Karen E. Peterson
· Stanley M. Garn Collegiate Professor of Nutritional Sciences, Professor of Global Public Health, Professor of Environmental Health SciencesVerifiedUniversity of Michigan · Nutritional Sciences
Active 1975–2026
About
Karen E. Peterson is the Stanley M. Garn Professor and inaugural Chair of Nutritional Sciences at the University of Michigan School of Public Health, where she also holds joint appointments as Professor of Environmental Health Sciences and Professor of Global Public Health. Her research considers the effects of diet and toxicants, such as metals and endocrine disrupting chemicals, on physical growth, maturation, and the development of cardiometabolic risk across the life course via epigenetic mechanisms. She also focuses on the design and evaluation of health behavior interventions aimed at reducing obesity in multi-ethnic women and children. Dr. Peterson has served as Contact Principal Investigator for the U-M Children's Environmental Health and Disease Protection Center and is currently the Contact PI for the project 'E3Gen,' which investigates multigenerational influences of social structure on toxicant exposures and health in the ELEMENT Cohort, following three generations of women and their offspring in Mexico City. She is also the Associate Director of the Michigan Nutrition and Obesity Research Center and directs the MNORC Nutrition Assessment Lab. Her expertise includes the validation of self-report instruments for dietary assessment, relative to unbiased nutrient biomarkers such as metabolomics. Additionally, she is the founding Director of the Momentum Center at the University of Michigan, which fosters transdisciplinary research to end child obesity.
Research topics
- Medicine
- Internal medicine
- Psychiatry
- Machine Learning
- Computer Science
- Demography
- Biology
- Gerontology
- Physiology
- Endocrinology
- Environmental health
- Chemistry
- Artificial Intelligence
- Mathematics
- Clinical psychology
- Algorithm
- Psychology
- Statistics
- Genetics
- Environmental chemistry
- Food science
Selected publications
Bridging the gap: Enhancing the generalizability of epigenetic clocks through transfer learning
The Annals of Applied Statistics · 2026-03-01
articleChanges in DNA methylation patterns exhibit a high correlation with chronological age. Epigenetic clocks, developed through statistical models that estimate epigenetic age using the methylation levels of cytosine-guanine dinucleotide (CpG) sites, have emerged as powerful tools for understanding aging and age-related diseases. Despite their popularity, the generalizability of these clocks across diverse populations remains a challenge. Some of the widely used epigenetic clocks, such as Horvath’s clock (Genome Biol. 14 (2013) 1–20) and the PedBE clock (Proc. Natl. Acad. Sci. USA 117 (2020) 23329–23335), are shown to perform poorly in our target cohort. This loss of prediction accuracy raises concerns about their viability in calculating biological age in distinct demographic and ethnic groups. Technically, the feature space of existing clocks is yielded with an obsolete technique, potentially leading to systematic bias in the analysis of all target data generated by the EPIC 850K array. To address both population heterogeneity and technological advances, we adopt a transfer learning framework to calibrate existing epigenetic clocks by borrowing shared knowledge from diverse datasets. Furthermore, our transfer learning is built on kriging- and DNN-based methods for feature adaptation, to close the gap between existing clocks and our target data. We analyze data collected from 523 blood samples from a cohort of children and adolescents in the Early Life Exposure in Mexico to Environmental Toxicants (ELEMENT) study and show that our proposed transfer learning methods significantly improve prediction performance compared to existing clocks. Performance is further enhanced by using the CpG sites profiled on the higher-resolution EPIC array. More importantly, calibrated clocks produce epigenetic age accelerations that correlate better with stages of sexual maturation. Our methodology demonstrates the potential to bridge the gap between different DNA methylation datasets and various profiling platforms, thereby enhancing the applicability of epigenetic clocks across diverse population groups and contributing to more accurate aging research.
