Predictor Of DNA Methylation Research Articles

Blood-based epigenome-wide analyses of chronic low-grade inflammation across diverse population cohorts

Chronic inflammation is a hallmark of age-related disease states. The effectiveness of inflammatory proteins including C-reactive protein (CRP) in assessing long-term inflammation is hindered by their phasic nature. DNA methylation (DNAm) signatures of CRP may act as more reliable markers of chronic inflammation. We show that inter-individual differences in DNAm capture 50% of the variance in circulating CRP (N= 17,936, Generation Scotland). We develop a series of DNAm predictors of CRP using state-of-the-art algorithms. An elastic-net-regression-based predictor outperformed competing methods and explained 18% of phenotypic variance in the Lothian Birth Cohort of 1936 (LBC1936) cohort, doubling that of existing DNAm predictors. DNAm predictors performed comparably in four additional test cohorts (Avon Longitudinal Study of Parents and Children, Health for Life in Singapore, Southall and Brent Revisited, and LBC1921), including for individuals of diverse genetic ancestry and different age groups. The best-performing predictor surpassed assay-measured CRP and a genetic score in its associations with 26 health outcomes. Our findings forge new avenues for assessing chronic low-grade inflammation in diverse populations.

Abstract 11956: Blood DNA Methylation Predicts Differential Metabolic Outcomes Induced by Different Types of Weight-Loss Intervention: Findings From an Epigenome-Wide Association Study

Background Previous epigenetic studies in obesity interventions have focused on the intervention-associated DNA methylation (DNAm) changes, yet whether the DNAm before intervention could be used as a predictor of prognosis has not been investigated. Here we hypothesize that baseline DNAm can predict differential metabolic outcomes, i.e., changes in fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c), induced by different weight-loss interventions. Methods The analyses included 75 individuals with obesity in the HeadsUp Study, a longitudinal study assessing the effectiveness of different weight-loss interventions, from the non-surgical intensive medical intervention (IMI), adjustable gastric band (BAND) and Roux-en-Y gastric bypass (RYGB) groups (n = 25 per group). Genome-wide DNAm was determined using baseline blood samples, while changes in FPG and HbA1c were measured at 1-year follow-up. Epigenome-wide association studies were performed on 455,983 CpGs to identify DNAm predictors of differential metabolic outcomes, adjusting for age, sex, BMI, FPG/HbA1c at baseline, weight loss and batch effect. Results DNAm levels at 3,216 and 117 CpG loci significantly predicted differential changes in FPG and HbA1c, respectively, when comparing RYGB vs. IMI groups, but not in other group comparisons (RYGB vs. BAND or BAND vs. IMI, with the exception of 6 CpGs identified for HbA1c changes when comparing RYGB vs. BAND), after FDR correction for multiple testing. Of these, altered DNAm at 79 CpGs significantly predicted changes in both FPG and HbA1c (RYGB vs. IMI). These CpGs were annotated to 68 genes, among which the most significant ones were SNX18, LNPK, CCDC178, ADAM7 and ZNF23 . These genes are mostly involved in adaptive thermogenesis, temperature homeostasis and regulation of cell population proliferation. Conclusions To our knowledge, this is the first study demonstrating that baseline blood DNAm predicts differential metabolic outcomes induced by different types of weight-loss interventions, independent of weight loss. Such results help explain, at least partly, the favorable metabolic outcomes of RYGB over other types of weight-loss intervention (e.g., BAND, IMI), and are likely to lead to precision strategies for obesity treatment.

Significant variation in the performance of DNA methylation predictors across data preprocessing and normalization strategies

BackgroundDNA methylation (DNAm)-based predictors hold great promise to serve as clinical tools for health interventions and disease management. While these algorithms often have high prediction accuracy, the consistency of their performance remains to be determined. We therefore conduct a systematic evaluation across 101 different DNAm data preprocessing and normalization strategies and assess how each analytical strategy affects the consistency of 41 DNAm-based predictors.ResultsOur analyses are conducted in a large EPIC DNAm array dataset from the Jackson Heart Study (N = 2053) that included 146 pairs of technical replicate samples. By estimating the average absolute agreement between replicate pairs, we show that 32 out of 41 predictors (78%) demonstrate excellent consistency when appropriate data processing and normalization steps are implemented. Across all pairs of predictors, we find a moderate correlation in performance across analytical strategies (mean rho = 0.40, SD = 0.27), highlighting significant heterogeneity in performance across algorithms. Successful or unsuccessful removal of technical variation furthermore significantly impacts downstream phenotypic association analysis, such as all-cause mortality risk associations.ConclusionsWe show that DNAm-based algorithms are sensitive to technical variation. The right choice of data processing strategy is important to achieve reproducible estimates and improve prediction accuracy in downstream phenotypic association analyses. For each of the 41 DNAm predictors, we report its degree of consistency and provide the best performing analytical strategy as a guideline for the research community. As DNAm-based predictors become more and more widely used, our work helps improve their performance and standardize their implementation.

