surgery and cosmetic dermatology, photographic data are an invaluable element of research and clinical practice. Additionally, the use of before and after images is a standard documentation method for procedures, and these images are particularly useful in consultations for effective communication with the patient. An artificial intelligence (AI)-based approach has been proven to have significant results in medical dermatology, plastic surgery, and antiaging procedures in recent years, with applications ranging from skin cancer screening to 3D face reconstructions, the prediction of biological age and perceived age. The increasing use of AI and computer vision methods is due to their noninvasive nature and their potential to provide remote diagnostics. This is especially helpful in instances where traveling to a physical office is complicated, as we have experienced in recent years with the global coronavirus pandemic. However, one question remains: how should the results of AI-based analysis be presented to enable personalization? In this paper, the author investigates the benefit of using gender- and age-specific scales to present skin parameter scores calculated using AI-based systems when analyzing image data.