Mapping the human epigenetic landscape across three generations: A DNA methylation resource from TMM BirThree

Mapping the human epigenetic landscape across three generations: A DNA methylation resource from TMM BirThree | 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 Article Mapping the human epigenetic landscape across three generations: A DNA methylation resource from TMM BirThree Atsushi Shimizu, Shiori Minabe, Hideki Ohmomo, Akira Takashima, and 28 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7314319/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Transgenerational epigenetic inheritance has been demonstrated in rodent models but not in humans. To address this gap, we established a comprehensive DNA methylation resource derived from 158 three-generation Japanese families. The dataset integrates genome-wide methylation profiles with extensive clinical and lifestyle data. Using targeted bisulfite sequencing, we profiled > 1 million CpG sites across the genome, covering the promoter and gene body regions of > 16,500 annotated genes, in 938 adult peripheral blood and 155 neonatal cord blood samples. To demonstrate the utility of this resource, we performed a representative analysis focusing on the intergenerational impact of maternal and grandmaternal pre-pregnancy smoking. We identified persistent methylation marks in neonates associated with ancestral smoking history, suggesting the potential transgenerational transmission of environmental effects in humans. This multigenerational epigenomic resource provides a valuable foundation for future studies on intergenerational epigenetic mechanisms and their role in shaping human health trajectories. Health sciences/Risk factors Health sciences/Biomarkers/Predictive markers Biological sciences/Genetics/Epigenetics/DNA methylation Biological sciences/Genetics/Epigenomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Understanding the origins of complex human diseases requires an integrated view of genetic inheritance and environmental exposure across the life course. The Developmental Origins of Health and Disease (DOHaD) concept has been well-established through numerous epidemiological studies linking early-life environments to adult disease risk 1 – 3 . Seminal human studies 4 , 5 , 6 have suggested that prenatal and even ancestral exposures can influence offspring and grand-offspring health outcomes. DNA methylation is thought to mediate these associations, at least in part 6 , 7 . Epigenetic mechanisms offer a plausible molecular basis for how early-life environmental cues become embedded and potentially transmitted across generations. Animal studies have demonstrated that prenatal exposure to endocrine disruptors 8 , nutritional restriction 9 , or maternal stress 10 can induce heritable methylation changes, providing evidence for transgenerational epigenetic inheritance. However, in humans, molecular evidence for such inheritance remains scarce, largely because of the difficulty of conducting large-scale and long-term cohort studies that can integrate multigenerational exposure and epigenomic data by collecting biological and socioeconomic information across generations. However, these studies face substantial logistical, financial, and ethical barriers 11 . Large national studies were eventually terminated owing to challenges in participant recruitment and the complexity of the study design 12 . The Tohoku Medical Megabank Project’s Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study) has been established as a prospective genomic cohort to support precision medicine and long-term health monitoring in post-disaster Japan 13 . Through longitudinal collection of biospecimens and detailed metadata on medical history, lifestyle, diet, and psychosocial stress, the TMM BirThree cohort provides a framework for investigating how early-life and ancestral exposure may influence molecular phenotypes, including DNA methylation. Its multigenerational structure and deep phenotyping make it particularly well-suited for examining research questions aligned with the DOHaD concept. To investigate the molecular basis of epigenetic inheritance and the DOHaD concept in humans, we established two types of DNA methylation resources: a three-generation dataset from 158 Japanese families (the Hepta-family dataset) 13 and a population-specific reference panel based on whole-genome bisulfite sequencing (WGBS) of nucleated red blood cells (nRBCs) from Japanese newborns used to develop a cell-type composition estimation model. The Hepta-family dataset includes methylation profiles obtained by targeted bisulfite sequencing from newborns (umbilical cord blood [CB]), their parents, and all four grandparents: maternal grandmother (MGM), maternal grandfather (MGF), paternal grandmother (PGM), and paternal grandfather (PGF), with peripheral blood (PB) used for all samples except newborns. To demonstrate the utility of this multigenerational resource, we performed a representative analysis focusing on the epigenetic impact of maternal and grandmaternal pre-pregnancy smoking, an environmental exposure with well-established and reproducible methylation signatures 14 . This analysis suggests that the epigenetic response to smoking may vary according to developmental stage and tissue context, highlighting the importance of considering stage-specific effects when investigating transgenerational epigenetic signals. Results Structure and characteristics of the three generation Hepta family The dataset comprised 158 mothers, 158 fathers, 157 MGMs, 156 MGFs, 155 PGMs, 154 PGFs, and 155 neonates (Fig. 1 ). Among them, 148 families were complete with all 7 members available for