The immune factors driving DNA methylation variation in human blood

Epigenetic changes are required for normal development, yet the nature and respective contribution of factors that drive epigenetic variation in humans remain to be fully characterized. Here, we assessed how the blood DNA methylome of 884 adults is affected by DNA sequence variation, age, sex and 139 factors relating to life habits and immunity. Furthermore, we investigated whether these effects are mediated or not by changes in cellular composition, measured by deep immunophenotyping. We show that DNA methylation differs substantially between naïve and memory T cells, supporting the need for adjustment on these cell-types. By doing so, we find that latent cytomegalovirus infection drives DNA methylation variation and provide further support that the increased dispersion of DNA methylation with aging is due to epigenetic drift. Finally, our results indicate that cellular composition and DNA sequence variation are the strongest predictors of DNA methylation, highlighting critical factors for medical epigenomics studies.

Blood-based epigenome-wide analyses of cognitive abilities

BackgroundBlood-based markers of cognitive functioning might provide an accessible way to track neurodegeneration years prior to clinical manifestation of cognitive impairment and dementia.ResultsUsing blood-based epigenome-wide analyses of general cognitive function, we show that individual differences in DNA methylation (DNAm) explain 35.0% of the variance in general cognitive function (g). A DNAm predictor explains ~4% of the variance, independently of a polygenic score, in two external cohorts. It also associates with circulating levels of neurology- and inflammation-related proteins, global brain imaging metrics, and regional cortical volumes.ConclusionsAs sample sizes increase, the ability to assess cognitive function from DNAm data may be informative in settings where cognitive testing is unreliable or unavailable.

Epigenetic prediction of 17β-estradiol and relationship to trauma-related outcomes in women

17β-estradiol (E2) levels in women correlate with multiple neuropsychiatric symptoms, including those that are stress-related. Furthermore, prior work from our group has demonstrated that E2 status influences DNA methylation (DNAm) across the genome. We developed and validated a DNAm-based predictor of E2 (one of four naturally occurring estrogens) using a training set of 183 females and a test set of 79 females from the same traumatized cohort. We showed that predicted E2 levels were highly correlated with measured E2 concentrations in our testing set (r ​= ​0.75, p ​= ​1.8e-15). We further demonstrated that predicted E2 concentrations, in combination with measured values, negatively correlated with current post-traumatic stress disorder (PTSD) (β ​= ​−0.38, p ​= ​0.01) and major depressive disorder (MDD) diagnoses (β ​= ​−0.45, p ​= ​0.02), as well as a continuous measure of PTSD symptom severity (β ​= ​−2.3, p ​= ​0.007) in females. Finally, we tested our predictor in an independent data set (n ​= ​85) also comprised of recently traumatized female subjects to determine if the predictor would generalize to a different population than the one on which it was developed. We found that the correlation between predicted and actual E2 concentrations in the external validation data set was also high (r ​= ​0.48, p ​= ​3.0e-6). While further validation is warranted, a DNAm predictor of E2 concentrations will advance our understanding of hormone-epigenetic interactions. Furthermore, such a DNAm predictor may serve as an epigenetic proxy for E2 concentrations and thus provide an important biomarker to better evaluate the contribution of E2 to current and potentially future psychiatric symptoms in samples for which E2 is not measured.

Blood-derived DNA methylation predictors of mortality discriminate tumor and healthy tissue in multiple organs.

