A Unified Framework for Systematic Curation and Evaluation of Aging Biomarkers

A Unified Framework for Systematic Curation and Evaluation of Aging Biomarkers | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Resource A Unified Framework for Systematic Curation and Evaluation of Aging Biomarkers Mahdi Moqri, Kejun Ying, Seth Paulson, Alec Eames, Alexander Tyshkovskiy, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4481437/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 04 Nov, 2025 Read the published version in Nature Aging → Version 1 posted You are reading this latest preprint version Abstract Aging biomarkers are essential for understanding and quantifying the aging process and developing targeted longevity interventions. However, validation of these tools has been hindered by the lack of standardized approaches for cross-population validation, disparate biomarker designs, and inconsistencies in dataset structures. To address these challenges, we developed Biolearn, an open-source library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers. Leveraging Biolearn, we conducted a comprehensive evaluation of various aging biomarkers across multiple datasets. Our systematic approach involved three key steps: ( 1 ) harmonizing existing and novel aging biomarkers in standardized formats; ( 2 ) unifying public datasets to ensure coherent structuring and formatting; and ( 3 ) applying computational methodologies to assess the harmonized biomarkers against the unified datasets. This evaluation yielded valuable insights into the performance, robustness, and generalizability of aging biomarkers across different populations and datasets. The Biolearn python library, which forms the foundation of this systematic evaluation, is freely available at https://Bio-Learn.github.io . Our work establishes a unified framework for the curation and evaluation of aging biomarkers, paving the way for more efficient and effective clinical validation and application in the field of longevity research. Biological sciences/Computational biology and bioinformatics/Computational models Biological sciences/Computational biology and bioinformatics/Software Biological sciences/Systems biology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Development and validation of robust biomarkers of aging (BoAs) have become key focal points in aging research, driven by the growing recognition of aging as a fundamental driver of chronic diseases and mortality. Numerous biomarkers have been proposed to quantify biological age and elucidate the biological processes underlying aging. However, clinical validation of BoAs remains a significant challenge due to heterogeneity in their formulations and disparate structures of validation datasets across populations 1,2 . Since the introduction of composite omic biomarkers of aging, exemplified by Horvath’s pioneering work on DNA methylation aging clocks 3 , the field has witnessed a rapid expansion in the repertoire of aging biomarkers. These biomarkers now span a wide array of omic modalities, including epigenomics, transcriptomics, and proteomics 4–9 . Omic biomarkers provide a comprehensive view of the molecular changes associated with aging, offering valuable insights into the aging process and its impact on human health. Among the various classes of omic biomarkers, DNA methylation-based clocks are currently the most advanced and robust tools for estimating biological age. These human clocks, such as the Horvath multi-tissue clock, DunedinPACE 5 , GrimAge 10 , PhenoAge 11 , causality-enriched DamAge/AdaptAge 12 , and the PRC2 clock 13 , have demonstrated significant associations with age-related conditions and mortality, highlighting the intricate relationship between epigenetic modifications and aging trajectories 6,10,14,15 . However, the diverse formulations of these biomarkers and inconsistencies in dataset structures across different populations pose substantial challenges for their systematic cross-population validation and benchmarking, which are crucial steps toward their clinical translation. Publicly available datasets, such as those from the Gene Expression Omnibus (GEO) 16 , the National Health and Nutrition Examination Survey (NHANES), and the Framingham Heart Study (FHS), hold immense potential for accelerating the validation of BoAs. However, the lack of a standardized framework that can accommodate the heterogeneous nature of these datasets hinders their effective utilization for this purpose. There is a pressing need for a unified platform that can seamlessly integrate and analyze various BoAs across datasets with harmonized structures. Such a platform would revolutionize the validation process, facilitate the discovery of novel biomarkers, and provide a structured avenue for community-driven efforts in advancing the field of aging biology. To address this need, we developed Biolearn, an open-source Python library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers (Fig. 1 a). Biolearn supports biomarkers based on multiple different biological data modalities and serves as an innovative tool that harmonizes existing BoAs, structures and formats human datasets and offers computational methodologies for assessing biomarkers against these datasets. By enabling the integration and analysis of diverse BoAs and datasets, Biolearn aims to accelerate the development and validation of BoAs, fostering a community-driven approach to aging research. Results Harmonization of Biomarkers of Aging and Datasets We harmonized a comprehensive set of 39 well-established epigenetic, transcriptomic, and clinical biomarkers (Table 1 ) and implemented these BoAs in Biolearn, representing the largest collection of BoAs in a single package to date. We have validated the implementation of these biomarkers with their respective developers to ensure accuracy and reliability. The epigenetic biomarkers encompass a wide range of categories, including: ( 1 ) Chronological clocks: Horvath’s multi-tissue clock and Hannum’s blood clock 3,17 ; ( 2 ) Healthspan and mortality-related clocks: GrimAge, GrimAge2, PhenoAge, and Zhang clock 10,11,18,19 ; ( 3 ) Biomarkers of the rate of aging: DunedinPoAm38 and DunedinPACE 5,20 ; ( 4 ) Causality-enriched clocks: Ying’s CausAge, DamAge, and AdaptAge 12 ; and ( 5 ) Various other clocks, including DNAm-based biomarkers and disease predictors, transcriptomic clocks, and clinical clocks (Table 1 ). Table 1 Harmonized biomarkers in Biolearn. Biomarker Year Tissue Predicts Omic Type HorvathV1 3 2013 Multi-tissue Age (Years) DNA Methylation Hannum 17 2013 Blood Age (Years) DNA Methylation Lin 39 2016 Blood Age (Years) DNA Methylation PhenoAge 11 2018 Blood Age (Years) DNA Methylation HorvathV2 40 2018 Skin, blood Age (Years) DNA Methylation PEDBE 41 2019 Buccal Age (Years) DNA Methylation Zhang_10 19 2019 Blood Mortality Risk DNA Methylation GrimAge 10 2019 Blood Age Adjusted by Mortality Risk (Years) DNA Methylation GrimAge2 18 2022 Blood Age Adjusted by Mortality Risk (Years) DNA Methylation DunedinPoAm38 20 2020 Blood Aging Rate (Years/Year) DNA Methylation DunedinPACE 5 2022 Blood Aging Rate (Years/Year) DNA Methylation DNAmTL 42 2019 Blood, Adipose Telomere Length DNA Methylation Knight 43 2016 Cord Blood Gestational Age DNA Methylation LeeControl 44 2019 Placenta Gestational Age DNA Methylation LeeRefinedRobust 44 2019 Placenta Gestational Age DNA Methylation LeeRobust 44 2019 Placenta Gestational Age DNA Methylation YingCausAge 12 2022 Blood Age (Years) DNA Methylation YingDamAge 12 2022 Blood Age (Years) DNA Methylation YingAdaptAge 12 2022 Blood Age (Years) DNA Methylation SmokingMcCartney 45 2018 Blood Smoking Status DNA Methylation AlcoholMcCartney 45 2018 Blood Alcohol Consumption DNA Methylation BMI_McCartney 45 2018 Blood BMI DNA Methylation EducationMcCartney 45 2018 Blood Educational Attainment DNA Methylation TotalCholesterolMcCartney 45 2018 Blood Total Cholesterol DNA Methylation HDLCholesterolMcCartney 45 2018 Blood HDL Cholesterol DNA Methylation LDLCholesterolMcCartney 45 2018 Blood LDL with Remnant Cholesterol DNA Methylation BodyFatMcCartney 45 2018 Blood Percentage Body Fat DNA Methylation BMI_Reed 46 2020 Blood BMI DNA Methylation ProstateCancerKirby 47 2017 Prostate Prostate Cancer Status DNA Methylation HepatoXu 48 2017 Circulating DNA Hepatocellular Carcinoma Status DNA Methylation CVD_Westerman 49 2020 Blood Coronary Heart Disease Status DNA Methylation AD_Bahado-Singh 50 2021 Blood Alzheimer’s Disease Status DNA Methylation DepressionBarbu 51 2021 Blood Depression Risk DNA Methylation Phenotypic Age 11 2018 Blood Phenotypic Age Clinical Biomarker Mahalanobis Distance 34 2023 Blood Mahalanobis Distance Clinical Biomarker Peters_tAge 30 2015 Blood Age (Years) RNA Multispecies-blood_tAge 2024 Blood Relative Age (Years) RNA Human-blood_tAge 2024 Blood Relative Age (Years) RNA Human-multitissue_tAge 2024 Multi-tissue Relative Age (Years) RNA The harmonization of these diverse biomarkers is crucial for enabling consistent and reproducible analyses across different datasets, facilitating cross-population validation studies, and advancing our understanding of the aging process. To achieve this, all biomarkers were formatted into standardized input structures, ensuring their consistent application across disparate datasets. The harmonization process involved collecting and unifying the annotation of clock specifications, such as tissue type, predicted age range, and source references. This meticulous approach guarantees transparent and reproducible analyses, enabling researchers to readily compare and interpret results across different studies and populations. To further support ongoing research in this field, we developed and implemented an open-source framework. This framework provides a standardized format for epigenetic biomarkers, facilitating the seamless integration and comparison of any future aging clocks and biomarkers that are developed. By providing a unified platform for the harmonization and analysis of aging biomarkers, this framework aims to foster collaboration and innovation. Researchers can easily contribute new biomarkers, compare their performance against existing ones, and explore their potential applications in various datasets. This collaborative approach is essential for accelerating progress in the field and developing more accurate and robust biomarkers of aging. To facilitate cross-population validation studies using publicly available data, we harnessed Biolearn’s capabilities to integrate and structure multiple public datasets (Table 2 ). The structured datasets were refined to enable a shared analysis platform, addressing the challenges of data heterogeneity and formatting inconsistencies 21,22 . With this capacity, Biolearn is used as the backend of ClockBase for epigenetic age computation 22 , enabling the systemic harmonization of over 200,000 human samples from Gene Expression Omnibus (GEO) array data. Table 2 Harmonized datasets in Biolearn. ID Title Format Samples Age Present Sex Present GSE40279 Genome-wide Methylation Profiles Reveal Quantitative Views o… Illumina450k 656 Yes Yes GSE19711 Genome wide DNA methylation profiling of United Kingdom Ovar… Illumina27k 540 Yes No GSE51057 Methylome Analysis and Epigenetic Changes Associated with Me… Illumina450k 329 Yes Yes GSE42861 Differential DNA methylation in Rheumatoid arthritis Illumina450k 689 Yes Yes GSE41169 Blood DNA methylation profiles in a Dutch population Illumina450k 95 Yes Yes GSE51032 EPIC-Italy at HuGeF Illumina450k 845 Yes No GSE73103 Many obesity-associated SNPs strongly associate with DNA met… Illumina450k 355 Yes Yes GSE69270 Aging-associated DNA methylation changes in middle-aged indi… Illumina450k 184 Yes No GSE36054 Methylation Profiling of Blood DNA from Healthy Children Illumina450k 192 No No GSE64495 DNA methylation profiles of human blood samples from a sever… Illumina450k 113 Yes Yes GSE30870 DNA methylomes of Newborns and Nonagenarians Illumina450k 40 Yes No GSE52588 Identification of a DNA methylation signature in blood from … Illumina450k 87 Yes Yes GSE157131 Methylation data from stored peripheral blood leukocytes fro… IlluminaEPIC 946 Yes Yes GSE132203 DNA Methylation (EPIC) from the Grady Trauma Project IlluminaEPIC 795 Yes Yes GSE134080 RNASeq whole blood of Dutch 500FG cohort IlluminaHiSeq2500 100 Yes Yes NHANES National Health and Nutrition Examination Survey Phenotypic 2877 Yes Yes FHS Framingham Heart Study Phenotypic 4434 Yes Yes Quality Control, Imputation, and Deconvolution Biolearn provides a comprehensive toolkit for data preprocessing, normalization, and cell-type deconvolution (Fig. 1 a). Quality control metrics, such as sample deviation from the population mean, missingness, and the number of sites with a high percentage of missingness, can be readily visualized (Fig. 1 b). This functionality enables researchers to identify potential outliers or problematic samples and make informed decisions about data inclusion and exclusion criteria. Moreover, Biolearn enables the prediction of sample sex from DNA methylation data with high accuracy, which can be compared against actual sex distributions (Fig. 1 c,d) 23 . This feature is particularly useful for identifying potential sample mislabeling or investigating sex-specific effects in aging research. We also implemented predictors of common traits, including smoking, BMI, and epigenetic scores for diseases like Down Syndrome (Fig. 1 e) 24 . These predictors allow researchers to explore the associations between aging biomarkers and various lifestyle factors or disease conditions, providing valuable insights into the complex interplay between aging and health. Missing DNA methylation data can be easily imputed with different methods in just a few lines of code (Fig. 1 f,g) 25 , ensuring that researchers can make the most of the available data and minimize the impact of missing values on their analyses. Biolearn also facilitates the integration of multiple datasets for large-scale analyses, such as the comparison of DNA methylation levels of CpG sites across different datasets (Fig. 1 h). This feature allows researchers to investigate the consistency and reproducibility of aging biomarkers across diverse populations and experimental settings, strengthening the robustness and generalizability of their findings. In addition to these features, Biolearn offers a deconvolution tool that estimates the proportion of cell types in a given sample based on a single bulk-level methylation measurement. Biolearn provides two modes for deconvolution, optimized for the 450K (DeconvoluteBlood450K) and EPIC (DeconvoluteBloodEPIC) methylation platforms, respectively 26 . These modes are designed for estimating cell proportions in blood methylation samples and account for technology-specific biases that can affect the accuracy of deconvolution 27 . The reference methylation matrices for each mode consist of methylation profiles for six cell types representing the most abundant cell types found in the blood: neutrophils, monocytes, natural killer cells, B cells, CD4 + T cells, and CD8 + T cells 28,29 . We benchmarked the accuracy of our deconvolution tool using datasets with known cell proportions assessed via fluorescence-activated cell sorting (FACS) and in vitro cell mixing 29 . Both deconvolution methods generated accurate predictions that matched known cell proportions (Fig. 1 i,j). These results demonstrate the reliability and utility of Biolearn’s deconvolution feature, which can help researchers account for cellular heterogeneity in their analyses and gain insights into the cell type-specific contributions to aging biomarkers. Systematic Evaluation of the Biomarkers of Aging To demonstrate the utility of Biolearn in facilitating the systematic evaluation of aging biomarkers, we conducted a comprehensive benchmarking analysis of