About
Kathryn L. Lunetta, PhD, is a Professor of Biostatistics and a statistical geneticist with more than 20 years of experience in statistics and genetics research. She has extensive experience with the design and analysis of genome-wide association studies and sequencing studies, collaborating on several large, national, and international genetic epidemiology studies at Boston University and with the Framingham Heart Study. Her methodological research focuses on statistical problems and study design and analysis issues in human genetics, particularly in the mapping of complex genetic diseases. Her collaborative work spans multiple fields, including the genetics of atrial fibrillation, healthy aging, reproduction, and Alzheimer’s disease. She is a member of the CHARGE consortium and participates on the Steering and Analysis committees, as well as a TOPMed investigator and analysis committee member. She serves as senior/primary statistician on several Alzheimer’s-related projects and for the Framingham Heart Study Atrial Fibrillation working groups, advising on methodology and providing analysis supervision. Additionally, she is a multi-Principal Investigator of a project investigating immune cell and brain aging biomarkers related to cognitive aging.
Research topics
- Biology
- Genetics
- Medicine
- Internal medicine
- Bioinformatics
- Computational biology
- Evolutionary biology
- Gerontology
- Computer Science
- Physiology
- Pathology
- Endocrinology
- Neuroscience
- Psychology
- Demography
- Psychiatry
Selected publications
Thymic health consequences in adults
Nature · 2026-03-18 · 3 citations
articleOpen accessThe thymus is essential for establishing T cell diversity early in life, but undergoes profound involution with age and has therefore traditionally been regarded as largely nonfunctional in adults1,2. Here we propose that preserving thymic functionality is integral to adult health and longevity. We developed a deep learning framework to quantify thymic health from routine radiographic images and evaluated its association with longevity and risk of major age-associated diseases in two large prospective cohorts of asymptomatic adults: the National Lung Screening Trial (n = 25,031) and the Framingham Heart Study (n = 2,581). In both cohorts, thymic health varied markedly across the population. In the National Lung Screening Trial, higher thymic health was consistently associated with lower all-cause mortality, reduced lung cancer incidence and lower cardiovascular mortality over 12 years of follow-up after adjustment for age, sex, smoking and comorbidities. In the independent Framingham Heart Study cohort, higher thymic health was significantly associated with reduced cardiovascular mortality, independent of age, sex and smoking. Thymic health was further linked to systemic inflammation and metabolic dysregulation, and associated with modifiable lifestyle factors including smoking, obesity and physical activity. Together, these findings reposition the thymus as a central regulator of immune-mediated ageing and disease susceptibility in adulthood, highlighting its potential as a target for preventive and regenerative strategies to promote healthy ageing and longevity. Assessing thymic function and health has highlighted the lifelong importance of the thymus as an organ that could be targeted to improve health outcomes, protect against disease and promote healthy ageing.
Alzheimer s & Dementia · 2026-04-28
articleOpen accessINTRODUCTION: The relationship between peripheral vascular health, including endothelia, cognitive decline, and Alzheimer's disease (AD) dementia risk is unclear. METHODS: In this study, 2844 dementia-free Framingham Offspring participants (mean age 60.6 years, 53.2% women) had baseline brachial artery flow-mediated dilation (FMD%) and reactive hyperemia (RH). Participants were then followed for a median of 17 years for incident AD and underwent plasma biomarker testing and brain magnetic resonance imaging. RESULTS: FMD% (hazard ratio [HR] = 0.83, 95% confidence interval [CI] 0.76 to 0.91, p < 0.001) and RH (HR = 0.89, 95% CI 0.79 to 0.99, p = 0.049) were negatively associated with incident AD dementia after adjusting for confounders. Associations were stronger in individuals with elevated C-reactive protein. Poor vascular function correlated with higher plasma AD biomarkers, smaller brain volumes, greater white matter injury, and increased cerebral microbleeds. DISCUSSION: Poor FMD% and RH may serve as a prognostic biomarker for cerebrovascular pathology, including endothelial dysfunction in the AD brain.