analysis. One family included twins, resulting in an eight-member structure (Supplementary Fig. 1). In addition, the dataset included 154 maternal-neonate pairs, 468 spousal pairs across three generations, and 464 parent–child trios. These trios comprised paternal (father, paternal grandparent, and neonate) and maternal (mother, maternal grandparent, and neonate) configurations (Supplementary Fig. 1). Table 1 summarizes the characteristics of the newborns based on the maternal and neonatal medical records at admission and parent-administered questionnaires completed by the mothers. Among the newborns, 54.2% were male. The mean gestational age at birth was 39.2 ± 1.4 weeks, and the mean birth weight was 3,078.0 ± 397.6 g. The proportion of low birthweight, defined by the World Health Organization as a birthweight of an infant of ≤ 2,499 g regardless of gestational age 15 , was 8.4% (13 of 155 newborns). The characteristics of the pregnant women (neonate mothers), fathers, and grandparents are shown in Table 2. Regarding household income, 23.4% of mothers reported an annual household income of < 4 million yen, and 35–39.0% of grandparents fell within this category. Household income data were partially missing, particularly for 6.4–7.1% of grandfathers and 15.5–16.6% of grandmothers, and were unavailable for fathers. Data on smoking and alcohol consumption were available for all pregnant mothers (Table 2). Among them, 35 were former smokers and none were current smokers. In contrast, 24.1% of mothers reported current alcohol consumption. Among MGMs, 4.5% were current smokers and 10.2% were former smokers, a trend that was similar among PGMs. Regarding grandfathers, 17.3% of MGFs and 29.9% of PGFs were current smokers, whereas 59.0% and 51.3%, respectively, were former smokers. Sequencing and mapping statistics of the Hepta-family DNA methylation dataset We analyzed > 1 million CpG sites spanning the promoter and gene body regions across > 16,500 annotated genes using the common DNA methylation variations (CDMV) 16 version 3 (CDMVv3) probe set probe set for captured methylation sequencing. The mean sequencing depth across target regions ranged from 26.2 to 26.9×, and ~ 98.0% of the regions were covered at least once, ensuring reliable methylation quantification and demonstrating efficient hybridization and sequencing performance (Supplementary Table 1). The consistency in sequencing depth across generations and between sexes further supports the reproducibility of our method. The detailed statistics are shown in Supplementary Table 1. Principal component analysis (PCA) clustering and biological annotation in the Hepta-family DNA methylation dataset PCA was performed using DNA methylation data from all 1,093 individuals included in the Hepta-family dataset. The first and second principal components (PCs) (PC1 and PC2, respectively) explained 10.9% and 5.9% of the variance, respectively, with neonates forming a distinct cluster clearly separated from adult family members (Fig. 2 A). The correlation heatmap in Fig. 2 B was used to assess the potential batch effects by evaluating the associations between PCs and various factors, including age, sample type (CB vs. PB), flow cell ID, sequencing batch, and sex. The results indicated no significant batch effects, confirming the robustness of the dataset for subsequent analyses. To further explore the biological relevance of the major axes of variation, PC1 and PC2, (Supplementary Fig. 2A), we performed gene enrichment analysis on CpG sites with high loading scores (≥ 0.4). Among the 823,553 CpG sites analyzed (call rate [CR] = 100%), 162,802 CpG sites for PC1 and 81,574 for PC2 met the cutoff (Supplementary Fig. 2B). The CpG sites contributing to PC1 were enriched in genes associated with neurodevelopmental processes, including axonogenesis, regulation of nervous system development, and sensory system formation (Supplementary Fig. 2C, upper ). Tissue enrichment analysis further supported these findings, revealing a strong overrepresentation of genes specifically expressed in the brain (Supplementary Fig. 2D, upper ). In contrast, CpG sites contributing to PC2 were enriched in genes involved in immune-related processes, such as lymphocyte differentiation, leukocyte activation, and cell–cell adhesion (Supplementary Fig. 2C, lower ). These findings were further supported by tissue enrichment analysis, which revealed a significant overlap with genes specifically expressed in the lymph nodes, spleen, and bone marrow (Supplementary Fig. 2D, lower ), indicating that PC2 captures immunologically driven interindividual epigenetic variability. Global and regional methylation levels across generations To investigate generational differences in DNA methylation levels, we compared mean methylation rates across newborns, parents, and grandparents using CpG sites with a CR ≥ 95% (Fig. 3 ). Significant generational differences in global methylation levels were observed (adjusted p < 0.001). Newborns had significantly higher global methylation levels than both parents and grandparents (adjusted p < 0.001), and a smaller yet significant difference was observed between parents and grandparents (adjusted p < 0.001). Significant effects of generation were also observed in the CpG islands, shores, shelves, and open-sea regions (all adjusted p < 0.001). Development and validation of a cell-type composition estimation method for CB in sequencing-based DNA methylation data To develop a reference panel for cell-type composition estimation in sequencing-based DNA methylation data from CB, nRBCs were isolated from CB samples collected from 24 newborns. All samples underwent cell sorting to isolate CD3⁻CD14⁻CD19⁻CD71⁺CD235⁺ cell populations. Of them, 15 samples passed quality control based on cell purity (≥ 94%) and DNA integrity and were included in downstream analyses. The average maternal age for the 15 newborns (10 males and 5 females) was 35.3 ± 4.8 years, and the mean gestational age at delivery was 37.9 ± 1.0 weeks. Extracted genomic DNA showed high integrity with a mean DNA Integrity Number of 8.2 ± 0.6, and nuclear staining confirmed the identity of the isolated nRBC. WGBS data from 15 newborn-derived nRBC samples revealed approximately 27 million CpG sites (≥ 1× depth) detected on both DNA strands. The raw sequencing depth exceeded 30× coverage, ensuring a sufficient read depth for downstream analysis (Supplementary Table S2 ). The DNA methylation profiles of nRBCs were compared with the WGBS data from eight other blood cell types (monocytes, neutrophils, CD4 + T cells, CD8 + T cells, NK cells, B cells, PB mononuclear cells, and leukocytes, n = 20) previously published in the iMETHYL database 17 . Moreover, WGBS data from erythroblasts (n = 2) obtained from the IHEC Data Portal and DNA methylation profiles of CBs obtained from 155 neonates were generated using targeted bisulfite sequencing in this study. PCA using 9,880 CpG sites detected across all datasets demonstrated that nRBCs formed a distinct cluster, indicating a unique DNA methylation profile (Supplementary Fig. 3). Next, we incorporated the nRBC WGBS data into a reference panel to estimate the cell-type composition in the CB samples. This model was applied to 938 adult PB and 155 CB samples from the Hepta-family cohort. Neutrophils were the predominant cell type in both sample types (69.7 ± 10.3% in PB, 54.9 ± 9.8% in CB). As expected, nRBCs were nearly absent in PB (0.6 ± 0.8%) but present at 12.8 ± 3.6% in CB, and this difference was statistically significant (t-test, p < 0.05) (Supplementary Fig. 4A and 4B). The coefficient sets derived from both CB- and adult-specific models are available in the iMETHYL database. To evaluate the performance of the CB-specific model, we conducted an epigenome-wide association studies (EWAS) using neonatal CB methylation data from the Hepta-family dataset with maternal smoking history as the exposure variable. Compared with the conventional adult-based estimation model (Model 1, Supplementary Fig. 5, right ), the CB-specific model (Model 2, Fig. 4 A, right ) demonstrated improved bias correction, as reflected by reduced genomic inflation factor (λ), which decreased from 1.027 (95% confidence interval [CI], 1.023–1.032) to 1.010 (95% CI, 1.006–1.015). Furthermore, Model 2 identified a greater number of CpG sites surpassing the genome-wide suggestive threshold ( p < 1.0 × 10⁻ 5 ), with 14 CpG sites detected (Fig. 4 A, left ) compared with 9 CpG sites by Model 1 (Supplementary Fig. 5, left ). Demonstration 1: Impact of maternal ancestral smoking history on DNA methylation in neonatal CB Characteristic features related to maternal smoking among the 153 MGM-mother–newborn families with available epigenomic data are summarized in Supplementary Table 3. Among the 153 mothers with available smoking history, 29 quit smoking before pregnancy. Of them, 27 had smoked for > 1 year prior to conception, and two had smoked for < 1 year. The average duration of smoking cessation before pregnancy was 7.1 ± 3.9 years. Information on smoking was available for 146 MGMs, among whom, 23 were identified as ever-smokers (either current or former smokers). The pre-pregnancy smoking status of each MGM was estimated by comparing the reported age at smoking initiation with the calculated age at delivery based on the birth dates of the grandmother and mother. Among the 23 ever-smokers, 16 had initiated smoking > 1 year before pregnancy. Of these, 12 were current smokers, and 4 were former smokers who had quit either before or during pregnancy. The average duration of smoking prior to pregnancy was 5.4 ± 3.3 years, and the mean duration since smoking cessation (for former smokers) was 2.5 ± 1.9 years. Information regarding whether smoking continued throughout pregnancy was unavailable. To investigate the effect of maternal pre-pregnancy smoking on DNA methylation in the offspring, we compared CB DNA methylation profiles (998,877 CpG sites) across three groups: never smokers (n = 114), former smokers who quit within 5 years prior to pregnancy (n = 12), and former smokers who quit ≥ 5 years prior to pregnancy (n = 19). The EWAS identified one CpG site that surpassed the Bonferroni-corrected significance threshold ( p < 5.1 × 10⁻⁸) and 13 CpG sites that reached the suggestive threshold ( p < 1.0 × 10⁻ 5 ) (Fig. 4 A). Among them, six CpG sites mapped to six distinct genes (Supplementary Table 4). Notably, several of these genes were previously implicated in EWAS associated with tobacco exposure. These genes included UBR4 , KIFC3 , MYOD1 , ADORA2A-AS1 , and CACNA1C , indicating their biological relevance 18 . Next, to examine the effect of MGMs’ pre-pregnancy smoking on DNA methylation in grandchildren, we compared the CB DNA methylation profiles (991,676 CpG sites) between newborns whose MGMs had smoked for > 1 year prior to pregnancy (n = 15) and those whose MGMs had never smoked before pregnancy (n = 122) (Fig. 4 B). EWAS did not identify any CpG sites that reached genome-wide significance; however, 19 CpG sites showed suggestive associations ( p < 1.0 × 10⁻ 5 ). Among them, 10 CpG sites mapped to 10 distinct genes (Supplementary Table 5). Several of these genes were previously implicated in EWAS related to tobacco exposure. These included ANXA6 , TERT , HSPG2 , OGDHL , and EHMT1 , indicating potential biological relevance 18 . Demonstration 2: EWAS of