Evidence has shown that certain methylation markers derived from blood can mirror corresponding methylation signatures in internal tissues. In the current study, we aimed to investigate two strong epigenetic predictors for life span, derived from blood DNA methylation data, in tissue samples of solid cancer patients. Using data from the Cancer Genome Atlas (TCGA) and the German DACHS study, we compared a mortality risk score (MRscore) and DNAmPhenoAge in paired tumor and adjacent normal tissue samples of patients with lung (N = 69), colorectal (n = 299), breast (n = 90), head/neck (n = 50), prostate (n = 50), and liver (n = 50) cancer. To explore the concordance across tissue and blood, we additionally assessed the two markers in blood samples of colorectal cancer (CRC) cases and matched controls (n = 93) in the DACHS+ study. The MRscore was significantly elevated in tumor tissues compared to normal tissues of all cancers except prostate cancer, for which an opposite pattern was observed. DNAmPhenoAge was consistently higher in all tumor tissues. The MRscore discriminated lung, colorectal, and prostate tumor tissues from normal tissues with very high accuracy [AUCs of 0.87, 0.99 (TCGA) /0.94 (DACHS), and 0.92, respectively]. DNAmPhenoAge accurately discriminated five types of tumor tissues from normal tissues (except prostate cancer), with AUCs of 0.82–0.93. The MRscore was also significantly higher in blood samples of CRC cases than in controls, with areas under the curve (AUC) of 0.74, whereas DNAmPhenoAge did not distinguish cases from controls, with AUC of 0.54. This study provides compelling evidence that blood‐derived DNAm markers could reflect methylation changes in less accessible tissues. Further research should explore the potential use of these findings for cancer diagnosis and early detection.

Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm.

Biological aging is the gradual, progressive decline in system integrity that occurs with advancing chronological age, causing morbidity and disability. Measurements of the pace of aging are needed as surrogate endpoints in trials of therapies designed to prevent disease by slowing biological aging. We report a blood-DNA-methylation measure that is sensitive to variation in pace of biological aging among individuals born the same year. We first modeled change-over-time in 18 biomarkers tracking organ-system integrity across 12 years of follow-up in n = 954 members of the Dunedin Study born in 1972-1973. Rates of change in each biomarker over ages 26-38 years were composited to form a measure of aging-related decline, termed Pace-of-Aging. Elastic-net regression was used to develop a DNA-methylation predictor of Pace-of-Aging, called DunedinPoAm for Dunedin(P)ace(o)f(A)ging(m)ethylation. Validation analysis in cohort studies and the CALERIE trial provide proof-of-principle for DunedinPoAm as a single-time-point measure of a person's pace of biological aging.

Schizophrenia and Epigenetic Aging Biomarkers: Increased Mortality, Reduced Cancer Risk, and Unique Clozapine Effects

BackgroundSchizophrenia (SZ) is associated with increased all-cause mortality, smoking, and age-associated proteins, yet multiple previous studies found no association between SZ and biological age using Horvath’s epigenetic clock, a well-established aging biomarker based on DNA methylation. However, numerous epigenetic clocks that may capture distinct aspects of aging have been developed. This study tested the hypothesis that altered aging in SZ manifests in these other clocks. MethodsWe performed a comprehensive analysis of 14 epigenetic clocks categorized according to what they were trained to predict: chronological age, mortality, mitotic divisions, or telomere length. To understand the etiology of biological age differences, we also examined DNA methylation predictors of smoking, alcohol, body mass index, serum proteins, and cell proportions. We independently analyzed 3 publicly available multiethnic DNA methylation data sets from whole blood, a total of 567 SZ cases and 594 nonpsychiatric controls. ResultsAll data sets showed accelerations in SZ for the 3 mortality clocks up to 5 years, driven by smoking and elevated levels of 6 age-associated proteins. The 2 mitotic clocks were decelerated in SZ related to antitumor natural killer and CD8T cells, which may help explain conflicting reports about low cancer rates in epidemiological studies of SZ. One cohort with available medication data showed that clozapine is associated with male-specific decelerations up to 7 years in multiple chronological age clocks. ConclusionsOur study demonstrates the utility of studying the various epigenetic clocks in tandem and highlights potential mechanisms by which mental illness influences long-term outcomes, including cancer and early mortality.