medRxiv · 2025-09-27
preprintOpen access-independent multi-ancestry AD PRS using genome-wide association study summary statistics from cohorts in the United States, Europe and East Asia that were applied to European ancestry (EA), African American (AA), Caribbean Hispanic (CH), and East Asian cohorts from the Alzheimer's Disease Genetics Consortium. PRS performance was evaluated in the multi-ancestry Alzheimer's Disease Sequencing Project (ADSP) dataset and validated in several additional multi-ancestry cohorts. The PRS was significantly associated with AD in the ADSP EA, AA, CH, and Native American Hispanic groups with adjusted odds ratios (ORs) between 1.14 and 1.52 per standard deviation of the PRS. PRS performance was validated in the replication cohorts (ORs 1.21-1.65). The PRS was also associated with poorer memory, executive function, and language performance; greater AD-related neuropathological burden (including CERAD, Braak stage, and Thal phase scores); reduced hippocampal volume; lower CSF Aβ42; and elevated total tau and phosphorylated tau (p-tau), with stronger p-tau associations observed in women. Longitudinal analyses revealed that individuals in the highest PRS decile exhibited the steepest cognitive decline, particularly among those who progressed to AD. Our findings demonstrate the utility of an ancestry-aware and APOE-independent PRS for advancing understanding of the genetic basis of AD across diverse populations. Associations observed with early biological and cognitive changes and potential sex-specific differences support the incorporation of a PRS in clinical trials and personalized intervention and prevention strategies.
Alzheimer s & Dementia · 2025-12-01
articleOpen accessAbstract Background Several viruses, including HSV‐1, can establish latent infections in neurons and have been linked to Alzheimer disease (AD). Method Whole genome sequence reads derived from brain (561 AD cases, 259 controls) and blood (5,033 AD cases, 4,907 controls) specimens obtained from European ancestry, African American, Native American Hispanic, Indian, and Caribbean Hispanic participants in the Alzheimer's Disease Sequencing Project that did not align to the human reference genome were aligned to viral reference genomes. We evaluated the association of HSV‐1 DNA presence with harmonized measures of cognitive performance in four cognitive domains (memory, executive functioning, language, and visual‐spatial perception) using regression models with covariates for AD diagnosis, APOE genotype, age, sex, tissue source, ancestry, and PCR amplification. One model evaluated the interaction between HSV‐1 and age on cognitive decline using Generalized Estimating Equations to account for repeated measures. A second model considered only the most recent cognitive scores. Both models were performed separately within AD cases, mild cognitively impaired (MCI), and controls. Result In the cross‐sectional; model, HSV‐1 was associated with lower memory performance in AD cases (b= ‐0.082, p = 0.002). HSV‐1 was significantly associated with decline across multiple domains, including executive function (b = ‐0.032, p = 0.012), language (b = ‐0.028, p = 0.027), and memory (b = ‐0.033, p = 0.008). Among AD cases, HSV‐1 was associated with reduced memory scores (b= ‐0.10, p = 4.35x10 ‐6 ). The HSV‐1 × age interaction term was significant for executive function (b = ‐0.003, p = 0.02) and memory (b = ‐0.003, p = 0.02) domains. In individuals with MCI (b = 0.012, p = 0.046) and controls (b = 0.091, p = 0.016), HSV‐1 was associated with higher visual‐spatial scores. Conclusion These findings provide further evidence for the role of HSV‐1 presence in AD pathogenesis and may exacerbate cognitive decline in individuals with AD.