smoking and smoking cessation in the grandparent generation To identify smoking-related DNA methylation markers detectable in the Hepta-family dataset, we conducted EWAS using epigenomic data from the grandparent generation. This cohort was selected because it included a substantial number of unrelated individuals and sufficient number of continuous smokers. Of the 622 grandparents, 39 with missing information on smoking status or household income were excluded. The analysis included 86 continuous smokers (Smk group) and 299 nonsmokers (Ctr group), with DNA methylation profiles assessed at 988,567 CpG sites using smoking status as the trait of interest. The EWAS identified 59 significant CpG sites surpassing the Bonferroni-corrected threshold ( p < 5.1 × 10⁻⁸) and 206 CpG sites with suggestive associations ( p < 1.0 × 10⁻ 5 ) (Supplementary Table 6). The Manhattan plot of smoking EWAS (Fig. 5 A) revealed peaks in genomic regions previously associated with smoking, including 2q37.1, 11q13.4, and genes such as GFI1 , KIF5C , EXOC2 , AHRR , ALPP , CNTNAP2 , MGAT3 , PLAT , MYOM1 , and F2RL3 18,19 . To explore the long-term epigenetic legacy of smoking after cessation, we conducted EWAS using a three-category variable for smoking status. The categories were never smokers (n = 302), former smokers who quit within 5 years (n = 31), and those who quit ≥ 5 years ago (n = 167). DNA methylation was assessed at 989,219 CpG sites. The EWAS identified three significant CpG sites surpassing the Bonferroni-corrected threshold ( p < 5.1 × 10⁻⁸) and 64 suggestive CpG sites ( p < 1.0 × 10⁻ 5 ) (Fig. 5 B; Supplementary Table 7). Manhattan plots revealed signal CpG sites at 2q37.1 and 11q13.4 loci and within KIF5C genes, consistent with those observed in smoking EWAS (Fig. 5 A, Supplementary Table 6). Demonstration 3: Comparison of smoking-associated CpG sites in PB and CB CpG sites showing differential DNA methylation in neonatal CB from mothers with a history of smoking did not overlap with known smoking- associated or smoking-cessation-associated CpG markers previously identified in adult PB (Fig. 6 A). Similarly, altered CpG sites in neonates whose MGMs had a history of smoking showed no overlap with adult-associated markers. When comparing cell-type-adjusted methylation levels at known smoking-associated CpG sites, we found notable differences between adult PB and neonatal CB. As shown in Fig. 6 B and Supplementary Table 8, F2RL3 , NFE2L2 , RPTOR , and PPP1R15A were hypermethylated in neonatal CB compared with adult PB, whereas 2q37.1, MGAT3 , and EXOC2 were hypomethylated. Similar methylation patterns were observed in publicly available datasets. Using the iMETHYL database, we compared cell-type-adjusted methylation levels between adult PB (control group of the Kidney Cancer study) and gestational age-specific CB samples 20 . Although most smoking-associated CpG sites exhibited age-dependent differences in DNA methylation between newborns and adults, a small subset showed stable methylation patterns regardless of developmental stage. These included the CpG sites located at chr1:92947586 and chr1:92947588 in GFI1 , chr2:149824190 in KIF5C , chr5:373651 in AHRR , chr7:145814670 in CNTNAP2 , chr12:53613078 in RARG , and chr5:17000346 in F2RL3 (Supplementary Table 8). These consistently methylated sites may serve as robust biomarkers of tobacco exposure and may be less influenced by age-related epigenetic variations. Discussion In this study, we constructed a three-generation epigenomic dataset comprising 1,093 samples with detailed family structures and cohort metadata. All experimental steps, including DNA library preparation, targeted bisulfite sequencing, and bioinformatics analysis, were performed at a single facility to minimize technical variability and ensure data consistency across samples. To further enhance the accuracy of EWAS, we generated WGBS data for nRBCs and developed a cell-type composition estimation model specifically for neonatal CB. We also applied an existing cell-type estimation model for adult PB, enabling cell-type adjustment in both neonatal and adult samples derived from the same cohort. Summary data from the Hepta-family dataset have been made publicly available via the iMETHYL database in the form of summary-level epigenomic data, serving as a reference panel for epigenetic studies in Japanese populations. Notably, the demographic and phenotypic characteristics of our participants closely mirror those of the original three-generation cohort 13 , indicating that the dataset is representative of the general Japanese population and suitable for use as a reference in future epigenetic studies. The family-based structure provides a framework for investigating intergenerational effects and environmental exposure within a controlled genetic framework. Substantial differences in DNA methylation profiles were observed between neonatal CB and adult PB. PCA revealed a clear separation between the two. PC1 reflected variations associated with neurodevelopment and organogenesis and was enriched in brain-specific gene sets, whereas PC2 captured immunological diversity associated with lymphoid tissue-expressed genes. These differences were not fully explained by cell-type composition alone, suggesting coordinated, developmentally regulated epigenetic programs. Moreover, neonates showed significantly lower global and region-specific methylation levels, particularly in CpG islands, than parents and grandparents. This widespread hypomethylation, together with functional enrichment patterns, supports the epigenomic plasticity of early development 21 and emphasizes the need to consider the developmental stage and cellular context