Epigenetic prediction of complex traits and death

BackgroundGenome-wide DNA methylation (DNAm) profiling has allowed for the development of molecular predictors for a multitude of traits and diseases. Such predictors may be more accurate than the self-reported phenotypes and could have clinical applications.ResultsHere, penalized regression models are used to develop DNAm predictors for ten modifiable health and lifestyle factors in a cohort of 5087 individuals. Using an independent test cohort comprising 895 individuals, the proportion of phenotypic variance explained in each trait is examined for DNAm-based and genetic predictors. Receiver operator characteristic curves are generated to investigate the predictive performance of DNAm-based predictors, using dichotomized phenotypes. The relationship between DNAm scores and all-cause mortality (n = 212 events) is assessed via Cox proportional hazards models. DNAm predictors for smoking, alcohol, education, and waist-to-hip ratio are shown to predict mortality in multivariate models. The predictors show moderate discrimination of obesity, alcohol consumption, and HDL cholesterol. There is excellent discrimination of current smoking status, poorer discrimination of college-educated individuals and those with high total cholesterol, LDL with remnant cholesterol, and total:HDL cholesterol ratios.ConclusionsDNAm predictors correlate with lifestyle factors that are associated with health and mortality. They may supplement DNAm-based predictors of age to identify the lifestyle profiles of individuals and predict disease risk.

DNA methylation as a predictor of fetal alcohol spectrum disorder

BackgroundFetal alcohol spectrum disorder (FASD) is a developmental disorder that manifests through a range of cognitive, adaptive, physiological, and neurobiological deficits resulting from prenatal alcohol exposure. Although the North American prevalence is currently estimated at 2–5%, FASD has proven difficult to identify in the absence of the overt physical features characteristic of fetal alcohol syndrome. As interventions may have the greatest impact at an early age, accurate biomarkers are needed to identify children at risk for FASD. Building on our previous work identifying distinct DNA methylation patterns in children and adolescents with FASD, we have attempted to validate these associations in a different clinical cohort and to use our DNA methylation signature to develop a possible epigenetic predictor of FASD.MethodsGenome-wide DNA methylation patterns were analyzed using the Illumina HumanMethylation450 array in the buccal epithelial cells of a cohort of 48 individuals aged 3.5–18 (24 FASD cases, 24 controls). The DNA methylation predictor of FASD was built using a stochastic gradient boosting model on our previously published dataset FASD cases and controls (GSE80261). The predictor was tested on the current dataset and an independent dataset of 48 autism spectrum disorder cases and 48 controls (GSE50759).ResultsWe validated findings from our previous study that identified a DNA methylation signature of FASD, replicating the altered DNA methylation levels of 161/648 CpGs in this independent cohort, which may represent a robust signature of FASD in the epigenome. We also generated a predictive model of FASD using machine learning in a subset of our previously published cohort of 179 samples (83 FASD cases, 96 controls), which was tested in this novel cohort of 48 samples and resulted in a moderately accurate predictor of FASD status. Upon testing the algorithm in an independent cohort of individuals with autism spectrum disorder, we did not detect any bias towards autism, sex, age, or ethnicity.ConclusionThese findings further support the association of FASD with distinct DNA methylation patterns, while providing a possible entry point towards the development of epigenetic biomarkers of FASD.

FKBP5 DNA methylation does not mediate the association between childhood maltreatment and depression symptom severity in the Detroit Neighborhood Health Study

Exposure to childhood maltreatment increases the risk of developing mental illness later in life. Childhood maltreatment and depression have both been associated with dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis—a key regulator of the body's stress response. Additionally, HPA axis dysregulation has been implicated in the etiology of a range of mental illnesses. A substantial body of work has shown history of childhood maltreatment alters DNA methylation levels within key HPA axis genes. We therefore investigated whether one of these key genes, FKBP5 mediates the relationship between childhood maltreatment and depression, and assessed FKBP5 DNA methylation and gene expression within 112 adults from the Detroit Neighborhood Health Study (DNHS). DNA methylation was assessed in 4 regions, including the upstream promoter, downstream promoter, and two glucocorticoid response elements (GREs) via pyrosequencing using whole blood derived DNA; Taqman assays measured relative RNA expression from leukocytes. Mediation analyses were conducted using sequential linear regression. Childhood maltreatment was significantly associated with depression symptom severity (FDR < 0.006), but was not a significant predictor of DNA methylation in any of the four loci examined. FKBP5 showed elevated expression levels in participants with vs. without a history of depression (p < 0.001); no significant difference in gene expression levels was observed in relation to childhood maltreatment (p > 0.05). Our results suggest DNA methylation does not mediate the childhood maltreatment-depression association in the DNHS.

Increased global placental DNA methylation levels are associated with gestational diabetes.