medRxiv · 2025-12-17
preprintOpen accessAbstract Background Telomere length (TL), a biomarker of biological aging, but its association with Alzheimer’s disease (AD) remains unclear. Methods We estimated TL in whole-genome sequencing data from 35,014 Alzheimer’s Disease Sequencing Project participants using TelSeq, which after quality control yielded a dataset including 6,973 persons of European ancestry (EA), 4,188 African Americans (AA), 4,005 Caribbean Hispanics (CH), and 4,170 Native American Hispanics (NAH). TL was log-transformed, adjusted for age and blood cell counts, and z-scaled. Scaled TL was dichotomized into long and short groups according to the median. An AD GWAS for the interaction of TL with variants having a minor allele count >20 was performed in each ancestry group using logistic regression models including SNP and TL main effects and a SNP×TL interaction term. Results AD risk was associated with shorter TL (β = -0.18, P < 2×10 -16 ). Longer TL was associated with dosages of APOE ε2 ( P <5.08×10 -8 ) and APOE ε4 ( P =2.10×10 -2 ). In the EA group, genome-wide significant (GWS) TLxSNP interactions were identified for variants in SEMA6A (P =1.42×10 -8 ) and LOC105378654 (P =4.17×10 -8 ), between IL15 and INPP4B (P =1.77×10 -8 ) and upstream of RP11-2N5.2 ( P =4.60×10 -8 ). In the NAH group, GWS interactions were observed with an intronic variant in BSN ( P =3.26×10 -8 ) and missense variant in MST1 ( P =3.26×10 -8 ). In the total sample, interactions with variants between CTD-2160D9.1 and EEF1A1P20 ( P <1.19×10 -8 ), in TBC1D22A ( P =1.06×10 -8 ) and in PLK1 ( P =3.28×10 -8 ) were GWS. Conclusion We identified variants that significantly impact AD risk through their interaction with TL, suggesting that TL maintenance pathways may be central to AD pathogenesis.
Common variant genetic background risk impacts Alzheimer's disease rare genetic variant associations
Alzheimer s & Dementia · 2025-12-01
articleOpen accessAbstract Background Polygenic risk scores (PRS) are measures of an individual's aggregate genetic risk of disease and are typically calculated using common variants. Rare variants (RV), though not captured by PRS, have been shown to play an important role in AD. We utilized an AD PRS to discover associations with novel RVs. Method PRS for non‐Hispanic white (NHW) samples ( n = 11,409; AD=6630, controls=4779) from the Alzheimer's Disease Sequencing Project (ADSP) R4 dataset were calculated using published GWAS summary statistics of NHW subjects, excluding the APOE region. Participants were classified into high ( n = 5,442) and low ( n = 5,967) PRS groups based on the median PRS. After quality control, 11,485,531 low frequency variants (LV) (1% < MAF < 5%) and RVs with a MAC>5 were included in the analysis. The association of each variant was evaluated in low and high PRS groups separately using a regression model including covariates for age, sex, and principal components of ancestry implemented in GENESIS. Result Genome‐wide significant (GWS) associations of RVs and LVs with AD were identified in both the low and high PRS groups. Risk variants were disproportionately enriched in the low group while protective ones were disproportionately enriched in the high group. Among the GWS, RVs in the coding region of NYNRIN (rs570910195) (β=4.06, p = 4.37x10 ‐10 ) and 52 kb upstream of LOC105373398 (rs572619714, β=1.67, p = 2.20x10 ‐8 ) increased AD risk in the low PRS group. A rare deletion located 27 kb downstream of SDHC (chr1:161390036) was associated with decreased AD risk (β=‐2.33, p = 7.45x10 ‐9 ) and a low frequency variant 20 kb downstream of SMNDC1 (rs150913764, MAF=0.02) was associated with increased AD risk (β=0.96, p = 3.13x10 ‐8 ) in the high PRS group. These variants were not associated with AD in the other respective PRS group. Conclusion Stratifying individuals into those with lower and higher risk for AD based on aggregated effects of common variants enhanced the detection of RV and LV associations. Our results suggest that a PRS calculated from common variants may not necessarily correspond to an individual's genetic relative risk for AD. These findings also provide unique opportunities to study RVs whose effects are opposite to the risk conferred by the genetic background.