various epigenetic aging clocks across multiple datasets. By leveraging Biolearn’s harmonized dataset library and standardized clock implementations, we assessed the performance, robustness, and generalizability of these clocks in diverse biological contexts and populations (Fig. 2 ). Our analysis included a wide range of datasets, spanning the Human Aging Rates Study (GSE40279, N = 656), EPIC-Italy (GSE51032, N = 845), RA Case-control Cohort (GSE42861, N = 689), Dutch Schizophrenia Case-control Cohort (GSE41169, N = 95), Obesity Genetics Study (GSE73103, N = 355), Developmental Disorder Study (GSE64495, N = 113), African American GENOA (GSE157131, N = 1218), and Grady Trauma Project (GSE132203, N = 795) (Fig. 2 a). These diverse datasets allowed us to evaluate the performance of the biomarkers across various age ranges, ethnicities, and disease states. Biolearn’s user-friendly interface and efficient data handling capabilities streamlined data loading and implementation of aging clocks (Fig. 2 b). This highlights the library’s potential to accelerate the development and validation of novel aging biomarkers by providing a standardized framework for their evaluation. The benchmarking results (Fig. 2 a-c) revealed that the HorvathV2 clock (i.e., skin and blood clock) exhibited the highest overall accuracy in terms of predicting chronological age, with a mean R 2 of 0.88 across all datasets, followed closely by the Hannum clock (R 2 = 0.81), the Horvath1 clock (R 2 = 0.78) and YingCausAge clock (R 2 = 0.77). These findings suggest that these four clocks are the most robust and generalizable across diverse biological contexts and age-related conditions. Note that the GrimAgeV1 and GrimAgeV2 clocks use the age of the sample as the predictor, therefore they cannot be compared directly to the other clocks. These results highlight the applicability of these clocks across diverse age ranges and biological contexts, further emphasizing the importance of systematic evaluation in identifying the most suitable biomarkers for specific research questions or clinical applications. Overall, our findings demonstrate the value of Biolearn in enabling the systematic evaluation of epigenetic aging clocks across multiple datasets. By providing a standardized framework for clock implementation and evaluation, Biolearn facilitates the identification of robust and generalizable aging biomarkers, paving the way for their translation into clinical settings and advancing our understanding of the aging process. Mortality and Morbidity Risk Analysis It is also important to evaluate the predictive power of epigenetic aging biomarkers in predicting aging-associated outcomes such as mortality risk. Here, we conducted the most comprehensive evaluation of 17 representative epigenetic clock models to the Normative Aging Study (NAS) dataset (N = 1,488, 38.8% deceased) and the Massachusetts General Brigham (MGB) cohort (N = 500, 8.8% deceased), comparing their performance in predicting mortality risk (Fig. 3 ). The biomarkers showed a strong correlation with chronological age in both NAS (Fig. 3 a) and MGB cohorts (Fig. 3 b). We then examined their performance in predicting mortality risk using Cox Proportional Hazards analysis, adjusted for age and sex (Fig. 3 c, d). In the NAS cohort (Fig. 3 c), the top-performing clock in terms of hazard ratio (per standard deviation increase of the biomarker) was DunedinPoAm38 (HR = 1.38, P = 9.48e-18), followed closely by GrimAgeV2 (HR = 1.35, P = 1.21e-18) and DunedinPACE (HR = 1.35, P = 3.04e-19). Other clocks with strong predictive value include Zhang_10 (HR = 1.28, P = 9.66e-12), GrimAgeV1 (HR = 1.29, P = 7.05e-14), YingDamAge (HR = 1.25, P = 1.48e-11), PhenoAge (HR = 1.20, P = 8.96e-08), Hannum (HR = 1.13, P = 3.91e-04), and YingCausAge (HR = 1.09, P = 7.37e-03). In the MGB cohort (Fig. 3 d), the top-performing clocks in predicting mortality risk were GrimAgeV2 (HR = 2.08, P = 2.64e-03), PhenoAge (HR = 2.03, P = 2.52e-02), and GrimAgeV1 (HR = 1.84, P = 1.46e-02). Followed by Zhang_10 (HR = 1.48, P = 7.96e-05), and DunedinPACE (HR = 1.46, P = 7.40e-04). We further assessed the association between the predictive power of the epigenetic clocks for chronological age and mortality risk. Interestingly, in both cohorts, we observed a negative but insignificant correlation between Pearson’s R with chronological age and hazard ratio of mortality risk (Fig. 3 e). This suggests that the predictive power of epigenetic clocks on mortality risk, after adjusting for age and sex, is independent of their ability to predict chronological age, highlighting the importance of interpreting the meaning of age deviation (AgeDev) with caution for aging biomarkers. We also observed strong heterogeneous associations of epigenetic clocks with mortality risk in different cohorts. The analysis demonstrates the utility of Biolearn in facilitating systematic evaluation of epigenetic clocks in predicting mortality across multiple cohorts, emphasizing the importance of systematic evaluation in identifying the most suitable biomarkers for specific applications. Besides mortality, it is also important to evaluate aging biomarkers in predicting various other clinically relevant aging outcomes. We analyzed the associations between 14 aging biomarkers and six event categories (Stroke, Dementia, Operation, Lifespan, Cancer, and Healthspan, which is defined by the first incidence of any event) in the NAS cohort (Fig. 4 a). To assess the predictive power of these biomarkers, we performed Cox Proportional Hazards analyses, adjusted for age, and calculated the hazard ratios (HR) per standard deviation increase for each biomarker (Fig. 4 b-f). Across five clinical outcomes tested, DunedinPACE was the strongest predictor for three of the outcomes, namely healthspan (HR = 1.18), dementia (HR = 1.40), and stroke (HR = 1.50). PhenoAge was the strongest predictor for surgery (HR = 1.24), and HorvathV2 was the strongest predictor for cancer (HR = 1.12). These results suggest considerable heterogeneity of aging biomarkers in predicting different clinical outcomes. We further investigated the associations between the AgeDev term of aging biomarkers after adjusting for age in the NAS cohort and found strong positive correlations among most epigenetic clocks (Fig. 4 g). We observed two main clusters: ( 1 ) PhenoAge, GrimAgeV1, GrimAgeV2, YingCausAge, HorvathV1, Lin, Hannum, and HorvathV2; and ( 2 ) YingDamAge, DunedinPoAm38, Zhang10, and DunedinPACE. The DNAmTL telomere length clock and YingAdaptAge do not cluster with other clocks, suggesting their unique biological underpinnings. Lastly, we compared the predictive power of aging biomarkers for healthspan and lifespan (Fig. 4 h). In general, we observed a significant positive correlation using an inverse-variance-weighted approach (weighted correlation coefficient = 0.58, P = 0.012). DunedinPoAm38, GrimAgeV2, GrimAgeV1, and Zhang_10 demonstrated strong associations with healthspan and lifespan, indicating their potential as comprehensive aging biomarkers. These findings highlight the utility of Biolearn in facilitating the systematic evaluation of aging biomarkers for predicting various health outcomes. The results provide insights into the comparative performance of these biomarkers and their potential applications in clinical settings and aging research. Transcriptomic and Clinical Biomarkers To demonstrate Biolearn’s multi-omic capabilities, we also evaluated the performance of transcriptomic and phenotypic aging biomarkers. We include four transcriptomic age predictors: tAge.Peters 30 , tAge.Multispecies.Blood, tAge.Human.Multi-tissue, and tAge.Human.Blood 31,32 . We applied these predictors to the JenAge RNA-Seq dataset (Jena Centre for Systems Biology of Ageing, Illumina TruSeq 2.0, Whole Blood, N = 62) 33 . The tAge predictor showed a strong correlation with chronological age, with Pearson’s R ranging from 0.68 (Peters) to 0.90 (Human.Multi-tissue) (Fig. 5 a), highlighting their potential as robust aging biomarkers. All these predictors are implemented in Biolearn with simple and easy-to-use functions (Fig. 5 b). Next, we investigated the performance of the Phenotypic Age predictor 11 , a blood-test-based biomarker, using the NHANES 2010 dataset. We calculated the Phenotypic Age for each individual using Biolearn’s implementation of the predictor and compared it to chronological age (Fig. 5 c). We found a strong linear relationship between Phenotypic Age and chronological age (Pearson’s R = 0.96, P < 2.2e-16), indicating the predictor’s ability to capture age-related changes in clinical biomarkers. To assess the predictive power of Phenotypic Age on mortality risk, we performed a survival analysis on the NHANES 2010 dataset. Individuals were stratified into five groups based on their AgeDev (Fig. 5 d,e). We found that individuals with higher AgeDev had a significantly higher mortality risk (HR = 1.58, 95% CI: 1.31–1.91, P = 1.08e-06), while those with lower AgeDev had a lower risk (HR = 0.60, 95% CI: 0.49–0.74, P = 6.19e-07). These results underscore the predictive power of Phenotypic Age in assessing mortality risk and demonstrate Biolearn’s capability to facilitate such analyses. To investigate the association between aging and cell type proportions, we performed deconvolution analysis on the NAS cohort blood samples. This analysis revealed distinct changes for different cell types over the life course (Extended Fig. 1 ). Neutrophil and natural killer cell proportions showed significant positive correlations with age (R = 0.09, P = 5.64e-04 and R = 0.14, P = 1.03e-07, respectively). In contrast, B cell, CD4 T cell, and CD8 T cell proportions exhibited significant negative correlations with age (R=-0.09, P = 2.45e-04; R=-0.15, P = 6.07e-09; and R=-0.05, P = 4.01e-02, respectively). Monocyte proportion did not show a significant correlation with age (R = 0.02, P = 4.07e-01). We assessed the predictive power of cell type proportions on various health outcomes in the NAS cohort (Extended Fig. 2 ). For healthspan, natural killer cell proportion was a significant protective factor (HR = 0.93, 95% CI: 0.88–0.99, P = 1.55e-02), while CD8 T cell proportion was a slight risk factor (HR = 1.06, 95% CI: 1.00-1.12, P = 4.65e-02). Similarly, for lifespan, natural killer cell proportion was a significant protective factor (HR = 0.91, 95% CI: 0.85–0.98, P = 1.03e-02), and CD8 T cell proportion was a significant risk factor (HR = 1.07, 95% CI: 1.01–1.14, P = 2.59e-02). Natural killer cell proportion was also a significant protective factor for dementia (HR = 0.87, 95% CI: 0.77–0.98, P = 2.03e-02). For stroke risk, neutrophil proportion was a significant risk factor (HR = 1.33, 95% CI: 1.11–1.60, P = 2.18e-03), while natural killer cell proportion was a significant protective factor (HR = 0.70, 95% CI: 0.55–0.89, P = 4.21e-03). Similarly, for surgical events, neutrophil proportion was a significant risk factor (HR = 1.16, 95% CI: 1.05–1.28, P = 3.99e-03), and natural killer cell proportion was a significant protective factor (HR = 0.89, 95% CI: 0.80–0.99, P = 2.58e-02). No significant associations were found between cell type proportions and cancer risk. The integration of transcriptomic and phenotypic biomarkers in Biolearn enables researchers to investigate aging processes from different biological perspectives. The strong performance of the tAge predictor and Phenotypic Age in their respective datasets showcases the potential of multi-omic approaches in uncovering the complex mechanisms underlying aging. By leveraging Biolearn’s comprehensive framework, researchers can gain valuable insights into the interplay between different biological layers and their contributions to the aging process, ultimately facilitating the development of targeted interventions and personalized aging management strategies. Discussion Among the most significant challenges in aging biomarker research is cross-population validation of proposed biomarkers 34 . To take steps to address this need and provide an open-source tool for validation efforts across the field, we built Biolearn, an open-source library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers across diverse datasets. By leveraging Biolearn, we conducted a comprehensive benchmarking analysis of various epigenetic aging clocks, transcriptomic predictors, and phenotypic biomarkers, demonstrating their performance, robustness, and generalizability in different biological contexts and populations. Systematic evaluation of epigenetic aging clocks across multiple datasets revealed that the Horvath skin and blood clock, Hannum clock, Horvath multi-tissue clock and Ying CausAge clock exhibited the highest overall accuracy in predicting chronological age. Notably, the predictive power of epigenetic clocks on mortality risk, after adjusting for age and sex, was independent of their ability to predict chronological age, highlighting the importance of interpreting AgeDev with caution and underscoring the need for further investigation into the biological mechanisms captured by these clocks. Evaluation of aging biomarkers in predicting various clinical outcomes in the NAS cohort revealed considerable heterogeneity, with different biomarkers showing strengths in predicting specific outcomes. For example, DunedinPACE was the strongest predictor for healthspan, dementia, and stroke, while PhenoAge and the Horvath skin and blood clock were the strongest predictors for surgery and cancer, respectively. These findings underscore the importance of selecting appropriate biomarkers based on the specific clinical outcomes of interest and suggest the potential for developing targeted interventions tailored to individual aging trajectories. Integration of transcriptomic and phenotypic biomarkers in Biolearn enables a multi-omic approach to investigating the aging process. The strong performance of the tAge predictor and Phenotypic Age in their respective datasets highlights the potential of combining information from different biological layers to gain a more comprehensive understanding of the aging process. By leveraging Biolearn’s unified framework, researchers can explore the interplay between various omic modalities and uncover novel insights into the complex mechanisms underlying aging. With Biolearn, we also harmonized and evaluated several well-established aging clocks, providing the opportunity for these biomarkers to be refined and potentially for new ones to be developed. The modular design of Biolearn encourages the addition of new models and datasets, making it a living library that will grow in tandem with the field itself. By centralizing resources and knowledge, Biolearn considerably reduces redundancy and accelerates biomarker development and validation efforts 5,9 . Our approach emphasizes transparency and reproducibility, core tenets of open science. By making Biolearn publicly available and maintaining detailed documentation and development guidelines, we established an ecosystem that supports open collaboration and knowledge sharing. This open-science framework ensures that findings and tools can be widely accessed, providing equitable opportunities for researchers globally to contribute to and benefit from the collective advances in aging research. Moreover, our hope is that the open-access nature of Biolearn will promote cross-fertilization between aging researchers and scientists currently outside the field, incentivizing the development of novel and innovative biomarker models and validation approaches. Previous efforts to harmonize biomarkers of aging, notably methylCIPHER and BioAge 35,36 , have been limited in scope, focusing on methylation or blood-based biomarkers only. Furthermore, using R packages is somewhat limiting. Biolearn supports biomarkers based on multiple different biological data modalities