Environmental Health · 2026-04-28
articleOpen accessPhthalates are widespread endocrine-disrupting chemicals (EDCs) that may affect bone metabolism, though human studies remain limited. This study investigated the long-term associations between measures of phthalate exposure during pregnancy, a period of heightened bone activity, and midlife bone mineral density (BMD). Participants were women in the Early Life Exposure in Mexico to Environmental Toxicants (ELEMENT) cohort who had gestational phthalate measurements and midlife BMD assessments (n = 180), including 60 who received calcium supplementation during pregnancy and first year postpartum as part of a clinical trial. Nine phthalate metabolites were measured in urine samples collected at up to three pregnancy trimesters, corrected for specific gravity, and a molar sum was calculated for di-2-ethylhexyl phthalate metabolites (ΣDEHP). BMD of the spine, hip, and forearm were assessed using dual-energy X-ray absorptiometry (DXA) at a mean age of 50.1 years. We used linear models to estimate differences in BMD per doubling of geometric mean concentrations of individual phthalates across pregnancy, adjusting for perinatal calcium supplementation status, age at midlife and pre-pregnancy BMI. Effect modification by perinatal calcium supplementation was tested through interaction terms and sensitivity analyses additionally adjusted for menopausal status. Quantile-based g-computation was used to evaluate the overall effect of the phthalate mixtures. Most phthalate metabolites were inversely associated with spine and hip BMD, whereas some showed significant positive associations with forearm BMD. Perinatal calcium supplementation significantly modified associations between phthalate metabolites and spine BMD, showing consistent negative trends in the placebo group and positive trends in the supplementation group (ΣDEHP, p-interaction = 0.03). For example, a doubling in ΣDEHP was associated with 0.12 lower spine BMD in the placebo group (95%CI= -0.27, 0.03) and 0.28 higher spine BMD in the supplementation group (95%CI: -0.03, 0.59). Similar patterns were seen for hip BMD. Mixture analysis results aligned with single-pollutant findings. Phthalate exposure during pregnancy was associated with midlife BMD in a site-specific manner. Calcium supplementation in pregnancy and first year postpartum modified these associations at the spine and hip, suggesting a potential protective role.
Nutrition Metabolism and Cardiovascular Diseases · 2026-01-23
articleOpen accessBACKGROUND AND AIMS: To examine the cross-sectional relationship between dietary intake and epigenetic age acceleration, as well as the prospective relationship between epigenetic age acceleration and cardiometabolic parameters measured two years later. METHODS AND RESULTS: ), fasting insulin (β = 0.86 μIU/mL), and insulin resistance (β = 0.21). Skin-Blood acceleration was associated with decreased HDL in males, and PedBE acceleration was associated with triglycerides in both sexes, though in opposing directions. CONCLUSION: Adolescent diet was not strongly associated with baseline epigenetic age acceleration. However, epigenetic age acceleration was associated prospectively with fat distribution and insulin resistance.
Environmental Epidemiology · 2026-01-22
articleOpen accessBackground: Heavy metals, like lead, arsenic, and cadmium, are linked to increased inflammation even in early life; however, anti-inflammatory diets may offset these effects. Methods: We evaluated associations between heavy metals and inflammatory biomarkers, and potential effect modification by diet in 399 adolescents aged 10 to 18 at baseline who attended two study visits 2 years apart (773 observations). At baseline, blood lead, urinary arsenic, and cadmium concentrations were measured, and the Children’s Dietary Inflammatory Index (C-DII) was calculated. Fasting serum interleukin (IL)-4, IL-10, IL-1β, IL-6, IL-8, high-sensitivity C-reactive protein, and tumor necrosis factor-alpha were measured for both visits. Generalized estimating equation models were fit to assess associations between metal concentrations and repeatedly measured inflammatory biomarkers, adjusting for study visit and baseline sociodemographic and lifestyle variables. Effect modification by diet was assessed by including metal and C-DII tertiles interaction terms. Results: At baseline, the median age was 13.6 years, 50.4% were females, and 51.1% had a low socioeconomic status. Overall, there were no associations between metals and inflammatory markers in the entire population. In the most anti-inflammatory diet group (C-DII T1), higher blood lead was associated with higher IL-4, IL-1β, IL-6, and IL-8 levels, whereas for the most pro-inflammatory diet (C-DII T3), these associations were inverse ( P -trend <0.05). Contrarily, higher urinary arsenic was associated with lower IL-4, IL-6, and IL-8 levels in the anti-inflammatory diet group and positively associated with these cytokines in the pro-inflammatory diet group ( P -trend <0.05). Conclusion: Anti-inflammatory diets may modify adolescents’ inflammatory response to lead and arsenic. Heavy metal toxicity mitigation by anti-inflammatory diets requires further research.