BackgroundGestational diabetes mellitus (GDM) is associated with adverse pregnancy outcomes. It is known that GDM is associated with an altered placental function and changes in placental gene regulation. More recent studies demonstrated an involvement of epigenetic mechanisms. So far, the focus regarding placental epigenetic changes in GDM was set on gene-specific DNA methylation analyses. Studies that robustly investigated placental global DNA methylation are lacking. However, several studies showed that tissue-specific alterations in global DNA methylation are independently associated with type 2 diabetes. Thus, the aim of this study was to characterize global placental DNA methylation by robustly measuring placental DNA 5-methylcytosine (5mC) content and to examine whether differences in placental global DNA methylation are associated with GDM.MethodsGlobal DNA methylation was quantified by the current gold standard method, LC-MS/MS. In total, 1030 placental samples were analyzed in this single-center birth cohort study.ResultsMothers with GDM displayed a significantly increased global placental DNA methylation (3.22 ± 0.63 vs. 3.00 ± 0.46 %; p = 0.013; ±SD). Bivariate logistic regression showed a highly significant positive correlation between global placental DNA methylation and the presence of GDM (p = 0.0009). Quintile stratification according to placental DNA 5mC levels revealed that the frequency of GDM was evenly distributed in quintiles 1–4 (2.9–5.3 %), whereas the frequency in the fifth quintile was significantly higher (10.7 %; p = 0.003). Bivariate logistic models adjusted for maternal age, BMI, ethnicity, recurrent miscarriages, and familiar diabetes predisposition clearly demonstrated an independent association between global placental DNA hypermethylation and GDM. Furthermore, an ANCOVA model considering known predictors of DNA methylation substantiated an independent association between GDM and placental DNA methylation.ConclusionsThis is the first study that employed a robust quantitative assessment of placental global DNA methylation in over a thousand placental samples. The study provides large scale evidence that placental global DNA hypermethylation is associated with GDM, independent of established risk factors.

Abstract P362: Association of DNA Methylation Age With Risk of Type 2 diabetes: The Atherosclerosis Risk in Communities (ARIC) Study

Introduction: It has been proposed that DNA methylation patterns can be used to characterize rate of aging, but it is not known whether methylation age associates with aging-related diseases and conditions independently of chronological age and other risk factors. Hypothesis: There is a positive association between age acceleration (defined as the difference between DNA methylation age and chronological age) and risk of type 2 diabetes (T2D). Methods: We estimated DNA methylation age in African American men and women of the Atherosclerosis Risk in Communities Study using two published predictors. Incident T2D was identified from follow-up visits and telephone interviews. Logistic and Cox proportional hazards regression models were used to estimate the association of age acceleration with incident T2D, modeling age acceleration both continuously and categorically in three percentile groups (0-25th, 25-75th, 75-100th). Regression models were adjusted for chronological age, sex, white blood cell distribution, education, smoking, alcohol, sport index, body mass index, waist to hip ratio, blood pressure, antihypertensive and lipid-lowering medication, HDL and LDL cholesterol, and triglycerides. Results: After adjustment for chronological age, sex, and white blood cell type proportions, those with greater age acceleration were more likely to have prevalent T2D (p-trend ≤ 0.003 across percentile categories of age acceleration for both DNA methylation predictors). Among 1,547 participants who were free of T2D at baseline (mean age: 56 years, range: 47-70), there were 170 incident cases of T2D over a mean 5.4 years of follow-up. T2D incidence was not associated with age acceleration in either crude or risk factor-adjusted models. In the fully-adjusted logistic regression models, odds ratios (95% confidence intervals) for T2D across the percentile categories of age acceleration were 1 (ref.), 0.92 (0.58, 1.45), and 1.42 (0.84, 2.40); and 1 (ref.), 0.87 (0.55, 1.39), and 0.93 (0.52, 1.64), respectively, for the two DNA methylation age predictors. Conclusion: In conclusion, DNA methylation age acceleration, proposed as a marker of the rate of aging, was associated cross-sectionally with prevalent diabetes, but did not predict risk of T2D independently of chronological age. Additional prospective studies are needed to explore relationships with development of other chronic conditions in order to determine if DNA methylation age can serve as a useful proxy for rate of aging.