and is written in Python, which has a broader reach. In comparison to PyAging 37 , a preliminary contemporaneous Python biomarker library, Biolearn is focused on ease of use and reproducibility through automated testing against reference data. Biolearn also offers several distinct advantages over existing biomarker libraries: it supports the easy loading of a larger and more diverse set of data, enabling researchers to work with a wide range of datasets and explore the performance of biomarkers across various populations and biological contexts; it includes models that are not directly used for age prediction but may be relevant to health and lifespan, providing a more comprehensive toolkit for aging research; and it includes tooling for exploring and understanding your data, such as quality reports, which facilitate data preprocessing and ensure the reliability of the results. While Biolearn represents a significant advance for the field, some limitations remain. Currently, the library is tailored to biomarkers derived from biological samples, predominantly DNA methylation data. Moving forward, the scope of Biolearn will continue to expand to encompass diverse biological modalities—such as proteomics, metabolomics, and microbiomics—broadening its applicability 7,13 . Moreover, integration with larger and more diverse population datasets will be vital in advancing cross-population validation efforts. As new datasets emerge, Biolearn will adapt to incorporate these resources, ensuring ongoing robustness and scalability 38 . Finally, bioinformatics tools, including Biolearn, depend on a user base proficient in programming and data analysis. Efforts to make these tools more accessible to a wider audience, including those with limited computational expertise, will be crucial. This could involve the development of graphical user interfaces (GUIs) or web-based platforms to streamline the user experience. We anticipate that Biolearn will become a key resource for the field and will transform many facets of aging biomarker studies. Our preliminary survival studies conducted using Biolearn demonstrate not only the power of this new platform but illuminate the real-world implications of validated biomarkers. Biolearn’s standardization and analysis capabilities stand to serve as pivotal tools for researchers seeking to bridge the gap between biomarker discovery and clinical implementation 34 . Methods Overview of Biolearn Library Biolearn is an open-source computational suite that facilitates the harmonization and analysis of biomarkers of aging (BoAs). It is written in Python and is readily accessible through the Python Package Index (PyPI). Biolearn is developed using modern software engineering practices, including automated testing to ensure correctness and adherence to software design principles that ensure the safe interchangeability of like components. The library is designed to be user-friendly while offering robust functionalities for researchers across various disciplines involved in aging studies. Additionally, the results of each model in Biolearn are portable and can be easily exported in formats such as CSV, allowing for seamless integration with other tools like R for further analysis and visualization. System Requirements and Installation Biolearn requires Python version 3.10 or newer. It can be installed using the Python package manager, pip, with the command pip install biolearn. The successful installation of the library can be verified through the import test of Biolearn’s core classes. The library is cross-platform and is compatible with major operating systems, including Windows, MacOS, and Linux. Data Library and Model Gallery Biolearn incorporates a data library capable of loading and structuring datasets from a multitude of public sources like Gene Expression Omnibus (GEO), National Health and Nutrition Examination Survey (NHANES), and Framingham Heart Study. The model gallery within Biolearn holds reference implementations for various aging clocks and biomarkers, presenting a unified interface for users to apply these models to their datasets. All models were verified to be correct by comparing the outputs on a reference data set against their original implementations where available. Harmonization Process We used Biolearn to harmonize several aging clocks. Clock definitions were standardized, specifying the name, publication year, applicable species, target tissue, and the biological aspect they predict (e.g., age, mortality risk). We provided sources for both the original publications and the coefficients necessary for clock applications. Coherence across biological modalities and datasets was assured through Biolearn’s systematic approach to data preprocessing, normalization, and imputation procedures. Integration with Public Datasets Biolearn’s ability to interface seamlessly with public datasets was tested by integrating and formatting data from GEO and NHANES. Preprocessing pipelines were developed to convert raw data into a harmonized format suitable for subsequent analysis. Particular attention was given to metadata structures, variable normalization, and missing data treatment, ensuring consistent input formats required by the aging models. Cell Type Deconvolution Biolearn’s deconvolution function estimates cell type proportions within a sample from bulk methylation data. It operates on the principle that bulk methylation is a composite of methylation profiles from various cell types, proportionate to their presence (B = X * P). Here, ‘B’ represents bulk methylation, ‘X’ is the matrix of cell-type-specific methylation profiles, and ‘P’ is the vector of cell-type proportions. With known ‘B’ and estimated ‘X,’ ‘P’ is computable through constrained quadratic programming, ensuring proportions remain within biological plausibility—between zero and one and summing to one 26 . We incorporated two deconvolution variants tailored for blood methylation analysis: DeconvoluteBlood450K for the 450K platform and DeconvoluteBloodEPIC for the EPIC platform, addressing platform-specific biases 27 . The methylation matrix for each mode derives from the corresponding platform, encapsulating six major blood cell types—neutrophils, monocytes, NK cells, B cells, CD4 + T cells, and CD8 + T cells. Selection of CpG sites for deconvolution relied on identifying the 50 most distinctively hyper- and hypo-methylated sites per cell type, prioritized by the significance of their differential methylation. This yielded 600 reference CpG sites per deconvolution mode 28,29 . Statistical Analysis All statistical analyses were performed using tools embedded within the Biolearn library or through integration with renowned Python statistics libraries such as statsmodels and seaborn for visualization. The robustness and reproducibility of the analysis were ensured through the use of randomized cross-validation techniques for model assessment and bootstrapping methods for estimating confidence intervals where applicable. Survival analyses were conducted using the Cox Proportional Hazards model, adjusting for age and other relevant covariates. The performance of aging clocks in predicting chronological age and mortality risk was evaluated using metrics such as R 2 , hazard ratios, and p-values. Declarations Acknowledgments We thank all Biomarkers of Aging Consortium members for their valuable feedback and suggestions. We thank S. Horvath and A. Lu for sharing the GrimAgeV1 and V2. This work was inspired by methylCIPHER, an R package for DNA methylation clocks 36 . Supported by grants from the National Institute on Aging, Hevolution Foundation, Methuselah Foundation, and VoLo Foundation. References Moqri M et al (2024) A framework for validation of omic biomarkers of aging. Nat Med Press Moqri M et al (2023) Biomarkers of aging for the identification and evaluation of longevity interventions. Cell 186:3758–3775 Horvath S (2013) DNA methylation age of human tissues and cell types. Genome Biol 14:R115 Bell CG et al (2019) DNA methylation aging clocks: challenges and recommendations. Genome Biol 20:249 Belsky DW et al (2022) DunedinPACE, a DNA methylation biomarker of the pace of aging. Elife 11:e73420 Chen BH et al (2016) DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging 8:1844–1865 Galkin F et al (2020) Biohorology and biomarkers of aging: Current state-of-the-art, challenges and opportunities. Ageing Res Rev 60:101050 Horvath S, Raj K (2018) DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet 19:371–384 Jylhävä J, Pedersen NL, Hägg S (2017) Biol Age Predictors EBioMedicine 21:29–36 Lu AT et al (2019) DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging 11:303–327 Levine ME et al (2018) An epigenetic biomarker of aging for lifespan and healthspan. Aging 10:573–591 Ying K et al (2024) Causality-enriched epigenetic age uncouples damage and adaptation. Nat Aging 1–16. 10.1038/s43587-023-00557-0 Moqri M et al (2022) PRC2 clock: a universal epigenetic biomarker of aging and rejuvenation. 06.03.494609 Preprint at https://doi.org/10.1101/2022.06.03.494609 (2022) Field AE et al (2018) DNA Methylation Clocks in Aging: Categories, Causes, and Consequences. Mol Cell 71:882–895 Marioni RE et al (2015) The epigenetic clock is correlated with physical and cognitive fitness in the Lothian Birth Cohort 1936. Int J Epidemiol 44:1388–1396 Edgar R (2002) Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30:207–210 Hannum G et al (2013) Genome-wide Methylation Profiles Reveal Quantitative Views of Human Aging Rates. Mol Cell 49:359–367 Lu AT et al (2022) DNA methylation GrimAge version 2. Aging 14, 9484–9549 Zhang Q et al (2019) Improved precision of epigenetic clock estimates across tissues and its implication for biological ageing. Genome Med 11:54 Belsky DW et al (2020) Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. eLife 9, e54870 Thompson MJ et al (2018) A multi-tissue full lifespan epigenetic clock for mice. Aging 10:2832–2854 Ying K et al (2023) ClockBase : a comprehensive platform for biological age profiling in human and mouse. Preprint at. https://doi.org/10.1101/2023.02.28.530532 Wang Y et al (2021) DNA methylation-based sex classifier to predict sex and identify sex chromosome aneuploidy. BMC Genomics 22:484 Horvath S et al (2015) Accelerated epigenetic aging in Down syndrome. Aging Cell 14:491–495 Lena PD, Sala C, Prodi A, Nardini C (2020) Methylation data imputation performances under different representations and missingness patterns. BMC Bioinformatics 21:268 Houseman EA et al (2012) DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics 13:86 Hicks SC, Irizarry RA (2019) methylCC: technology-independent estimation of cell type composition using differentially methylated regions. Genome Biol 20:261 Reinius LE et al (2012) Differential DNA Methylation in Purified Human Blood Cells: Implications for Cell Lineage and Studies on Disease Susceptibility. PLoS ONE 7:e41361 Salas LA et al (2018) An optimized library for reference-based deconvolution of whole-blood biospecimens assayed using the Illumina HumanMethylationEPIC BeadArray. Genome Biol 19:64 Peters MJ et al (2015) The transcriptional landscape of age in human peripheral blood. Nat Commun 6:8570 Tyshkovskiy A et al (2019) Identification and Application of Gene Expression Signatures Associated with Lifespan Extension. Cell Metabol 30:573–593e8 Tyshkovskiy A et al (2023) Distinct longevity mechanisms across and within species and their association with aging. Cell 186:2929–2949e20 Aramillo Irizar P et al (2018) Transcriptomic alterations during ageing reflect the shift from cancer to degenerative diseases in the elderly. Nat Commun 9:327 Li Q et al (2023) Biomarkers of aging for the identification and evaluation of longevity interventions. Popul Health Metrics 186:3758–3775 Kwon D, Belsky DW (2021) A toolkit for quantification of biological age from blood chemistry and organ function test data: BioAge. GeroScience 43, 2795–2808 Thrush KL, Higgins-Chen AT, Liu Z, Levine ME (2022) R methylCIPHER: A Methylation Clock Investigational Package for Hypothesis-Driven Evaluation & Research. 07.13.499978 Preprint at https://doi.org/10.1101/2022.07.13.499978 (2022) Camillo LP (2023) de L. pyaging: a Python-based compendium of GPU-optimized aging clocks. 11.28.569069 Preprint at https://doi.org/10.1101/2023.11.28.569069 (2023) Yang J et al (2020) Prevalence of comorbidities and its effects in patients infected with SARS-CoV-2: a systematic review and meta-analysis. Int J Infect Dis 94:91–95 Lin Q et al (2016) DNA methylation levels at individual age-associated CpG sites can be indicative for life expectancy. Aging 8:394–401 Horvath S et al (2018) Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria Syndrome and ex vivo studies. Aging 10:1758–1775 McEwen LM et al (2020) The PedBE clock accurately estimates DNA methylation age in pediatric buccal cells. Proc. Natl. Acad. Sci. U.S.A. 117, 23329–23335 Lu AT et al (2019) DNA methylation-based estimator of telomere length. Aging 11:5895–5923 Knight AK et al (2016) An epigenetic clock for gestational age at birth based on blood methylation data. Genome Biol 17:206 Lee Y et al (2019) Placental epigenetic clocks: estimating gestational age using placental DNA methylation levels. Aging 11:4238–4253 McCartney DL et al (2018) Epigenetic prediction of complex traits and death. Genome Biol 19:136 Reed ZE, Suderman MJ, Relton CL, Davis OSP, Hemani G (2020) The association of DNA methylation with body mass index: distinguishing between predictors and biomarkers. Clin Epigenetics 12:50 Kirby MK et al (2017) Genome-wide DNA methylation measurements in prostate tissues uncovers novel prostate cancer diagnostic biomarkers and transcription factor binding patterns. BMC Cancer 17:273 Xu R et al (2017) Circulating tumour DNA methylation markers for diagnosis and prognosis of hepatocellular carcinoma. Nat Mater 16:1155–1161 Westerman K et al (2020) Epigenomic Assessment of Cardiovascular Disease Risk and Interactions With Traditional Risk Metrics. J Am Heart Assoc 9:e015299 Bahado-Singh RO et al (2021) Artificial intelligence and leukocyte epigenomics: Evaluation and prediction of late-onset Alzheimer’s disease. PLoS ONE 16:e0248375 Barbu MC et al (2021) Epigenetic prediction of major depressive disorder. Mol Psychiatry 26:5112–5123 Additional Declarations There is NO Competing Interest. Supplementary Files supplementarymaterial.docx Cite Share Download PDF Status: Published Journal Publication published 04 Nov, 2025 Read the published version in Nature Aging → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4481437","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Resource","associatedPublications":[],"authors":[{"id":330806555,"identity":"08f35182-c248-4fb6-af04-b8dcc7c0f0ac","order_by":0,"name":"Mahdi Moqri","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYFACHgaJBDCDsfEBiMtHipZmAxCXjSgtUBYbmEFQi2772YM3HjBsk+O7fbit8muOnQwbA/PDRzfwaDE7k5dskcBw21jyXGLbbdltyUCHsRkb5+DTciDHDOiX24kbzjC23ZbcxgzUwsMmjVfL+TcILcWS2+qJ0HIDyRbGj9sOE6PljbFFggHQL2cYm6UZtx3nYWMm5JfzOYY3f1TcluM7w/7w489t1fb87M0PH+PTAgGgODzAwMDMA+IwE1QOA0AtjD+IVj0KRsEoGAUjCQAAcNRHkXPXhsAAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-6675-0566","institution":"Harvard Medical School and Brigham and Women's Hospital","correspondingAuthor":true,"prefix":"","firstName":"Mahdi","middleName":"","lastName":"Moqri","suffix":""},{"id":330806556,"identity":"141a7333-c8a2-44e5-865e-d93e9fc1c068","order_by":1,"name":"Kejun Ying","email":"","orcid":"https://orcid.org/0000-0002-1791-6176","institution":"Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Kejun","middleName":"","lastName":"Ying","suffix":""},{"id":330806557,"identity":"df9842d9-34a8-4ab1-9ac7-f4f0cd2e40e3","order_by":2,"name":"Seth Paulson","email":"","orcid":"","institution":"Methuselah