Archives of Medical Research · 2026-03-31
articleOpen accessBACKGROUND: Pregnancy induces significant changes in calcium metabolism and bone turnover. Genetic variation in estrogen (ESR1) and vitamin D (VDR) receptors may modulate maternal skeletal adaptation, but evidence in pregnancy and the postpartum period is limited. AIMS: To evaluate the association between single nucleotide polymorphisms (SNPs) in the ESR1 and VDR genes and maternal bone resorption during pregnancy and early postpartum in a Mexican cohort. METHODS: We analyzed data from 575 women participating in the Early Life Exposures in Mexico to ENvironmental Toxicants (ELEMENT) cohort. Urinary N-telopeptide of type I collagen (NTx) was measured in each trimester and at one month postpartum to assess bone resorption. Six SNPs were genotyped (ESR1: rs9340799, rs2234693, and rs3798577; VDR: rs731236, rs1544410, and rs7975232) and associations were estimated using random-effects regression and generalized additive models, adjusting for age, primigravidity, hematocrit, calcium and energy intake, breastfeeding status, and pregnancy stage. RESULTS: The rs731236 CC genotype (VDR, TaqI) was associated with a significantly lower NTx level (≈ 24% reduction; p = 0.005), indicating a protective effect. None of the other loci were found to have a significant effect on bone resorption. CONCLUSION: Genetic variation in VDR (rs731236) was associated with lower maternal bone resorption during pregnancy and early postpartum. This finding highlights the potential relevance of specific genetic markers in understanding individual susceptibility to pregnancy-related bone loss.
Environmental Research · 2026-03-25
articleSenior authorSleep Medicine · 2026-02-05
articleOpen accessOBJECTIVES: Circadian disruption has been linked to adverse metabolic health. Adolescents are particularly susceptible to circadian disruptors, such as delayed sleep onset and social jetlag, which may have sex-specific effects. However, evidence linking these disruptors with circadian gene expression and subsequent cardiometabolic risk remains limited. METHODS: Our study included 203 adolescents (53% females, median age 13.6 years) from the ELEMENT cohort in Mexico City. Sleep was assessed via 7-day wrist actigraphy. A fasting venipuncture blood sample was collected between 8:00 a.m. and 12:00 p.m. RNA was isolated from blood leukocytes and sequenced to determine the relative expression of genes. We conducted differential gene expression analysis for 12 core clock genes in relation to sleep midpoint and social jetlag, adjusting for sleep duration and other potential confounders. We further evaluated how circadian gene expression associated with changes in adiposity, glucose metabolism, blood pressure, and lipid profiles over two years using linear regression. RESULTS: Later sleep midpoint (per 1-h increase) was associated with reduced mid-morning expression of four circadian genes: RORA (log2 fold change [LFC]: -0.190; P value: 0.001), RORC (LFC: -0.147; P value: 0.039), CLOCK (LFC: -0.141; P value: 0.019), and NR1D2 (LFC: -0.093; P value: 0.029). Additionally, expression levels of several clock genes (CRY1, NR1D2, BMAL1, and PER1-3) were associated with changes in metabolic biomarkers over two years in sex-specific patterns. For instance, NR1D2 showed a negative association with fasting glucose among females (β: -0.0012; P value: 0.020), while demonstrating positive associations with LDL cholesterol (β: 0.0023; P value: 0.002) and total cholesterol (β: 0.0016; P value: 0.028) among males. CONCLUSIONS: Expression of core clock genes was linked to circadian disruption and changes in cardiometabolic risk factors in a sex-specific manner among adolescents. Our findings provide novel insights into potential biological mechanisms underlying associations of circadian disruption with cardiometabolic health.
International Journal of Environmental Research and Public Health · 2025-04-26
articleOpen accessDental fluorosis indicates past fluoride intake. People living in Mexico City are exposed to fluoridated salt, which contributes significantly to fluoride intake. This study aimed to (1) estimate fluoride intake during early childhood and fluorosis prevalence in permanent teeth in adolescence and (2) identify intake windows associated with higher fluorosis scores in upper central incisors (UCIs). We included 432 participants from the ELEMENT project (Early-Life Exposures in Mexico to Environmental Toxicants), with data on fluoride intake at ages 1-5 and fluorosis (TFI) at adolescence. Median intakes ranged from 0.56 at age 1 to 1.14 mg/day at age 5, exceeding recommendations. All adolescents had some level of fluorosis, predominantly mild (62% with TFI 2). For every 0.1 mg of daily fluoride intake at age 1, the odds of higher TFI in UCIs were 1.08 [95% CI: 1.00-1.17]. At age 2, the odds were marginally significant [OR: 1.07; 95% CI: 1.00-1.16]. In conclusion, for participants of ELEMENT: (1) fluoride intake during early childhood exceeded recommendations and the prevalence of mild fluorosis in adolescence was high, and (2) fluorosis in UCIs was associated with dietary exposure during the first two years of life and may be used in future ELEMENT studies as exposure biomarkers.