when interpreting methylation data across generations. Recent studies have further demonstrated that DNA methylation profiles differ markedly between neonatal and adult immune cells, even within the same cell type, reflecting lineage-specific developmental maturation 21 . These findings underscore that even with appropriate cell-type adjustment, comparisons between neonatal and adult samples must be interpreted with caution. Even after adjusting for cell-type composition, methylation levels at almost all smoking marker CpGs, such as those in F2RL3 and 2q37.1 loci 22 , differed markedly between newborns and adults. Moreover, although our dataset replicated several well-established smoking-associated CpG sites 22 found in the PB, the sites associated with maternal or grandmaternal smoking history in the descendants did not overlap with these known markers. These observations suggest that the epigenetic response to smoking exposure differs depending on the developmental stage and tissue type. CpG sites that respond to environmental exposure may vary depending on the tissue and developmental context, and extrapolating adult-derived findings to neonatal samples may lead to misinterpretations. This is particularly important for studies investigating transgenerational epigenetic inheritance. In EWAS using blood-derived DNA methylation data, it is standard practice to adjust for potential confounders to reduce bias in statistical associations. This adjustment is particularly important in studies using CB, which contains nRBCs 23 . Consequently, conventional estimation models based on adult blood cell profiles are insufficient to accurately correct cell-type heterogeneity in CB-derived methylation data 24 , 25 . In this study, we applied a newly developed cell-type composition estimation model specifically optimized for CB and compatible with sequencing-based methylation data. This model revealed a significantly higher proportion of nRBCs in newborn samples than in adult PB, which is biologically consistent with the well-documented enrichment of nRBCs in the CB 21 . When applied to the neonatal EWAS, the CB-specific model outperformed the conventional adult PB model. Specifically, it yielded a lower inflation factor, indicating improved bias correction, and identified a greater number of CpG sites, surpassing the suggested threshold. Thus, the Hepta-family epigenomic dataset included resources for cell composition-related bias correction models tailored to both peripheral and CB tissues, enabling reliable and biologically appropriate epigenome-wide analyses across different life stages and sample types. As a representative application, we explored the potential effect of maternal and grandmaternal pre-pregnancy smoking on neonatal DNA methylation. Changes in DNA methylation were observed in the CB of neonates born to mothers who had smoked for over a year and quit within 5 years before conception, including at CpG sites near genes previously associated with smoking exposure. Epidemiological studies have reported a higher incidence of congenital anomalies in the offspring of former smokers, even when cessation occurred before pregnancy 26 , and epigenetic changes have also been detected in the granulosa cells of former smokers undergoing assisted reproductive technology 27 , suggesting persistent molecular signatures of smoking in the oocyte microenvironment. Thus, taken together with these previous studies, our dataset provides valuable insights for future studies on the epigenetic inheritance and germline persistence of environmental signals. In contrast, the epigenetic impact of maternal smoking diminished substantially compared with that observed in current smokers, with only a single significant CpG site detected. Moreover, no significant methylation changes were observed in relation to pre-pregnant grandmaternal smoking, suggesting that the intergenerational signal might not be retained in the third generation at the neonatal stage. These findings support the notion that a longer interval between exposure and conception may attenuate epigenetic alterations in offspring. Given that humans require a substantially longer period to reach reproductive maturity than rodent models 28 , this extended developmental window may allow for partial epigenetic recovery from adverse effects induced by early-life or pre-conceptional environmental exposures. As a family-based multigenerational resource, this dataset enables human research on transgenerational epigenetic effects, including the timing of exposure, potential for epigenetic recovery, and implications for health outcomes. In conclusion, we established a multigenerational DNA methylation dataset with detailed cohort information, offering a valuable resource for exploring how environmental exposure and genetic factors shape the human epigenome. By including both adult and neonatal samples, the dataset captures key developmental differences in methylation patterns, highlighting the stage-specific nature of epigenetic regulation. This study provides a foundation for future studies on transgenerational epigenetic inheritance and life-course epigenomics in humans. Methods Ethics This study was approved by two independent research protocols reviewed by the institutional ethics committee. One of them involved the construction of a DNA methylation dataset of the Hepta family and their detailed analysis with family history of smoking as part of the Tohoku Medical Megabank Organization’s birth cohort study, which was approved by the Ethics Committees of Tohoku University (Approval ID: 2020-4-058; September 14, 2020) and Iwate Medical University (IMU) (Approval ID: HG2020-008; June 4, 