Foundation","correspondingAuthor":false,"prefix":"","firstName":"Seth","middleName":"","lastName":"Paulson","suffix":""},{"id":330806558,"identity":"bc5a8423-76b2-4bbf-9849-f1151b8d13bf","order_by":3,"name":"Alec Eames","email":"","orcid":"","institution":"Division of Genetics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Alec","middleName":"","lastName":"Eames","suffix":""},{"id":330806559,"identity":"ba338db5-7879-46a4-9b43-9a480a4ea1fa","order_by":4,"name":"Alexander Tyshkovskiy","email":"","orcid":"https://orcid.org/0000-0002-6215-190X","institution":"Brigham and Women’s Hospital, Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"","lastName":"Tyshkovskiy","suffix":""},{"id":330806560,"identity":"957a6e3a-a497-40d7-8dbd-af2814981a23","order_by":5,"name":"Siyuan Li","email":"","orcid":"https://orcid.org/0009-0000-9254-7115","institution":"Division of Genetics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Siyuan","middleName":"","lastName":"Li","suffix":""},{"id":330806561,"identity":"ff9668ee-48af-4bdc-b19a-adefa1996b5f","order_by":6,"name":"Martin Perez-Guevara","email":"","orcid":"","institution":"Mind Operating Systems","correspondingAuthor":false,"prefix":"","firstName":"Martin","middleName":"","lastName":"Perez-Guevara","suffix":""},{"id":330806562,"identity":"a6cfda1d-bfe5-4bfc-a066-880723ff98bd","order_by":7,"name":"Mehrnoosh Emamifar","email":"","orcid":"","institution":"Department of Bioengineering, Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Mehrnoosh","middleName":"","lastName":"Emamifar","suffix":""},{"id":330806563,"identity":"38565a8e-46c1-44f8-ae1b-e4a0c229fc26","order_by":8,"name":"Maximiliano Casas Martinez","email":"","orcid":"","institution":"Division of Exact Sciences, Department of Mathematics, Instituto Tecnológico Autónomo de México","correspondingAuthor":false,"prefix":"","firstName":"Maximiliano","middleName":"Casas","lastName":"Martinez","suffix":""},{"id":330806564,"identity":"fb7244d5-250e-4b1b-a945-18e631e00ec7","order_by":9,"name":"Dayoon Kwon","email":"","orcid":"","institution":"Department of Epidemiology, UCLA Fielding School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Dayoon","middleName":"","lastName":"Kwon","suffix":""},{"id":330806565,"identity":"7c8c72bf-9ceb-4929-a38b-8fb8582566a3","order_by":10,"name":"Anna Kosheleva","email":"","orcid":"","institution":"T. H. Chan School of Public Health, Harvard University","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Kosheleva","suffix":""},{"id":330806566,"identity":"b7014c2d-9048-4b80-95bf-49dc159cac95","order_by":11,"name":"Michael Snyder","email":"","orcid":"https://orcid.org/0000-0003-0784-7987","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"","lastName":"Snyder","suffix":""},{"id":330806567,"identity":"10932789-b2b0-4d24-b872-69c616e023e9","order_by":12,"name":"Dane Gobel","email":"","orcid":"","institution":"Methuselah Foundation","correspondingAuthor":false,"prefix":"","firstName":"Dane","middleName":"","lastName":"Gobel","suffix":""},{"id":330806568,"identity":"6d7a7bdb-d51a-4412-8174-221cc9e79c26","order_by":13,"name":"Chiara Herzog","email":"","orcid":"https://orcid.org/0000-0002-1572-498X","institution":"University of Innsbruck","correspondingAuthor":false,"prefix":"","firstName":"Chiara","middleName":"","lastName":"Herzog","suffix":""},{"id":330806569,"identity":"0f97a73e-af8f-4c51-9e6d-42bc8bbf9ceb","order_by":14,"name":"Jesse Poganik","email":"","orcid":"https://orcid.org/0000-0003-3098-8550","institution":"Harvard University","correspondingAuthor":false,"prefix":"","firstName":"Jesse","middleName":"","lastName":"Poganik","suffix":""},{"id":330806570,"identity":"6307630c-f44f-4e65-8631-9e8d7010c65b","order_by":15,"name":"Vadim Gladyshev","email":"","orcid":"https://orcid.org/0000-0002-0372-7016","institution":"Brigham and Women's Hospital and Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Vadim","middleName":"","lastName":"Gladyshev","suffix":""}],"badges":[],"createdAt":"2024-05-26 21:55:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4481437/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4481437/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s43587-025-00987-y","type":"published","date":"2025-11-04T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":60983677,"identity":"d054db3a-c151-47f6-a39f-caf7d14574bf","added_by":"auto","created_at":"2024-07-24 09:35:46","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":4247831,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBiolearn: A harmonization framework for aging biomarkers and biological datasets. a.\u003c/strong\u003e Overview of Biolearn’s functionalities, including biomarker discovery, dataset harmonization, and imputation. \u003cstrong\u003eb.\u003c/strong\u003e Quality control results are visualized as a ridge density plot showing the distribution of sample deviations from the mean of the population. The accompanying table lists key metadata for each dataset, including sample count, age statistics, and disease status. \u003cstrong\u003ec.\u003c/strong\u003e Comparison of actual vs. predicted sex distributions in four datasets (GSE51057, GSE42861, GSE41169, and GSE64495), with prediction accuracies listed in the accompanying table. \u003cstrong\u003ed.\u003c/strong\u003e Stacked bar graphs displaying the actual and predicted sex distributions for each dataset, with female, male and other categories shown. \u003cstrong\u003ee.\u003c/strong\u003e Epigenetic score distribution of healthy individuals compared to those with Down Syndrome (P \u0026lt; 2.2e-16, Wilcoxon rank-sum test). \u003cstrong\u003ef.\u003c/strong\u003e Python code snippet illustrating the usage of Biolearn to import, preprocess and impute missing data in a target dataset using a reference dataset. \u003cstrong\u003eg.\u003c/strong\u003e Scatter plot showing a significant inverse correlation between the Horvath epigenetic age predictions and imputed Hannum age predictions across multiple datasets (Pearson’s R = 0.999, MAE = 0.325). \u003cstrong\u003eh.\u003c/strong\u003e DNA methylation levels of three age-associated CpG sites (cg16867657, cg21572722, cg24724428) plotted against chronological age for 9 datasets, with points colored by dataset. Robust linear regressions are shown, with the correlation coefficient (R) and p-value displayed for each CpG site. \u003cstrong\u003ei,j.\u003c/strong\u003e Deconvolution results for the DeconvoluteBloodEPIC method applied to a dataset with known cell proportions from in vitro mixes (\u003cstrong\u003ei\u003c/strong\u003e) and with FACS (\u003cstrong\u003ej\u003c/strong\u003e). The estimated cell type proportions are compared to the true proportions.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/ecabda92f2294bbec8f62ae2.png"},{"id":60982936,"identity":"14293533-3c39-4cf4-81a8-dab635cf8744","added_by":"auto","created_at":"2024-07-24 09:27:46","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":6589631,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSystematic evaluation of epigenetic aging clocks across diverse datasets using Biolearn. a.\u003c/strong\u003e Benchmarking of aging clocks across eight datasets, with mean R\u003csup\u003e2\u003c/sup\u003e values indicating the average performance of each clock. The dots colored based on the dataset represent the R\u003csup\u003e2\u003c/sup\u003e measured in each individual dataset. The datasets span various ethnicities and populations, enabling a comprehensive assessment of clock robustness and generalizability. \u003cstrong\u003eb.\u003c/strong\u003e Code snippets demonstrating the streamlined curated data loading and implementation of aging clocks using the Biolearn library, highlighting its user-friendly interface and efficient data handling. \u003cstrong\u003ec.\u003c/strong\u003e Scatter plots depicting the relationship between predicted epigenetic age and chronological age for six representative clocks across different datasets represented by different colors. The plots showcase the performance of each clock, with Pearson’s R and RMSE provided. The datasets cover a wide age range, allowing for a thorough evaluation of clock accuracy and applicability across diverse age groups and biological contexts. The datasets shown include the Human Aging Rates Study (GSE40279), UKOPS (GSE19711), Dutch Schizophrenia Case-control Cohort (GSE41169), Obesity Genetics Study (GSE73103), Young Finns Study (GSE69270), and Newborns and Nonagenarians Study (GSE30870).\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/fb7447f181a5a1fbc1537c55.png"},{"id":60983678,"identity":"d20c9423-6844-4ecb-a646-b89e410eb3b1","added_by":"auto","created_at":"2024-07-24 09:35:46","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3977780,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComprehensive analysis of epigenetic clocks in predicting mortality. a-b.\u003c/strong\u003e Scatter plots illustrate the relationship between DNA methylation age and chronological age for the NAS cohort (N = 1,488, 38.8% deceased, \u003cstrong\u003ea\u003c/strong\u003e) and the MGB cohort (N = 500, 8.8% deceased, \u003cstrong\u003eb\u003c/strong\u003e). The five most representative clock models are shown for each cohort. The accuracy of each clock is indicated by Pearson’s R value, with associated P-values based on two-sided tests. \u003cstrong\u003ec-d.\u003c/strong\u003e Forest plot presenting the hazard ratios (HR) of Cox proportional hazards regression models predicting mortality risk for the NAS cohort (\u003cstrong\u003ec\u003c/strong\u003e) and MGB cohort (\u003cstrong\u003ed\u003c/strong\u003e), as predicted by 14 epigenetic clock models, adjusted for age. The hazard ratios are shown per standard deviation increase of the biomarker, with 95% confidence intervals displayed as error bars. P-values are based on two-sided tests. The insignificant results (P \u0026gt; 0.05) are colored gray. \u003cstrong\u003ee.\u003c/strong\u003e Scatter plot comparing the Pearson’s correlation coefficient with chronological age (x-axis) and the hazard ratio of mortality risk (y-axis) for each aging clock in NAS and MGB cohort. The points are colored based on the cohorts. The regression line and 95% confidence intervals are shown. Pearson’s R-value and P-value are shown at the top of the plot.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/ace917ddd06744e0195a4047.png"},{"id":60982941,"identity":"d3f78ab1-244b-499d-a9a3-625f08322411","added_by":"auto","created_at":"2024-07-24 09:27:46","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1848854,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComprehensive analysis of aging biomarkers in predicting clinical outcomes. a.\u003c/strong\u003e Bar plot displaying the number of people with events across six categories (Stroke, Dementia, Operation, Lifespan, Cancer, and Healthspan) in the NAS cohort. \u003cstrong\u003eb-f.\u003c/strong\u003e Forest plots presenting the hazard ratios (HR) per standard deviation increase for various aging biomarkers in predicting specific clinical outcomes: Healthspan (b), Operation (c), Cancer (d), Dementia (e), Stroke (f), and individual event types (g). The 95% confidence intervals are displayed as error bars, and P-values are based on two-sided tests. \u003cstrong\u003eg.\u003c/strong\u003e Correlation matrix showcasing the associations between different aging biomarkers in the NAS cohort, after adjusting for age. The color intensity represents the strength of the correlation, with red indicating positive correlations and blue indicating negative correlations. The significant pairs are annotated with black dots. \u003cstrong\u003eh.\u003c/strong\u003e Scatter plot comparing the Healthspan hazard ratios (log scale) against the log2-transformed lifespan hazard ratios for various aging biomarkers. The plot reveals the relationship between the predictive power of biomarkers for healthspan and lifespan. The size of the points represents the significance of the P-values, with larger points indicating higher significance. The inverse variance weighted correlation and P-values are shown.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/a37720dd9fa18fd0e49d8ac0.png"},{"id":60982939,"identity":"4310364b-3908-4f4f-9fcb-63f652c8fedf","added_by":"auto","created_at":"2024-07-24 09:27:46","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1732730,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTranscriptomic and clinical biomarkers. a.\u003c/strong\u003e Scatter plots of chronological age vs. predicted age for various transcriptomic clocks (tAge) on JenAge RNA-Seq dataset (Jena Centre for Systems Biology of Ageing, Illumina TruSeq 2.0, Whole Blood, N=62, GSE103232 \u0026amp; GSE75337). Pearson’s R values are shown at the top of the plot. \u003cstrong\u003eb.\u003c/strong\u003e Overview of Biolearn’s transcriptomic and phenotypic biomarker functionalities. The code snippet shows that transcriptomic age predictors like Peters (tAge) can be calculated with a few lines of code using the Biolearn library. \u003cstrong\u003ec.\u003c/strong\u003eScatter plot showing the correlation between Phenotypic Age prediction (X-axis) and chronological age (Y-axis). The dots are colored by AgeDev. \u003cstrong\u003ed.\u003c/strong\u003eSurvival analysis of the NHANES 2010 dataset (N = 2877), stratified by biological age discrepancies (marked by different colors) based on Phenotypic Age. The survival curves demonstrate that individuals with accelerated biological aging (red) have a lower survival probability compared to those with decelerated aging (blue). The shaded areas in c represent the standard error of the survival estimates. \u003cstrong\u003ee.\u003c/strong\u003e Forest plot displaying the hazard ratios (HR) and 95% confidence intervals (CI) for the association between biological age discrepancies based on the Phenotypic Age metrics and all-cause mortality.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/42ad397e65fc336dd89f315a.png"},{"id":60982938,"identity":"aba8d8b6-6f98-4732-aa28-9f8af468eaa2","added_by":"auto","created_at":"2024-07-24 09:27:46","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1330094,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-4481437/v1/85b1f39cdef3eca46f8dc517.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"A Unified Framework for Systematic Curation and Evaluation of Aging Biomarkers","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDevelopment and validation of robust biomarkers of aging (BoAs) have become key focal points in aging research, driven by the growing recognition of aging as a fundamental driver of chronic diseases and mortality. Numerous biomarkers have been proposed to quantify biological age and elucidate the biological processes underlying aging. However, clinical validation of BoAs remains a significant challenge due to heterogeneity in their formulations and disparate structures of validation datasets across populations \u003csup\u003e1,2\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSince the introduction of composite omic biomarkers of aging, exemplified by Horvath\u0026rsquo;s pioneering work on DNA methylation aging clocks \u003csup\u003e3\u003c/sup\u003e, the field has witnessed a rapid expansion in the repertoire of aging biomarkers. These biomarkers now span a wide array of omic modalities, including epigenomics, transcriptomics, and proteomics \u003csup\u003e4\u0026ndash;9\u003c/sup\u003e. Omic biomarkers provide a comprehensive view of the molecular changes associated with aging, offering valuable insights into the aging process and its impact on human health.