S.3.2 Genetic predictors of illness and treatment response

Exposure to childhood maltreatment increases the risk of developing mental illness later in life. Childhood maltreatment and depression have both been associated with dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis—a key regulator of the body's stress response. Additionally, HPA axis dysregulation has been implicated in the etiology of a range of mental illnesses. A substantial body of work has shown history of childhood maltreatment alters DNA methylation levels within key HPA axis genes. We therefore investigated whether one of these key genes, FKBP5 mediates the relationship between childhood maltreatment and depression, and assessed FKBP5 DNA methylation and gene expression within 112 adults from the Detroit Neighborhood Health Study (DNHS). DNA methylation was assessed in 4 regions, including the upstream promoter, downstream promoter, and two glucocorticoid response elements (GREs) via pyrosequencing using whole blood derived DNA; Taqman assays measured relative RNA expression from leukocytes. Mediation analyses were conducted using sequential linear regression. Childhood maltreatment was significantly associated with depression symptom severity (FDR < 0.006), but was not a significant predictor of DNA methylation in any of the four loci examined. FKBP5 showed elevated expression levels in participants with vs. without a history of depression (p < 0.001); no significant difference in gene expression levels was observed in relation to childhood maltreatment (p > 0.05). Our results suggest DNA methylation does not mediate the childhood maltreatment-depression association in the DNHS.

Association of breast cancer DNA methylation profiles with hormone receptor status and response to tamoxifen.

We have generated DNA methylation profiles of 148 human breast tumors and found significant differences in hormone receptor (HR) status between clusters of DNA methylation profiles. Of 35 DNA methylation markers analyzed, the ESR1 gene, encoding estrogen receptor alpha, proved to be the best predictor of progesterone receptor status, whereas methylation of the PGR gene, encoding progesterone receptor, was the best predictor of estrogen receptor status. ESR1 methylation outperformed HR status as a predictor of clinical response in patients treated with the antiestrogen tamoxifen, whereas promoter methylation of the CYP1B1 gene, encoding a tamoxifen- and estradiol-metabolizing cytochrome p450, predicted response differentially in tamoxifen-treated and nontamoxifen-treated patients. High levels of promoter methylation of the ARHI gene, encoding a RAS-related small G-protein, were strongly predictive of good survival in patients who had not received tamoxifen therapy. Our results reveal an as yet unrecognized degree of interaction between DNA methylation and HR biology in breast cancer cells and suggest potentially clinically useful novel DNA methylation predictors of response to hormonal and non-hormonal breast cancer therapy.

Cloudy central corneal dystrophy of François. Five cases in the same family.

<h3>Abstract</h3> <h3>Background</h3> Genome-wide DNA methylation (DNAm) profiling has allowed for the development of molecular predictors for a multitude of traits and diseases. Such predictors may be more accurate than the self-reported phenotypes, and could have clinical applications. Here, penalised regression models were used to develop DNAm predictors for body mass index (BMI), smoking status, alcohol consumption, and educational attainment in a cohort of 5,100 individuals. Using an independent test cohort comprising 906 individuals, the proportion of phenotypic variance explained in each trait was examined for DNAm-based and genetic predictors. Receiver operator characteristic curves were generated to investigate the predictive performance of DNAm-based predictors, using dichotomised phenotypes. The relationship between DNAm scores and all-cause mortality (n = 214 events) was assessed via Cox proportional-hazards models. <h3>Results</h3> The DNAm-based predictors explained different proportions of the phenotypic variance for BMI (12%), smoking (60%), alcohol consumption (12%) and education (3%). The combined genetic and DNAm predictors explained 20% of the variance in BMI, 61% in smoking, 13% in alcohol consumption, and 6% in education. DNAm predictors for smoking, alcohol, and education but not BMI predicted mortality in univariate models. The predictors showed moderate discrimination of obesity (AUC=0.67) and alcohol consumption (AUC=0.75), and excellent discrimination of current smoking status (AUC=0.98). There was poorer discrimination of college-educated individuals (AUC=0.59). <h3>Conclusions</h3> DNAm predictors correlate with lifestyle factors that are associated with health and mortality. They may supplement DNAm-based predictors of age to identify the lifestyle profiles of individuals and predict disease risk. <h3>List of abbreviations</h3> DNAm: DNA methylation BMI: Body mass index AUC: Area under the curve CpG: Cytosine phosphate Guanine dinucleotide EWAS: Epigenome-wide association study GS:SFHS: Generation Scotland: The Scottish family health study LBC1936: Lothian birth cohort 1936 LASSO: Least absolute shrinkage and selector operator HR: Hazard ratio CI: Confidence interval STRADL: Stratifying resilience and depression longitudinally