2020). All adult participants provided written informed consent before participation. For neonatal participants with an insufficient ability to understand the study protocol at any age, informed consent was obtained from their guardians, with the approval of the ethics committee. Another study involving the WGBS analysis of purified nRBCs and the development of a model of cell-type composition estimation for EWAS with umbilical CB was approved by the Ethics Committee of the School of Medicine, IMU (Approval ID: HG2019-025; June 4, 2020). Written informed consent was obtained from pregnant women who delivered at IMU Hospital. Study design and setting The sample for Hepta-family study was drawn from the TMM BirThree Cohort Study in Japan; details of the TMM BirThree Cohort Study have been described elsewhere 13 , 29 , 30 . The study population consisted of 158 Hepta families, comprising 1,107 individuals. DNA methylation data were analyzed for 1,093 individuals who provided informed consent for EWAS analyses incorporating diverse environmental factors and were included in the Hepta-family dataset for this study (Fig. 1 A). The Hepta families comprised a family lineage of seven members, including parents and grandparents, as viewed from the center of the neonate (Fig. 1 B). For the nRBC WGBS analysis, Japanese pregnant women who planned to deliver at the IMU Hospital were recruited, and written informed consent was obtained from 24 participants between March 2020 and November 2020. The inclusion criteria were mothers who had a singleton birth, were scheduled for caesarean section, and had no perinatal complications. The exclusion criteria were as follows: (1) multiple births, (2) mothers with underlying medical conditions, (3) fetuses with congenital anomalies, (4) withdrawal of consent to participate in this study, and (5) CB samples that required extended collection time. All statistical analyses were performed using Tohoku Medical Megabank Organization Supercomputer System. Captured methylation sequencing and DNA methylation profiling of Hepta family Genomic DNA samples derived from the PB of adult participants and umbilical CB of neonates were retrieved from the TMM biobank. Library preparation was performed using 1.1 µg of DNA per sample with a targeted bisulfite sequencing method with Agilent SureSelect Human Methyl-Seq Custom Capture Kits with customized probes (i.e., CDMV 16 version 3 probe set) on an Agilent Bravo system (Agilent Technologies, Santa Clara, CA, USA), following previously described protocols. Sequencing and subsequent data processing, including quality control of raw reads and methylation calling, were conducted, as described previously 20 , 31 . CpG sites with a call rate ≥ 95% were annotated using R package annotatr, with promoter regions (< 1 kb upstream of the TSS) and gene bodies, defined as 5'UTRs, 3'UTRs, exons, and introns. PCA and functional enrichment analysis of Hepta-family methylation data set Principal component analysis was performed on unadjusted beta values from the sequencing data to explore variability and detect batch effects or population structures. CpG sites with missing values (NA) were excluded to ensure the availability of complete data for analysis. The analysis was performed using the prcomp function in R, and the results were visualized using the ggplot2 package to highlight clustering patterns and associations with variables, such as age, sex, and sequencing batch. Eigenvalues derived from the PCA were used to generate a scree plot, which identified the PC1 and PC2 as the most informative contributors to variance. For downstream analysis, CpG sites were selected based on their PC loading scores, with thresholds of ≥ 0.4. Only CpG sites with a CR of 100% were included in the analysis. CpG sites in the associated genes were annotated using the R package annotatr. Gene ontology analysis of the biological process terms was performed using the clusterProfiler package, and tissue enrichment analysis was performed using the TissueEnrich package. Generational differences in DNA methylation levels and variability across genomic regions To investigate DNA methylation levels and interindividual variability across generations and genomic contexts, we calculated the mean β-value at each CpG site for newborns, parents, and grandparents. CpGs with a CR ≥ 95% were retained. Genomic annotations, including CpG island-related categories (CpG islands, shores, shelves, and open sea) and gene-based features (promoters, exons, introns, untranslated regions, and intergenic areas), were assigned using the R packages AnnotationHub and GenomicRanges. Group-wise comparisons across generations within each genomic category were performed using nonparametric methods: the Kruskal–Wallis rank-sum test, followed by Dunn’s post-hoc test with Benjamini–Hochberg correction for multiple testing. Smoking and other variables in Hepta-family methylation dataset Cohort information, including variables for EWAS and baseline characteristics, was obtained from a curated version of the TMM BirThree Cohort Study dataset 13 , 29 , 30 . For this study, data corresponding to 1,095 individuals selected for DNA methylation analysis were extracted, excluding those who withdrew consent. Collected variables included demographic characteristics (age, sex, body mass index), lifestyle factors (e.g., smoking status and alcohol consumption obtained from individual questionnaires), and clinical data. These variables were used to describe the baseline characteristics of the study population and covariates or confounding factors in the EWAS analysis. Cell composition estimation for sequencing-based methylation data of neonatal CB To