\u003c/p\u003e \u003cp\u003eAmong the various classes of omic biomarkers, DNA methylation-based clocks are currently the most advanced and robust tools for estimating biological age. These human clocks, such as the Horvath multi-tissue clock, DunedinPACE \u003csup\u003e5\u003c/sup\u003e, GrimAge \u003csup\u003e10\u003c/sup\u003e, PhenoAge \u003csup\u003e11\u003c/sup\u003e, causality-enriched DamAge/AdaptAge \u003csup\u003e12\u003c/sup\u003e, and the PRC2 clock \u003csup\u003e13\u003c/sup\u003e, have demonstrated significant associations with age-related conditions and mortality, highlighting the intricate relationship between epigenetic modifications and aging trajectories \u003csup\u003e6,10,14,15\u003c/sup\u003e. However, the diverse formulations of these biomarkers and inconsistencies in dataset structures across different populations pose substantial challenges for their systematic cross-population validation and benchmarking, which are crucial steps toward their clinical translation.\u003c/p\u003e \u003cp\u003ePublicly available datasets, such as those from the Gene Expression Omnibus (GEO) \u003csup\u003e16\u003c/sup\u003e, the National Health and Nutrition Examination Survey (NHANES), and the Framingham Heart Study (FHS), hold immense potential for accelerating the validation of BoAs. However, the lack of a standardized framework that can accommodate the heterogeneous nature of these datasets hinders their effective utilization for this purpose. There is a pressing need for a unified platform that can seamlessly integrate and analyze various BoAs across datasets with harmonized structures. Such a platform would revolutionize the validation process, facilitate the discovery of novel biomarkers, and provide a structured avenue for community-driven efforts in advancing the field of aging biology.\u003c/p\u003e \u003cp\u003eTo address this need, we developed Biolearn, an open-source Python library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Biolearn supports biomarkers based on multiple different biological data modalities and serves as an innovative tool that harmonizes existing BoAs, structures and formats human datasets and offers computational methodologies for assessing biomarkers against these datasets. By enabling the integration and analysis of diverse BoAs and datasets, Biolearn aims to accelerate the development and validation of BoAs, fostering a community-driven approach to aging research.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eHarmonization of Biomarkers of Aging and Datasets\u003c/h2\u003e \u003cp\u003eWe harmonized a comprehensive set of 39 well-established epigenetic, transcriptomic, and clinical biomarkers (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and implemented these BoAs in Biolearn, representing the largest collection of BoAs in a single package to date. We have validated the implementation of these biomarkers with their respective developers to ensure accuracy and reliability. The epigenetic biomarkers encompass a wide range of categories, including: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Chronological clocks: Horvath\u0026rsquo;s multi-tissue clock and Hannum\u0026rsquo;s blood clock \u003csup\u003e3,17\u003c/sup\u003e; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) Healthspan and mortality-related clocks: GrimAge, GrimAge2, PhenoAge, and Zhang clock \u003csup\u003e10,11,18,19\u003c/sup\u003e; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) Biomarkers of the rate of aging: DunedinPoAm38 and DunedinPACE \u003csup\u003e5,20\u003c/sup\u003e; (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) Causality-enriched clocks: Ying\u0026rsquo;s CausAge, DamAge, and AdaptAge \u003csup\u003e12\u003c/sup\u003e; and (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) Various other clocks, including DNAm-based biomarkers and disease predictors, transcriptomic clocks, and clinical clocks (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHarmonized biomarkers in Biolearn.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiomarker\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTissue\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePredicts\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOmic Type\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHorvathV1 \u003csup\u003e3\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMulti-tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHannum \u003csup\u003e17\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLin \u003csup\u003e39\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePhenoAge \u003csup\u003e11\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHorvathV2 \u003csup\u003e40\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSkin, blood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEDBE \u003csup\u003e41\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBuccal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZhang_10 \u003csup\u003e19\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMortality Risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrimAge \u003csup\u003e10\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge Adjusted by Mortality Risk (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrimAge2 \u003csup\u003e18\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge Adjusted by Mortality Risk (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDunedinPoAm38 \u003csup\u003e20\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAging Rate (Years/Year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDunedinPACE \u003csup\u003e5\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAging Rate (Years/Year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDNAmTL \u003csup\u003e42\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood, Adipose\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTelomere Length\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKnight \u003csup\u003e43\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCord Blood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGestational Age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeeControl \u003csup\u003e44\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePlacenta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGestational Age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeeRefinedRobust \u003csup\u003e44\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePlacenta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGestational Age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeeRobust \u003csup\u003e44\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePlacenta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGestational Age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYingCausAge \u003csup\u003e12\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYingDamAge \u003csup\u003e12\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYingAdaptAge \u003csup\u003e12\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmokingMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSmoking Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlcoholMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAlcohol Consumption\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI_McCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducationMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEducational Attainment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotalCholesterolMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTotal Cholesterol\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHDLCholesterolMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHDL Cholesterol\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDLCholesterolMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLDL with Remnant Cholesterol\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBodyFatMcCartney \u003csup\u003e45\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePercentage Body Fat\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI_Reed \u003csup\u003e46\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProstateCancerKirby \u003csup\u003e47\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eProstate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eProstate Cancer Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHepatoXu \u003csup\u003e48\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCirculating DNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHepatocellular Carcinoma Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCVD_Westerman \u003csup\u003e49\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCoronary Heart Disease Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAD_Bahado-Singh \u003csup\u003e50\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAlzheimer\u0026rsquo;s Disease Status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepressionBarbu \u003csup\u003e51\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDepression Risk\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDNA Methylation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePhenotypic Age \u003csup\u003e11\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePhenotypic Age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eClinical Biomarker\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMahalanobis Distance \u003csup\u003e34\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMahalanobis Distance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eClinical Biomarker\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeters_tAge \u003csup\u003e30\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAge (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRNA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMultispecies-blood_tAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRelative Age (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRNA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman-blood_tAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBlood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRelative Age (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRNA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman-multitissue_tAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMulti-tissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRelative Age (Years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRNA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe harmonization of these diverse biomarkers is crucial for enabling consistent and reproducible analyses across different datasets, facilitating cross-population validation studies, and advancing our understanding of the aging process. To achieve this, all biomarkers were formatted into standardized input structures, ensuring their consistent application across disparate datasets. The harmonization process involved collecting and unifying the annotation of clock specifications, such as tissue type, predicted age range, and source references. This meticulous approach guarantees transparent and reproducible analyses, enabling researchers to readily compare and interpret results across different studies and populations.\u003c/p\u003e \u003cp\u003eTo further support ongoing research in this field, we developed and implemented an open-source framework. This framework provides a standardized format for epigenetic biomarkers, facilitating the seamless integration and comparison of any future aging clocks and biomarkers that are developed. By providing a unified platform for the harmonization and analysis of aging biomarkers, this framework aims to foster collaboration and innovation. Researchers can easily contribute new biomarkers, compare their performance against existing ones, and explore their potential applications in various datasets. This collaborative approach is essential for accelerating progress in the field and developing more accurate and robust biomarkers of aging.\u003c/p\u003e \u003cp\u003eTo facilitate cross-population validation studies using publicly available data, we harnessed Biolearn\u0026rsquo;s capabilities to integrate and structure multiple public datasets (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The structured datasets were refined to enable a shared analysis platform, addressing the challenges of data heterogeneity and formatting inconsistencies \u003csup\u003e21,22\u003c/sup\u003e. With this capacity, Biolearn is used as the backend of ClockBase for epigenetic age computation \u003csup\u003e22\u003c/sup\u003e, enabling the systemic harmonization of over 200,000 human samples from Gene Expression Omnibus (GEO) array data.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHarmonized datasets in Biolearn.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTitle\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFormat\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSamples\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAge Present\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSex Present\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE40279\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGenome-wide Methylation Profiles Reveal Quantitative Views o\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e656\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE19711\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGenome wide DNA methylation profiling of United Kingdom Ovar\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina27k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e540\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE51057\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMethylome Analysis and Epigenetic Changes Associated with Me\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e329\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE42861\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDifferential DNA methylation in Rheumatoid arthritis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE41169\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBlood DNA methylation profiles in a Dutch population\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE51032\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEPIC-Italy at HuGeF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e845\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE73103\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMany obesity-associated SNPs strongly associate with DNA met\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e355\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE69270\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAging-associated DNA methylation changes in middle-aged indi\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e184\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE36054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMethylation Profiling of Blood DNA from Healthy Children\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e192\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE64495\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDNA methylation profiles of human blood samples from a sever\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE30870\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDNA methylomes of Newborns and Nonagenarians\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE52588\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIdentification of a DNA methylation signature in blood from \u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIllumina450k\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE157131\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMethylation data from stored peripheral blood leukocytes fro\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIlluminaEPIC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e946\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE132203\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDNA Methylation (EPIC) from the Grady Trauma Project\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIlluminaEPIC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e795\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE134080\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRNASeq whole blood of Dutch 500FG cohort\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIlluminaHiSeq2500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNHANES\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNational Health and Nutrition Examination Survey\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePhenotypic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2877\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFHS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFramingham Heart Study\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePhenotypic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4434\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eQuality Control, Imputation, and Deconvolution\u003c/h2\u003e \u003cp\u003eBiolearn provides a comprehensive toolkit for data preprocessing, normalization, and cell-type deconvolution (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Quality control metrics, such as sample deviation from the population mean, missingness, and the number of sites with a high percentage of missingness, can be readily visualized (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). This functionality enables researchers to identify potential outliers or problematic samples and make informed decisions about data inclusion and exclusion criteria.