Alcohol consumption and immune cell profiles: Insights from the Framingham Heart Study
Alcohol Clinical and Experimental Research · 2025-07-29
articleOpen accessBACKGROUND: Alcohol consumption affects immune function, with excessive intake linked to immune suppression and inflammation. However, its impact on immune cell phenotypes remains unclear. This study investigates the association between alcohol consumption and immune cell profiles in a well-characterized Framingham Heart Study (FHS) cohort while examining sex-specific differences in alcohol-immune cell associations. METHODS: We analyzed data from 996 participants of the FHS Offspring cohort who underwent immune cell phenotyping and completed an alcohol questionnaire during Exam 7 (1998-2001). Alcohol intake was categorized as abstainer, moderate, at-risk, or heavy drinking. Linear mixed-effects models examined associations between alcohol intake and 15 immune cell phenotypes, adjusting for age, sex, and Cytomegalovirus (CMV) (Model 1) and additional covariates (Model 2). False discovery rate (FDR) correction was applied for multiple testing. RESULTS: The CD4+ Tn/Tm ratio showed a significant nonlinear relationship with alcohol categories in Model 1 (p = 0.002, FDR = 0.03), with higher ratios in moderate (β = 0.26) and at-risk drinkers (β = 0.26) compared with abstainers; effects were smaller in Model 2 (β = 0.23 and β = 0.23, respectively). Sex-stratified analyses revealed that among males, alcohol consumption was associated with several immune cell phenotypes in Model 1, and with CD8+ Tn/Tm ratio in Model 2 (p = 0.0001, FDR = 0.002), where moderate drinking was associated with higher CD8+ Tn/Tm ratio compared with abstainers (β = 0.29). Among male drinkers, consumption level was also associated with CD8+ Tn/Tm ratio in both models: at-risk and heavy consumption showed significantly lower CD8+ Tn/Tm ratio compared with moderate drinkers (β = -0.43 and β = -0.46, respectively, in Model 2). CONCLUSIONS: Alcohol consumption exhibits a nonlinear relationship with certain immune cells, with moderate intake potentially benefiting immunity, while higher consumption may compromise immune homeostasis. Given the study's cross-sectional design, causality cannot be inferred; nonetheless, our sex-specific, dose-dependent findings merit confirmation in longitudinal cohorts.
Alzheimer s & Dementia · 2025-12-01
articleOpen accessAbstract Background Genome‐wide association studies (GWAS) have identified over 1,000 blood pressure (BP) loci and over 80 Alzheimer's disease (AD) loci. Considering BP is an AD risk factor, identifying pleiotropy in BP and cognitive performance may indicate mechanistic links between BP and AD. Method GWAS for pleiotropy in four BP variables—systolic (SBP), diastolic (DBP), mean arterial (MAP), pulse pressure (PP)—and co‐calibrated scores for cognitive domains (executive function, language, memory) were performed using generalized linear mixed models and 116,075 longitudinal measures from 25,726 participants of clinic‐based and prospective cohorts. Pleiotropy GWAS was conducted using PLACO to estimate SNP's main effect, SNP×age interaction, and their joint effect on pleiotropy. Effects of genome‐wide significant (GWS) pleiotropic SNPs on cognition as direct or mediated through BP were evaluated using Mendelian randomization. Potential contribution of genes in top‐ranked pleiotropic loci to cognitive resilience was assessed by comparing their expression in brain tissue from pathologically confirmed AD cases with and without clinical symptoms. Result Pleiotropy GWAS identified GWS associations with APOE and 11 novel loci. GWS pleiotropy was identified for SBP and language with SNPs in JPH2 ( P Joint =6.09×10 ‐9 ) and GATA3 ( P G×Age =1.42×10 ‐8 ) in the total sample and with RTN4 ( P G×Age =1.49×10 ‐8 ) in prospective cohorts. GWS pleiotropy was observed for PP and language with ADAMTS3 ( P G =2.37×10 ‐8 ) in clinic‐based cohorts. ULK2 was pleiotropic for DBP and executive function ( P Joint =2.85×10 ‐8 ) in prospective cohorts. In the total sample, PAX2 ( P G×Age =4.22×10 ‐8 ) and LOC105371656 ( P G×Age =1.75×10 ‐8 ) had GWS pleiotropy for MAP paired with executive function and language, respectively, and SUFU was pleiotropic for DBP and language ( P G =2.10×10 ‐8 ). LINC02946 was associated with memory paired with SBP ( P G×Age =3.47×10 ‐8 ) in clinic‐based cohorts. GWS pleiotropy was found in prospective cohorts for PP and memory with SORBS2 ( P G =2.33×10 ‐8 ) and for DBP and memory with LOC100128993 ( P G×Age =2.81×10 ‐8 ). Five of the GWS pleiotropic loci influence cognition directly. Genes at six pleiotropic loci were differentially expressed between pathologically confirmed AD cases with and without clinical symptoms. Conclusion Our results provide insight into the underlying mechanisms of BP and AD. Ongoing efforts to harmonize BP and cognitive measures across several cohorts will improve the power of discovering, replicating, and generalizing novel associations with pleiotropic loci.