Uncertainty Quantification in Epigenetic Clocks via Conformalized Quantile Regression
Genetic Epidemiology · 2025-03-27 · 1 citations
articleOpen accessDNA methylation (DNAm) is a chemical modification of DNA that can be influenced by various factors, including age, the environment, and lifestyle. An epigenetic clock is a predictive tool that measures biological age based on DNAm levels. It can provide insights into an individual's biological age, which may differ from their chronological age. This difference, known as the epigenetic age acceleration, may reflect health status and the risk for age-related diseases. Moreover, epigenetic clocks are used in studies of aging to assess the effectiveness of antiaging interventions and to understand the underlying mechanisms of aging and disease. Various epigenetic clocks have been developed using samples from different populations, tissues, and cell types, typically by training high-dimensional linear regression models with an elastic net penalty. While these models can predict mean biological age based on DNAm with high precision, there is a lack of uncertainty quantification which is important for interpreting the precision of age estimations and for clinical decision-making. To understand the distribution of a biological age clock beyond its mean, we propose a general pipeline for training epigenetic clocks, based on an integration of high-dimensional quantile regression and conformal prediction, to effectively reveal population heterogeneity and construct prediction intervals. Our approach produces adaptive prediction intervals not only achieving nominal coverage but also accounting for the inherent variability across individuals. By using the data collected from 728 blood samples in 11 DNAm data sets from children, we find that our quantile regression-based prediction intervals are narrower than those derived from conventional mean regression-based epigenetic clocks. This observation demonstrates an improved statistical efficiency over the existing pipeline for training epigenetic clocks. In addition, the resulting intervals have a synchronized varying pattern to age acceleration, effectively revealing cellular evolutionary heterogeneity in age patterns in different developmental stages during individual childhoods and adolescent cohort. Our findings suggest that conformalized high-dimensional quantile regression can produce valid prediction intervals and uncover underlying population heterogeneity. Although our methodology focuses on the distribution of measures of biological aging in children, it is applicable to a broader range of age groups to improve understanding of epigenetic age beyond the mean. This inference-based toolbox could provide valuable insights for future applications of epigenetic interventions for age-related diseases.
Pediatric Obesity · 2025-10-21 · 1 citations
articleOpen accessBACKGROUND: Adequate sleep duration is a prevention factor for paediatric obesity, but the role of timing is still unclear. OBJECTIVES: To investigate associations of sleep duration and timing with BMI in a large paediatric database. METHODS: Medical chart and survey data were collected from 29 409 children aged 2-18 years who attended well-child visits between Jan 2019 and Dec 2023 (repeated-measures cross-sectional design). Logistic and linear mixed effects regression models accounting for repeated measures estimated odds of overweight/obesity and continuous BMI-for-age CDC-based percentiles for each additional/later hour of sleep duration, midpoint (median of bedtime and wake time), and bedtime, adjusted for potential confounders and stratified by age groups. RESULTS: Among young children (2-5 years), shorter sleep duration but not sleep timing was related to higher odds of overweight/obesity (21% higher odds with 95% CI: 3% to 36%). In mid-childhood (6-12 years), shorter sleep duration and later midpoint were associated with higher odds of overweight/obesity (18%, 95% CI = 9%, 26%; 32%, 95% CI = 17%, 49%). Among adolescents (13-18 years), each hour of later sleep midpoint equated to 12% higher odds of living with overweight/obesity (95% CI: 1% to 23%). Linear models were similar. CONCLUSIONS: Shorter sleep duration at younger ages and later sleep timing in adolescence were associated with higher BMI.
Recent grants
NIH · $1.5M · 2017–2022
NIH · $3.4M · 2006
NIH · $10.9M · 2010–2030
NIH · $358k · 2017–2022
NIH · $5.7M · 2013
Frequent coauthors
- 249 shared
Martha María Téllez‐Rojo
Instituto Nacional de Salud Pública
- 183 shared
Howard Hu
Keck Hospital of USC
- 158 shared
Alejandra Cantoral
Ibero American University
- 123 shared
Brisa N. Sánchez
Drexel University
- 102 shared
John D. Meeker
University of Michigan–Ann Arbor
- 85 shared
Adrienne S. Ettinger
Rutgers, The State University of New Jersey
- 78 shared
Erica C. Jansen
- 76 shared
Dana C. Dolinoy
- Resume-aware match score
- Save to shortlist
- AI-drafted outreach
See your match with Karen E. Peterson
PhdFit ranks faculty by your research interests, methods, and publications — grounded in their actual work, not templates.
- Free to start
- No credit card
- 30-second signup