\u003c/p\u003e \u003cp\u003eMoreover, Biolearn enables the prediction of sample sex from DNA methylation data with high accuracy, which can be compared against actual sex distributions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec,d)\u003csup\u003e23\u003c/sup\u003e. This feature is particularly useful for identifying potential sample mislabeling or investigating sex-specific effects in aging research. We also implemented predictors of common traits, including smoking, BMI, and epigenetic scores for diseases like Down Syndrome (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee)\u003csup\u003e24\u003c/sup\u003e. These predictors allow researchers to explore the associations between aging biomarkers and various lifestyle factors or disease conditions, providing valuable insights into the complex interplay between aging and health.\u003c/p\u003e \u003cp\u003eMissing DNA methylation data can be easily imputed with different methods in just a few lines of code (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ef,g) \u003csup\u003e25\u003c/sup\u003e, ensuring that researchers can make the most of the available data and minimize the impact of missing values on their analyses. Biolearn also facilitates the integration of multiple datasets for large-scale analyses, such as the comparison of DNA methylation levels of CpG sites across different datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eh). This feature allows researchers to investigate the consistency and reproducibility of aging biomarkers across diverse populations and experimental settings, strengthening the robustness and generalizability of their findings.\u003c/p\u003e \u003cp\u003eIn addition to these features, Biolearn offers a deconvolution tool that estimates the proportion of cell types in a given sample based on a single bulk-level methylation measurement. Biolearn provides two modes for deconvolution, optimized for the 450K (DeconvoluteBlood450K) and EPIC (DeconvoluteBloodEPIC) methylation platforms, respectively \u003csup\u003e26\u003c/sup\u003e. These modes are designed for estimating cell proportions in blood methylation samples and account for technology-specific biases that can affect the accuracy of deconvolution \u003csup\u003e27\u003c/sup\u003e. The reference methylation matrices for each mode consist of methylation profiles for six cell types representing the most abundant cell types found in the blood: neutrophils, monocytes, natural killer cells, B cells, CD4\u0026thinsp;+\u0026thinsp;T cells, and CD8\u0026thinsp;+\u0026thinsp;T cells \u003csup\u003e28,29\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe benchmarked the accuracy of our deconvolution tool using datasets with known cell proportions assessed via fluorescence-activated cell sorting (FACS) and \u003cem\u003ein vitro\u003c/em\u003e cell mixing \u003csup\u003e29\u003c/sup\u003e. Both deconvolution methods generated accurate predictions that matched known cell proportions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ei,j). These results demonstrate the reliability and utility of Biolearn\u0026rsquo;s deconvolution feature, which can help researchers account for cellular heterogeneity in their analyses and gain insights into the cell type-specific contributions to aging biomarkers.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eSystematic Evaluation of the Biomarkers of Aging\u003c/h2\u003e \u003cp\u003eTo demonstrate the utility of Biolearn in facilitating the systematic evaluation of aging biomarkers, we conducted a comprehensive benchmarking analysis of various epigenetic aging clocks across multiple datasets. By leveraging Biolearn\u0026rsquo;s harmonized dataset library and standardized clock implementations, we assessed the performance, robustness, and generalizability of these clocks in diverse biological contexts and populations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Our analysis included a wide range of datasets, spanning the Human Aging Rates Study (GSE40279, N\u0026thinsp;=\u0026thinsp;656), EPIC-Italy (GSE51032, N\u0026thinsp;=\u0026thinsp;845), RA Case-control Cohort (GSE42861, N\u0026thinsp;=\u0026thinsp;689), Dutch Schizophrenia Case-control Cohort (GSE41169, N\u0026thinsp;=\u0026thinsp;95), Obesity Genetics Study (GSE73103, N\u0026thinsp;=\u0026thinsp;355), Developmental Disorder Study (GSE64495, N\u0026thinsp;=\u0026thinsp;113), African American GENOA (GSE157131, N\u0026thinsp;=\u0026thinsp;1218), and Grady Trauma Project (GSE132203, N\u0026thinsp;=\u0026thinsp;795) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). These diverse datasets allowed us to evaluate the performance of the biomarkers across various age ranges, ethnicities, and disease states. Biolearn\u0026rsquo;s user-friendly interface and efficient data handling capabilities streamlined data loading and implementation of aging clocks (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). This highlights the library\u0026rsquo;s potential to accelerate the development and validation of novel aging biomarkers by providing a standardized framework for their evaluation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe benchmarking results (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-c) revealed that the HorvathV2 clock (i.e., skin and blood clock) exhibited the highest overall accuracy in terms of predicting chronological age, with a mean R\u003csup\u003e2\u003c/sup\u003e of 0.88 across all datasets, followed closely by the Hannum clock (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.81), the Horvath1 clock (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.78) and YingCausAge clock (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.77). These findings suggest that these four clocks are the most robust and generalizable across diverse biological contexts and age-related conditions. Note that the GrimAgeV1 and GrimAgeV2 clocks use the age of the sample as the predictor, therefore they cannot be compared directly to the other clocks. These results highlight the applicability of these clocks across diverse age ranges and biological contexts, further emphasizing the importance of systematic evaluation in identifying the most suitable biomarkers for specific research questions or clinical applications.\u003c/p\u003e \u003cp\u003eOverall, our findings demonstrate the value of Biolearn in enabling the systematic evaluation of epigenetic aging clocks across multiple datasets. By providing a standardized framework for clock implementation and evaluation, Biolearn facilitates the identification of robust and generalizable aging biomarkers, paving the way for their translation into clinical settings and advancing our understanding of the aging process.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eMortality and Morbidity Risk Analysis\u003c/h2\u003e \u003cp\u003eIt is also important to evaluate the predictive power of epigenetic aging biomarkers in predicting aging-associated outcomes such as mortality risk. Here, we conducted the most comprehensive evaluation of 17 representative epigenetic clock models to the Normative Aging Study (NAS) dataset (N\u0026thinsp;=\u0026thinsp;1,488, 38.8% deceased) and the Massachusetts General Brigham (MGB) cohort (N\u0026thinsp;=\u0026thinsp;500, 8.8% deceased), comparing their performance in predicting mortality risk (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). The biomarkers showed a strong correlation with chronological age in both NAS (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea) and MGB cohorts (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe then examined their performance in predicting mortality risk using Cox Proportional Hazards analysis, adjusted for age and sex (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, d). In the NAS cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec), the top-performing clock in terms of hazard ratio (per standard deviation increase of the biomarker) was DunedinPoAm38 (HR\u0026thinsp;=\u0026thinsp;1.38, P\u0026thinsp;=\u0026thinsp;9.48e-18), followed closely by GrimAgeV2 (HR\u0026thinsp;=\u0026thinsp;1.35, P\u0026thinsp;=\u0026thinsp;1.21e-18) and DunedinPACE (HR\u0026thinsp;=\u0026thinsp;1.35, P\u0026thinsp;=\u0026thinsp;3.04e-19). Other clocks with strong predictive value include Zhang_10 (HR\u0026thinsp;=\u0026thinsp;1.28, P\u0026thinsp;=\u0026thinsp;9.66e-12), GrimAgeV1 (HR\u0026thinsp;=\u0026thinsp;1.29, P\u0026thinsp;=\u0026thinsp;7.05e-14), YingDamAge (HR\u0026thinsp;=\u0026thinsp;1.25, P\u0026thinsp;=\u0026thinsp;1.48e-11), PhenoAge (HR\u0026thinsp;=\u0026thinsp;1.20, P\u0026thinsp;=\u0026thinsp;8.96e-08), Hannum (HR\u0026thinsp;=\u0026thinsp;1.13, P\u0026thinsp;=\u0026thinsp;3.91e-04), and YingCausAge (HR\u0026thinsp;=\u0026thinsp;1.09, P\u0026thinsp;=\u0026thinsp;7.37e-03). In the MGB cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed), the top-performing clocks in predicting mortality risk were GrimAgeV2 (HR\u0026thinsp;=\u0026thinsp;2.08, P\u0026thinsp;=\u0026thinsp;2.64e-03), PhenoAge (HR\u0026thinsp;=\u0026thinsp;2.03, P\u0026thinsp;=\u0026thinsp;2.52e-02), and GrimAgeV1 (HR\u0026thinsp;=\u0026thinsp;1.84, P\u0026thinsp;=\u0026thinsp;1.46e-02). Followed by Zhang_10 (HR\u0026thinsp;=\u0026thinsp;1.48, P\u0026thinsp;=\u0026thinsp;7.96e-05), and DunedinPACE (HR\u0026thinsp;=\u0026thinsp;1.46, P\u0026thinsp;=\u0026thinsp;7.40e-04).\u003c/p\u003e \u003cp\u003eWe further assessed the association between the predictive power of the epigenetic clocks for chronological age and mortality risk. Interestingly, in both cohorts, we observed a negative but insignificant correlation between Pearson\u0026rsquo;s R with chronological age and hazard ratio of mortality risk (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee). This suggests that the predictive power of epigenetic clocks on mortality risk, after adjusting for age and sex, is independent of their ability to predict chronological age, highlighting the importance of interpreting the meaning of age deviation (AgeDev) with caution for aging biomarkers. We also observed strong heterogeneous associations of epigenetic clocks with mortality risk in different cohorts. The analysis demonstrates the utility of Biolearn in facilitating systematic evaluation of epigenetic clocks in predicting mortality across multiple cohorts, emphasizing the importance of systematic evaluation in identifying the most suitable biomarkers for specific applications.\u003c/p\u003e \u003cp\u003eBesides mortality, it is also important to evaluate aging biomarkers in predicting various other clinically relevant aging outcomes. We analyzed the associations between 14 aging biomarkers and six event categories (Stroke, Dementia, Operation, Lifespan, Cancer, and Healthspan, which is defined by the first incidence of any event) in the NAS cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). To assess the predictive power of these biomarkers, we performed Cox Proportional Hazards analyses, adjusted for age, and calculated the hazard ratios (HR) per standard deviation increase for each biomarker (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb-f).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAcross five clinical outcomes tested, DunedinPACE was the strongest predictor for three of the outcomes, namely healthspan (HR\u0026thinsp;=\u0026thinsp;1.18), dementia (HR\u0026thinsp;=\u0026thinsp;1.40), and stroke (HR\u0026thinsp;=\u0026thinsp;1.50). PhenoAge was the strongest predictor for surgery (HR\u0026thinsp;=\u0026thinsp;1.24), and HorvathV2 was the strongest predictor for cancer (HR\u0026thinsp;=\u0026thinsp;1.12). These results suggest considerable heterogeneity of aging biomarkers in predicting different clinical outcomes.\u003c/p\u003e \u003cp\u003eWe further investigated the associations between the AgeDev term of aging biomarkers after adjusting for age in the NAS cohort and found strong positive correlations among most epigenetic clocks (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eg). We observed two main clusters: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) PhenoAge, GrimAgeV1, GrimAgeV2, YingCausAge, HorvathV1, Lin, Hannum, and HorvathV2; and (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) YingDamAge, DunedinPoAm38, Zhang10, and DunedinPACE. The DNAmTL telomere length clock and YingAdaptAge do not cluster with other clocks, suggesting their unique biological underpinnings.\u003c/p\u003e \u003cp\u003eLastly, we compared the predictive power of aging biomarkers for healthspan and lifespan (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eh). In general, we observed a significant positive correlation using an inverse-variance-weighted approach (weighted correlation coefficient\u0026thinsp;=\u0026thinsp;0.58, P\u0026thinsp;=\u0026thinsp;0.012). DunedinPoAm38, GrimAgeV2, GrimAgeV1, and Zhang_10 demonstrated strong associations with healthspan and lifespan, indicating their potential as comprehensive aging biomarkers. These findings highlight the utility of Biolearn in facilitating the systematic evaluation of aging biomarkers for predicting various health outcomes. The results provide insights into the comparative performance of these biomarkers and their potential applications in clinical settings and aging research.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eTranscriptomic and Clinical Biomarkers\u003c/h2\u003e \u003cp\u003eTo demonstrate Biolearn\u0026rsquo;s multi-omic capabilities, we also evaluated the performance of transcriptomic and phenotypic aging biomarkers. We include four transcriptomic age predictors: tAge.Peters \u003csup\u003e30\u003c/sup\u003e, tAge.Multispecies.Blood, tAge.Human.Multi-tissue, and tAge.Human.Blood \u003csup\u003e31,32\u003c/sup\u003e. We applied these predictors to the JenAge RNA-Seq dataset (Jena Centre for Systems Biology of Ageing, Illumina TruSeq 2.0, Whole Blood, N\u0026thinsp;=\u0026thinsp;62) \u003csup\u003e33\u003c/sup\u003e. The tAge predictor showed a strong correlation with chronological age, with Pearson\u0026rsquo;s R ranging from 0.68 (Peters) to 0.90 (Human.Multi-tissue) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea), highlighting their potential as robust aging biomarkers. All these predictors are implemented in Biolearn with simple and easy-to-use functions (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, we investigated the performance of the Phenotypic Age predictor \u003csup\u003e11\u003c/sup\u003e, a blood-test-based biomarker, using the NHANES 2010 dataset. We calculated the Phenotypic Age for each individual using Biolearn\u0026rsquo;s implementation of the predictor and compared it to chronological age (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). We found a strong linear relationship between Phenotypic Age and chronological age (Pearson\u0026rsquo;s R\u0026thinsp;=\u0026thinsp;0.96, P\u0026thinsp;\u0026lt;\u0026thinsp;2.2e-16), indicating the predictor\u0026rsquo;s ability to capture age-related changes in clinical biomarkers.