Translational Psychiatry · 2025-08-19 · 2 citations
articleOpen accessMonocyte chemoattractant protein-1 (MCP-1), a cytokine involved in peripheral inflammation, has been shown to modulate established Alzheimer’s disease (AD) loci. In this study, we hypothesized that blood MCP-1 levels may impact the associations of other genetic variants with AD risk beyond the well-established AD loci. We performed a genome-wide association study (GWAS) using logistic regression with the generalized estimating equation (GEE) and Cox proportional hazards models to examine the combined effects of single nucleotide polymorphisms (SNPs) and blood MCP-1 levels on AD. Three datasets were used: the Framingham Heart Study (FHS), Religious Orders Study/Memory and Aging Project (ROSMAP), and Alzheimer’s Disease Neuroimaging Initiative (ADNI). We identified SNPs in two genes in the meta-analysis, namely, neuron navigator 3 (NAV3, also named unc-53 homolog 3, rs696468) (p < 7.55 × 10−9) and the homolog unc-5 netrin receptor c (UNC5C rs72659964) (p < 1.07 × 10−8), which are modified by blood MCP-1 concentration for AD risk. Elevated blood MCP-1 concentrations increased AD risk and brain AD pathology in individuals with NAV3 (rs696468-CC) and UNC5C (rs72659964-AT + TT) genotypes. Given that NAV3 and UNC5C are involved in regulating neurite outgrowth and guidance, increased MCP-1 levels may disturb the functions of vulnerable gene carriers to increase AD risk.
Learning Gaussian graphical models from correlated data
Frontiers in Systems Biology · 2025-07-03 · 1 citations
articleOpen accessGaussian Graphical Models (GGMs) are a type of network modeling that uses partial correlation rather than correlation for representing complex relationships among multiple variables. The advantage of using partial correlation is to show the relation between two variables after "adjusting" for the effects of other variables and leads to more parsimonious and interpretable models. There are well established procedures to build GGMs from a sample of independent and identical distributed observations. However, many studies include clustered and longitudinal data that result in correlated observations and ignoring this correlation among observations can lead to inflated Type I error. In this paper, we propose a cluster-based bootstrap algorithm to infer GGMs from correlated data. We use extensive simulations of correlated data from family-based studies to show that the proposed bootstrap method does not inflate the Type I error while retaining statistical power compared to alternative solutions when there are sufficient number of clusters. We apply our method to learn the GGM that represents complex relations between 47 Polygenic Risk Scores generated using genome-wide genotype data from the Long Life Family Study. By comparing it to the conventional methods that ignore within-cluster correlation, we show that our method controls the Type I error well without power loss.
Recent grants
NIH · $474k · 2018
IDENTIFICATION OF COMMON GENETIC VARIANTS FOR ATRIAL FIBRILLATION AND PR INTERVAL
NIH · $13.6M · 2009–2026
IDENTIFICATION OF COMMON GENETIC VARIANTS FOR ATRIAL FIBRILLATION AND PR INTERVAL
NIH · $7.0M · 2009–2018
Role of Peripheral Immune Cells in Cognitive Aging: The Framingham Offspring Study
NIH · $3.4M · 2020–2026
NIH · $35.9M · 2018
Frequent coauthors
- 730 shared
Emelia J. Benjamin
Boston Medical Center
- 532 shared
Patrick T. Ellinor
- 508 shared
Steven A. Lubitz
Broad Institute
- 368 shared
Joanne M. Murabito
- 313 shared
Lu‐Chen Weng
Brigham and Women's Hospital
- 297 shared
Honghuang Lin
University of Massachusetts Chan Medical School
- 295 shared
Daniel Levy
National Heart Lung and Blood Institute
- 289 shared
André G. Uitterlinden
Education
- 1996
PhD, Biostatistics
University of Michigan
- 1993
MS, Biostatistics
University of Michigan
- 1989
BA
Grinnell College
- Resume-aware match score
- Save to shortlist
- AI-drafted outreach
See your match with Kathryn L. Lunetta
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