address cell-type heterogeneity in CB-derived DNA methylation data, we developed a cell-type composition estimation model optimized for sequencing-based data. This model incorporates reference methylation profiles for seven blood cell types, including WGBS data obtained from purified nRBCs from the CB of Japanese newborns. Reference profiles for six other cell types (natural killer cells, B cells, CD4 + T cells, CD8 + T cells, monocytes, and neutrophils) were obtained from previously established datasets. The estimation was implemented by modifying the “estimateCellCounts” function from the R package minfi. The DNA methylation profiles of nRBCs were compared with those of other blood cell types to confirm their distinct epigenetic signatures (see Supplementary Information for full methods and results). A cell composition estimation model for adult PB was developed and utilized in our previous study 31 . The WGBS summary data for nRBCs are publicly available in the iMETHYL database 17 . Epigenome‑wide association study To investigate the potential transmission of smoking-related DNA methylation signatures across generations, we conducted an EWAS focusing on maternal and grandmaternal pre-pregnancy smoking. Specifically, we examined whether smoking exposure in the preceding generations was associated with neonatal DNA methylation patterns and compared CpG sites identified across generations to assess potential transgenerational epigenetic inheritance. CpG sites with a CR of < 95% were excluded. All EWAS were performed using linear regression models, with DNA methylation beta values as the dependent variable and smoking-related traits as the independent variables. The Bonferroni method was used to correct multiple tests. The results were visualized using the qqman and ggplot2 R packages, and CpG sites were annotated based on genomic features using the AnnotationHub, GenomicRanges, rtracklayer, and org.Hs.eg.db R packages. Neonatal EWAS1 – Maternal pre-pregnancy smoking Maternal smoking status before pregnancy was categorized into three groups: never smokers, former smokers who quit within 5 years, and those who quit ≥ 5 years ago. The model was adjusted for the estimated cell-type composition (natural killer cells, B cells, CD4 + T cells, CD8 + T cells, monocytes, neutrophils, and nRBCs), maternal parity (primiparous or multiparous), sex, household income (< 4 or ≥ 4 million yen), and partner smoking status. Additionally, to assess the performance of the cell-type composition model developed for CB, an EWAS for pre-pregnant maternal smoking was conducted using cell-type estimates derived from an adult PB model. Lambda (λ) values from both models were compared to evaluate the effectiveness of cell-type adjustment in controlling statistical inflation. Neonatal EWAS2 – Maternal grandmother’s (MGM’s) pre-pregnancy smoking MGM’s smoking status before pregnancy was categorized as a binary variable (never smokers vs ever smokers), as only one participant had a grandmother who had quit smoking ≥ 5 years before pregnancy. This model was further adjusted for maternal pre-pregnancy smoking, in addition to the covariates described in Neonatal EWAS1. Grandparental EWAS1 – Own smoking status Smoking status was treated as a binary trait, comparing ever smokers (including former and current smokers) with never smokers in an EWAS conducted among grandparents (MGM, MGF, PGM, and PGF). The model was adjusted for the estimated cell-type composition (natural killer cells, B cells, CD4 + T cells, CD8 + T cells, monocytes, and neutrophils), age, sex, and household income. If the household income information for an individual was unavailable, the corresponding partner’s household income data were used as a proxy for adjustment. Grandparental EWAS2 – Own smoking cessation status Smoking cessation was treated as a three-category variable representing smoking cessation status (never smokers, former smokers who quit within 5 years, and those who quit ≥ 5 years ago). These categorizations were based on Fang et al. (2023) 32 . The EWAS was performed using the same covariates as in Grandparental EWAS1. Statistical analyses of smoking-associated CpG sites To evaluate the DNA methylation status of smoking-related CpG sites across generations, we compared the CpG sites identified in each EWAS (smoking, smoking cessation, pre-pregnant maternal smoking, and pre-pregnant MGM’s smoking). We visualized their overlap using Venn diagrams implemented in R (GenomicRanges, rtracklayer, VennDiagram, and UpSetR packages). To compare the methylation levels between neonates and adults (grandparents), we focused on the CpG sites identified in this study as smoking-associated DNA methylation markers. Four groups were included: neonates born to mothers with a history of pre-pregnancy smoking (M_Smk, n = 31), neonates born to nonsmoking mothers (M_Ctr, n = 114), currently smoking grandparents (Smk, n = 86), and never-smoking grandparents (Ctr, n = 299). Cell-type composition bias was corrected using a linear regression model fitted separately for each CpG site, with the DNA methylation β-value as the dependent variable and estimated proportions of six or seven cell types as the independent variables. The adjusted methylation value was calculated as the model residual plus the original mean β-value (residual + mean method). The cell-type composition was estimated using reference models appropriate for CB and adult PB. Two-way analysis of variance was performed on the adjusted β-values for each CpG site, testing for the main effects of smoking status (smoking: yes/no), age group (AgeGroup: adult/neonatal), and their interaction. P -values were corrected for multiple testing