\u003c/p\u003e \u003cp\u003eTo assess the predictive power of Phenotypic Age on mortality risk, we performed a survival analysis on the NHANES 2010 dataset. Individuals were stratified into five groups based on their AgeDev (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed,e). We found that individuals with higher AgeDev had a significantly higher mortality risk (HR\u0026thinsp;=\u0026thinsp;1.58, 95% CI: 1.31\u0026ndash;1.91, P\u0026thinsp;=\u0026thinsp;1.08e-06), while those with lower AgeDev had a lower risk (HR\u0026thinsp;=\u0026thinsp;0.60, 95% CI: 0.49\u0026ndash;0.74, P\u0026thinsp;=\u0026thinsp;6.19e-07). These results underscore the predictive power of Phenotypic Age in assessing mortality risk and demonstrate Biolearn\u0026rsquo;s capability to facilitate such analyses.\u003c/p\u003e \u003cp\u003eTo investigate the association between aging and cell type proportions, we performed deconvolution analysis on the NAS cohort blood samples. This analysis revealed distinct changes for different cell types over the life course (Extended Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Neutrophil and natural killer cell proportions showed significant positive correlations with age (R\u0026thinsp;=\u0026thinsp;0.09, P\u0026thinsp;=\u0026thinsp;5.64e-04 and R\u0026thinsp;=\u0026thinsp;0.14, P\u0026thinsp;=\u0026thinsp;1.03e-07, respectively). In contrast, B cell, CD4 T cell, and CD8 T cell proportions exhibited significant negative correlations with age (R=-0.09, P\u0026thinsp;=\u0026thinsp;2.45e-04; R=-0.15, P\u0026thinsp;=\u0026thinsp;6.07e-09; and R=-0.05, P\u0026thinsp;=\u0026thinsp;4.01e-02, respectively). Monocyte proportion did not show a significant correlation with age (R\u0026thinsp;=\u0026thinsp;0.02, P\u0026thinsp;=\u0026thinsp;4.07e-01).\u003c/p\u003e \u003cp\u003eWe assessed the predictive power of cell type proportions on various health outcomes in the NAS cohort (Extended Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). For healthspan, natural killer cell proportion was a significant protective factor (HR\u0026thinsp;=\u0026thinsp;0.93, 95% CI: 0.88\u0026ndash;0.99, P\u0026thinsp;=\u0026thinsp;1.55e-02), while CD8 T cell proportion was a slight risk factor (HR\u0026thinsp;=\u0026thinsp;1.06, 95% CI: 1.00-1.12, P\u0026thinsp;=\u0026thinsp;4.65e-02). Similarly, for lifespan, natural killer cell proportion was a significant protective factor (HR\u0026thinsp;=\u0026thinsp;0.91, 95% CI: 0.85\u0026ndash;0.98, P\u0026thinsp;=\u0026thinsp;1.03e-02), and CD8 T cell proportion was a significant risk factor (HR\u0026thinsp;=\u0026thinsp;1.07, 95% CI: 1.01\u0026ndash;1.14, P\u0026thinsp;=\u0026thinsp;2.59e-02). Natural killer cell proportion was also a significant protective factor for dementia (HR\u0026thinsp;=\u0026thinsp;0.87, 95% CI: 0.77\u0026ndash;0.98, P\u0026thinsp;=\u0026thinsp;2.03e-02). For stroke risk, neutrophil proportion was a significant risk factor (HR\u0026thinsp;=\u0026thinsp;1.33, 95% CI: 1.11\u0026ndash;1.60, P\u0026thinsp;=\u0026thinsp;2.18e-03), while natural killer cell proportion was a significant protective factor (HR\u0026thinsp;=\u0026thinsp;0.70, 95% CI: 0.55\u0026ndash;0.89, P\u0026thinsp;=\u0026thinsp;4.21e-03). Similarly, for surgical events, neutrophil proportion was a significant risk factor (HR\u0026thinsp;=\u0026thinsp;1.16, 95% CI: 1.05\u0026ndash;1.28, P\u0026thinsp;=\u0026thinsp;3.99e-03), and natural killer cell proportion was a significant protective factor (HR\u0026thinsp;=\u0026thinsp;0.89, 95% CI: 0.80\u0026ndash;0.99, P\u0026thinsp;=\u0026thinsp;2.58e-02). No significant associations were found between cell type proportions and cancer risk.\u003c/p\u003e \u003cp\u003eThe integration of transcriptomic and phenotypic biomarkers in Biolearn enables researchers to investigate aging processes from different biological perspectives. The strong performance of the tAge predictor and Phenotypic Age in their respective datasets showcases the potential of multi-omic approaches in uncovering the complex mechanisms underlying aging. By leveraging Biolearn\u0026rsquo;s comprehensive framework, researchers can gain valuable insights into the interplay between different biological layers and their contributions to the aging process, ultimately facilitating the development of targeted interventions and personalized aging management strategies.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eAmong the most significant challenges in aging biomarker research is cross-population validation of proposed biomarkers \u003csup\u003e34\u003c/sup\u003e. To take steps to address this need and provide an open-source tool for validation efforts across the field, we built Biolearn, an open-source library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers across diverse datasets. By leveraging Biolearn, we conducted a comprehensive benchmarking analysis of various epigenetic aging clocks, transcriptomic predictors, and phenotypic biomarkers, demonstrating their performance, robustness, and generalizability in different biological contexts and populations.\u003c/p\u003e \u003cp\u003eSystematic evaluation of epigenetic aging clocks across multiple datasets revealed that the Horvath skin and blood clock, Hannum clock, Horvath multi-tissue clock and Ying CausAge clock exhibited the highest overall accuracy in predicting chronological age. Notably, the predictive power of epigenetic clocks on mortality risk, after adjusting for age and sex, was independent of their ability to predict chronological age, highlighting the importance of interpreting AgeDev with caution and underscoring the need for further investigation into the biological mechanisms captured by these clocks. Evaluation of aging biomarkers in predicting various clinical outcomes in the NAS cohort revealed considerable heterogeneity, with different biomarkers showing strengths in predicting specific outcomes. For example, DunedinPACE was the strongest predictor for healthspan, dementia, and stroke, while PhenoAge and the Horvath skin and blood clock were the strongest predictors for surgery and cancer, respectively. These findings underscore the importance of selecting appropriate biomarkers based on the specific clinical outcomes of interest and suggest the potential for developing targeted interventions tailored to individual aging trajectories.\u003c/p\u003e \u003cp\u003eIntegration of transcriptomic and phenotypic biomarkers in Biolearn enables a multi-omic approach to investigating the aging process. The strong performance of the tAge predictor and Phenotypic Age in their respective datasets highlights the potential of combining information from different biological layers to gain a more comprehensive understanding of the aging process. By leveraging Biolearn\u0026rsquo;s unified framework, researchers can explore the interplay between various omic modalities and uncover novel insights into the complex mechanisms underlying aging.\u003c/p\u003e \u003cp\u003eWith Biolearn, we also harmonized and evaluated several well-established aging clocks, providing the opportunity for these biomarkers to be refined and potentially for new ones to be developed. The modular design of Biolearn encourages the addition of new models and datasets, making it a living library that will grow in tandem with the field itself. By centralizing resources and knowledge, Biolearn considerably reduces redundancy and accelerates biomarker development and validation efforts \u003csup\u003e5,9\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOur approach emphasizes transparency and reproducibility, core tenets of open science. By making Biolearn publicly available and maintaining detailed documentation and development guidelines, we established an ecosystem that supports open collaboration and knowledge sharing. This open-science framework ensures that findings and tools can be widely accessed, providing equitable opportunities for researchers globally to contribute to and benefit from the collective advances in aging research. Moreover, our hope is that the open-access nature of Biolearn will promote cross-fertilization between aging researchers and scientists currently outside the field, incentivizing the development of novel and innovative biomarker models and validation approaches.\u003c/p\u003e \u003cp\u003ePrevious efforts to harmonize biomarkers of aging, notably methylCIPHER and BioAge \u003csup\u003e35,36\u003c/sup\u003e, have been limited in scope, focusing on methylation or blood-based biomarkers only. Furthermore, using R packages is somewhat limiting. Biolearn supports biomarkers based on multiple different biological data modalities and is written in Python, which has a broader reach. In comparison to PyAging \u003csup\u003e37\u003c/sup\u003e, a preliminary contemporaneous Python biomarker library, Biolearn is focused on ease of use and reproducibility through automated testing against reference data. Biolearn also offers several distinct advantages over existing biomarker libraries: it supports the easy loading of a larger and more diverse set of data, enabling researchers to work with a wide range of datasets and explore the performance of biomarkers across various populations and biological contexts; it includes models that are not directly used for age prediction but may be relevant to health and lifespan, providing a more comprehensive toolkit for aging research; and it includes tooling for exploring and understanding your data, such as quality reports, which facilitate data preprocessing and ensure the reliability of the results.\u003c/p\u003e \u003cp\u003eWhile Biolearn represents a significant advance for the field, some limitations remain. Currently, the library is tailored to biomarkers derived from biological samples, predominantly DNA methylation data. Moving forward, the scope of Biolearn will continue to expand to encompass diverse biological modalities\u0026mdash;such as proteomics, metabolomics, and microbiomics\u0026mdash;broadening its applicability \u003csup\u003e7,13\u003c/sup\u003e. Moreover, integration with larger and more diverse population datasets will be vital in advancing cross-population validation efforts. As new datasets emerge, Biolearn will adapt to incorporate these resources, ensuring ongoing robustness and scalability \u003csup\u003e38\u003c/sup\u003e. Finally, bioinformatics tools, including Biolearn, depend on a user base proficient in programming and data analysis. Efforts to make these tools more accessible to a wider audience, including those with limited computational expertise, will be crucial. This could involve the development of graphical user interfaces (GUIs) or web-based platforms to streamline the user experience.\u003c/p\u003e \u003cp\u003eWe anticipate that Biolearn will become a key resource for the field and will transform many facets of aging biomarker studies. Our preliminary survival studies conducted using Biolearn demonstrate not only the power of this new platform but illuminate the real-world implications of validated biomarkers. Biolearn\u0026rsquo;s standardization and analysis capabilities stand to serve as pivotal tools for researchers seeking to bridge the gap between biomarker discovery and clinical implementation \u003csup\u003e34\u003c/sup\u003e.\u003c/p\u003e "},{"header":"Methods","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003eOverview of Biolearn Library\u003c/h2\u003e \u003cp\u003eBiolearn is an open-source computational suite that facilitates the harmonization and analysis of biomarkers of aging (BoAs). It is written in Python and is readily accessible through the Python Package Index (PyPI). Biolearn is developed using modern software engineering practices, including automated testing to ensure correctness and adherence to software design principles that ensure the safe interchangeability of like components. The library is designed to be user-friendly while offering robust functionalities for researchers across various disciplines involved in aging studies. Additionally, the results of each model in Biolearn are portable and can be easily exported in formats such as CSV, allowing for seamless integration with other tools like R for further analysis and visualization.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSystem Requirements and Installation\u003c/h2\u003e \u003cp\u003eBiolearn requires Python version 3.10 or newer. It can be installed using the Python package manager, pip, with the command pip install biolearn. The successful installation of the library can be verified through the import test of Biolearn\u0026rsquo;s core classes. The library is cross-platform and is compatible with major operating systems, including Windows, MacOS, and Linux.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eData Library and Model Gallery\u003c/h2\u003e \u003cp\u003eBiolearn incorporates a data library capable of loading and structuring datasets from a multitude of public sources like Gene Expression Omnibus (GEO), National Health and Nutrition Examination Survey (NHANES), and Framingham Heart Study. The model gallery within Biolearn holds reference implementations for various aging clocks and biomarkers, presenting a unified interface for users to apply these models to their datasets. All models were verified to be correct by comparing the outputs on a reference data set against their original implementations where available.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eHarmonization Process\u003c/h2\u003e \u003cp\u003eWe used Biolearn to harmonize several aging clocks. Clock definitions were standardized, specifying the name, publication year, applicable species, target tissue, and the biological aspect they predict (e.g., age, mortality risk). We provided sources for both the original publications and the coefficients necessary for clock applications. Coherence across biological modalities and datasets was assured through Biolearn\u0026rsquo;s systematic approach to data preprocessing, normalization, and imputation procedures.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eIntegration with Public Datasets\u003c/h2\u003e \u003cp\u003eBiolearn\u0026rsquo;s ability to interface seamlessly with public datasets was tested by integrating and formatting data from GEO and NHANES. Preprocessing pipelines were developed to convert raw data into a harmonized format suitable for subsequent analysis. Particular attention was given to metadata structures, variable normalization, and missing data treatment, ensuring consistent input formats required by the aging models.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eCell Type Deconvolution\u003c/h2\u003e \u003cp\u003eBiolearn\u0026rsquo;s deconvolution function estimates cell type proportions within a sample from bulk methylation data. It operates on the principle that bulk methylation is a composite of methylation profiles from various cell types, proportionate to their presence (B\u0026thinsp;=\u0026thinsp;X * P). Here, \u0026lsquo;B\u0026rsquo; represents bulk methylation, \u0026lsquo;X\u0026rsquo; is the matrix of cell-type-specific methylation profiles, and \u0026lsquo;P\u0026rsquo; is the vector of cell-type proportions. With known \u0026lsquo;B\u0026rsquo; and estimated \u0026lsquo;X,\u0026rsquo; \u0026lsquo;P\u0026rsquo; is computable through constrained quadratic programming, ensuring proportions remain within biological plausibility\u0026mdash;between zero and one and summing to one \u003csup\u003e26\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe incorporated two deconvolution variants tailored for blood methylation analysis: DeconvoluteBlood450K for the 450K platform and DeconvoluteBloodEPIC for the EPIC platform, addressing platform-specific biases \u003csup\u003e27\u003c/sup\u003e. The methylation matrix for each mode derives from the corresponding platform, encapsulating six major blood cell types\u0026mdash;neutrophils, monocytes, NK cells, B cells, CD4\u0026thinsp;+\u0026thinsp;T cells, and CD8\u0026thinsp;+\u0026thinsp;T cells. Selection of CpG sites for deconvolution relied on identifying the 50 most distinctively hyper- and hypo-methylated sites per cell type, prioritized by the significance of their differential methylation. This yielded 600 reference CpG sites per deconvolution mode \u003csup\u003e28,29\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed using tools embedded within the Biolearn library or through integration with renowned Python statistics libraries such as statsmodels and seaborn for visualization. The robustness and reproducibility of the analysis were ensured through the use of randomized cross-validation techniques for model assessment and bootstrapping methods for estimating confidence intervals where applicable. Survival analyses were conducted using the Cox Proportional Hazards model, adjusting for age and other relevant covariates. The performance of aging clocks in predicting chronological age and mortality risk was evaluated using metrics such as R\u003csup\u003e2\u003c/sup\u003e, hazard ratios, and p-values.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eWe thank all Biomarkers of Aging Consortium members for their valuable feedback and suggestions. We thank S. Horvath and A. Lu for sharing the GrimAgeV1 and V2. This work was inspired by methylCIPHER, an R package for DNA methylation clocks \u003csup\u003e36\u003c/sup\u003e. Supported by grants from the National Institute on Aging, Hevolution Foundation, Methuselah Foundation, and VoLo Foundation.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMoqri M et al (2024) A framework for validation of omic biomarkers of aging. Nat Med Press\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoqri M et al (2023) Biomarkers of aging for the identification and evaluation of longevity interventions. Cell 186:3758\u0026ndash;3775\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorvath S (2013) DNA methylation age of human tissues and cell types. Genome Biol 14:R115\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBell CG et al (2019) DNA methylation aging clocks: challenges and recommendations. Genome Biol 20:249\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBelsky DW et al (2022) DunedinPACE, a DNA methylation biomarker of the pace of aging. Elife 11:e73420\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen BH et al (2016) DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging 8:1844\u0026ndash;1865\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGalkin F et al (2020) Biohorology and biomarkers of aging: Current state-of-the-art, challenges and opportunities. Ageing Res Rev 60:101050\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorvath S, Raj K (2018) DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet 19:371\u0026ndash;384\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJylh\u0026auml;v\u0026auml; J, Pedersen NL, H\u0026auml;gg S (2017) Biol Age Predictors EBioMedicine 21:29\u0026ndash;36\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu AT et al (2019) DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging 11:303\u0026ndash;327\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLevine ME et al (2018) An epigenetic biomarker of aging for lifespan and healthspan. Aging 10:573\u0026ndash;591\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYing K et al (2024) Causality-enriched epigenetic age uncouples damage and adaptation. Nat Aging 1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s43587-023-00557-0\u003c/span\u003e\u003cspan address=\"10.1038/s43587-023-00557-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoqri M et al (2022) PRC2 clock: a universal epigenetic biomarker of aging and rejuvenation. 06.03.494609 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2022.06.03.494609\u003c/span\u003e\u003cspan address=\"10.1101/2022.06.03.494609\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eField AE et al (2018) DNA Methylation Clocks in Aging: Categories, Causes, and Consequences. Mol Cell 71:882\u0026ndash;895\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarioni RE et al (2015) The epigenetic clock is correlated with physical and cognitive fitness in the Lothian Birth Cohort 1936. Int J Epidemiol 44:1388\u0026ndash;1396\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEdgar R (2002) Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30:207\u0026ndash;210\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHannum G et al (2013) Genome-wide Methylation Profiles Reveal Quantitative Views of Human Aging Rates. Mol Cell 49:359\u0026ndash;367\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu AT et al (2022) DNA methylation GrimAge version 2. \u003cem\u003eAging\u003c/em\u003e 14, 9484\u0026ndash;9549\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Q et al (2019) Improved precision of epigenetic clock estimates across tissues and its implication for biological ageing. Genome Med 11:54\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBelsky DW et al (2020) Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. eLife 9, e54870\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThompson MJ et al (2018) A multi-tissue full lifespan epigenetic clock for mice. Aging 10:2832\u0026ndash;2854\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYing K et al (2023) \u003cem\u003eClockBase\u003c/em\u003e: a comprehensive platform for biological age profiling in human and mouse. Preprint at. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2023.02.28.530532\u003c/span\u003e\u003cspan address=\"10.1101/2023.02.28.530532\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Y et al (2021) DNA methylation-based sex classifier to predict sex and identify sex chromosome aneuploidy. BMC Genomics 22:484\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorvath S et al (2015) Accelerated epigenetic aging in Down syndrome. Aging Cell 14:491\u0026ndash;495\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLena PD, Sala C, Prodi A, Nardini C (2020) Methylation data imputation performances under different representations and missingness patterns. BMC Bioinformatics 21:268\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHouseman EA et al (2012) DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics 13:86\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHicks SC, Irizarry RA (2019) methylCC: technology-independent estimation of cell type composition using differentially methylated regions. Genome Biol 20:261\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReinius LE et al (2012) Differential DNA Methylation in Purified Human Blood Cells: Implications for Cell Lineage and Studies on Disease Susceptibility. PLoS ONE 7:e41361\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalas LA et al (2018) An optimized library for reference-based deconvolution of whole-blood biospecimens assayed using the Illumina HumanMethylationEPIC BeadArray. Genome Biol 19:64\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeters MJ et al (2015) The transcriptional landscape of age in human peripheral blood. Nat Commun 6:8570\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTyshkovskiy A et al (2019) Identification and Application of Gene Expression Signatures Associated with Lifespan Extension. Cell Metabol 30:573\u0026ndash;593e8\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTyshkovskiy A et al (2023) Distinct longevity mechanisms across and within species and their association with aging. Cell 186:2929\u0026ndash;2949e20\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAramillo Irizar P et al (2018) Transcriptomic alterations during ageing reflect the shift from cancer to degenerative diseases in the elderly. Nat Commun 9:327\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi Q et al (2023) Biomarkers of aging for the identification and evaluation of longevity interventions. Popul Health Metrics 186:3758\u0026ndash;3775\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwon D, Belsky DW (2021) A toolkit for quantification of biological age from blood chemistry and organ function test data: BioAge. \u003cem\u003eGeroScience\u003c/em\u003e 43, 2795\u0026ndash;2808\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThrush KL, Higgins-Chen AT, Liu Z, Levine ME (2022) R methylCIPHER: A Methylation Clock Investigational Package for Hypothesis-Driven Evaluation \u0026amp; Research. 07.13.499978 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2022.07.13.499978\u003c/span\u003e\u003cspan address=\"10.1101/2022.07.13.499978\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCamillo LP (2023) de L. pyaging: a Python-based compendium of GPU-optimized aging clocks. 11.28.569069 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2023.11.28.569069\u003c/span\u003e\u003cspan address=\"10.1101/2023.11.28.569069\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J et al (2020) Prevalence of comorbidities and its effects in patients infected with SARS-CoV-2: a systematic review and meta-analysis. Int J Infect Dis 94:91\u0026ndash;95\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin Q et al (2016) DNA methylation levels at individual age-associated CpG sites can be indicative for life expectancy. Aging 8:394\u0026ndash;401\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorvath S et al (2018) Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria Syndrome and ex vivo studies. Aging 10:1758\u0026ndash;1775\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcEwen LM et al (2020) The PedBE clock accurately estimates DNA methylation age in pediatric buccal cells. \u003cem\u003eProc. Natl. Acad. Sci. U.S.A.\u003c/em\u003e 117, 23329\u0026ndash;23335\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu AT et al (2019) DNA methylation-based estimator of telomere length. Aging 11:5895\u0026ndash;5923\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKnight AK et al (2016) An epigenetic clock for gestational age at birth based on blood methylation data. Genome Biol 17:206\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee Y et al (2019) Placental epigenetic clocks: estimating gestational age using placental DNA methylation levels. Aging 11:4238\u0026ndash;4253\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcCartney DL et al (2018) Epigenetic prediction of complex traits and death. Genome Biol 19:136\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReed ZE, Suderman MJ, Relton CL, Davis OSP, Hemani G (2020) The association of DNA methylation with body mass index: distinguishing between predictors and biomarkers. Clin Epigenetics 12:50\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKirby MK et al (2017) Genome-wide DNA methylation measurements in prostate tissues uncovers novel prostate cancer diagnostic biomarkers and transcription factor binding patterns. BMC Cancer 17:273\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu R et al (2017) Circulating tumour DNA methylation markers for diagnosis and prognosis of hepatocellular carcinoma. Nat Mater 16:1155\u0026ndash;1161\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWesterman K et al (2020) Epigenomic Assessment of Cardiovascular Disease Risk and Interactions With Traditional Risk Metrics. J Am Heart Assoc 9:e015299\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBahado-Singh RO et al (2021) Artificial intelligence and leukocyte epigenomics: Evaluation and prediction of late-onset Alzheimer\u0026rsquo;s disease. PLoS ONE 16:e0248375\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarbu MC et al (2021) Epigenetic prediction of major depressive disorder. Mol Psychiatry 26:5112\u0026ndash;5123\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":false,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4481437/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4481437/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAging biomarkers are essential for understanding and quantifying the aging process and developing targeted longevity interventions. However, validation of these tools has been hindered by the lack of standardized approaches for cross-population validation, disparate biomarker designs, and inconsistencies in dataset structures. To address these challenges, we developed Biolearn, an open-source library that provides a unified framework for the curation, harmonization, and systematic evaluation of aging biomarkers. Leveraging Biolearn, we conducted a comprehensive evaluation of various aging biomarkers across multiple datasets. Our systematic approach involved three key steps: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) harmonizing existing and novel aging biomarkers in standardized formats; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) unifying public datasets to ensure coherent structuring and formatting; and (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) applying computational methodologies to assess the harmonized biomarkers against the unified datasets. This evaluation yielded valuable insights into the performance, robustness, and generalizability of aging biomarkers across different populations and datasets. The Biolearn python library, which forms the foundation of this systematic evaluation, is freely available at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://Bio-Learn.github.io\u003c/span\u003e\u003cspan address=\"https://Bio-Learn.github.io\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Our work establishes a unified framework for the curation and evaluation of aging biomarkers, paving the way for more efficient and effective clinical validation and application in the field of longevity research.\u003c/p\u003e","manuscriptTitle":"A Unified Framework for Systematic Curation and Evaluation of Aging Biomarkers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-24 09:27:41","doi":"10.21203/rs.3.rs-4481437/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-aging","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nataging","sideBox":"Learn more about [Nature Aging](https://www.nature.com/nataging/)","snPcode":"","submissionUrl":"","title":"Nature Aging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e92f00f9-a1ce-4250-ada3-9ff6114123c9","owner":[],"postedDate":"July 24th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":35020918,"name":"Biological sciences/Computational biology and bioinformatics/Computational models"},{"id":35020919,"name":"Biological sciences/Computational biology and bioinformatics/Software"},{"id":35020920,"name":"Biological sciences/Systems biology"}],"tags":[],"updatedAt":"2025-11-05T08:08:48+00:00","versionOfRecord":{"articleIdentity":"rs-4481437","link":"https://doi.org/10.1038/s43587-025-00987-y","journal":{"identity":"nature-aging","isVorOnly":false,"title":"Nature Aging"},"publishedOn":"2025-11-04 05:00:00","publishedOnDateReadable":"November 4th, 2025"},"versionCreatedAt":"2024-07-24 09:27:41","video":"","vorDoi":"10.1038/s43587-025-00987-y","vorDoiUrl":"https://doi.org/10.1038/s43587-025-00987-y","workflowStages":[]},"version":"v1","identity":"rs-4481437","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4481437","identity":"rs-4481437","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}