using the Bonferroni method (n = 156), with adjusted p < 0.05 considered statistically significant. Declarations Data availability Summary data of DNA methylation profiles in the Hepta-family cohort and WGBS summary data of nRBCs from the CB of Japanese newborns are available in the iMETHYL database (https://imethyl.ihec-epigenomes.org/). Owing to ethical considerations, including the protection of participant privacy and the prevention of unintended identification, individual-level data from the TMM BirThree Cohort Study are not publicly available. Access to these datasets may be granted upon request and is subject to approval from the Ethics Committee of IMU and the Materials and Information Distribution Review Committee of the TMM Project. Researchers interested in accessing these datasets must contact the corresponding author to initiate the request process. A cknowledgments We thank all the participants who provided specimens and data. We are also grateful to Miyuki Horie, Yukino Nakamura, Hiroko Nakamura, and Anna Kudo for their assistance with the experiments. This study was supported by the Tohoku Medical Megabank Project (Special Account for the Reconstruction of the Great East Japan Earthquake) of the Ministry of Education, Culture, Sports, Science and Technology and Japan Agency for Medical Research and Development (AMED) (grant numbers JP19km0105004 and JP20km0105004). Supercomputer resources were provided by an AMED Research Grant (JP20km0405001). This work was also supported by the Japan Endocrine Society Grant for Promising Investigator and JSPS KAKENHI Grant-in-Aid for Young Scientists (grant number 23K14437). Author contributions S. Minabe designed the study, performed analyses, and wrote the manuscript. H.O. managed the Hepta-family experiments and drafted the nRBC methods. S.U. coordinated WGBS experiments. E.K., K. Kikuchi, G. H., M. T., C. I., H. Kawamura, T. S., S. H. obtained informed consent and collected umbilical cord blood. R. O. coordinated scheduling with obstetricians and participants. A.T. processed sequencing data. K.O., T.A., and K.F. conducted laboratory work. S. Komaki managed the iMETHYL database. K. Kumada, S. Mizuno, and H. Kudo handled biospecimen provision. S.T., M.I., and T.O. managed cohort data. F.K., S.O., and K. Kinoshita. supported supercomputer infrastructure. T. B. and A. S. conceived the plan for using nRBC in this study and managed its implementation. M.Y. and S. Kuriyama planned the Hepta-family project and oversaw the cohort. H.O., A.T., K.O., T.A., K.F., S. Komaki, Y.S., Y.O., A.S. reviewed and revised the manuscript. A.S. supervised the project. All authors reviewed and approved the final manuscript. Competing interest The authors declare no competing interests. Disclosure Statement The authors have nothing to disclose. References Barker DJ (2007) The origins of the developmental origins theory. J Intern Med 261:412–417 Gluckman PD, Hanson MA, Cooper C, Thornburg KL (2008) Effect of in utero and early-life conditions on adult health and disease. N Engl J Med 359:61–73 van Dijk SJ et al (2015) Epigenetics and human obesity. Int J Obes (Lond) 39:85–97 Pembrey M, Saffery R, Bygren LO, Network (2014) in Epigenetic, E. & Network in Epigenetic, E. Human transgenerational responses to early-life experience: potential impact on development, health and biomedical research. J Med Genet 51, 563 – 72 Pembrey ME (2010) Male-line transgenerational responses in humans. Hum Fertil (Camb) 13:268–271 Heijmans BT et al (2008) Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A 105:17046–17049 Joubert BR et al (2016) DNA Methylation in Newborns and Maternal Smoking in Pregnancy: Genome-wide Consortium Meta-analysis. Am J Hum Genet 98:680–696 Anway MD, Cupp AS, Uzumcu M, Skinner MK (2005) Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 308:1466–1469 Hoile SP, Lillycrop KA, Thomas NA, Hanson MA, Burdge GC (2011) Dietary protein restriction during F0 pregnancy in rats induces transgenerational changes in the hepatic transcriptome in female offspring. PLoS ONE 6:e21668 Franklin TB et al (2010) Epigenetic transmission of the impact of early stress across generations. Biol Psychiatry 68, 408 – 15 Hebbring S (2019) Genomic and Phenomic Research in the 21st Century. Trends Genet 35:29–41 Pearson H (2015) Massive UK baby study cancelled. Nature 526:620–621 Kuriyama S et al (2020) Cohort Profile: Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study): rationale, progress and perspective. Int J Epidemiol 49:18–19m Miyake K et al (2018) Association between DNA methylation in cord blood and maternal smoking: The Hokkaido Study on Environment and Children's Health. Sci Rep 8:5654 World Health O (2004) United Nations Children's. F. Low birthweight: country, regional and global estimates. World Health Organization, Geneva Hachiya T et al (2017) Genome-wide identification of inter-individually variable DNA methylation sites improves the efficacy of epigenetic association studies. NPJ Genom Med 2:11 Komaki S et al (2018) iMETHYL: an integrative database of human DNA methylation, gene expression, and genomic variation. Hum Genome Var 5:18008 Hoang TT et al (2024) Comprehensive evaluation of smoking exposures and their interactions on DNA methylation. EBioMedicine 100:104956 Ohmomo H et al (2022) DNA Methylation Abnormalities and Altered Whole Transcriptome Profiles after Switching from Combustible Tobacco Smoking to Heated Tobacco Products. Cancer Epidemiol Biomarkers Prev 31:269–279 Ohmomo H et al (2022) Potential DNA methylation biomarkers for the detection of clear cell renal cell carcinoma identified by a whole blood-based epigenome-wide association study. Epigenetics Commun 2:1–11 Jones MJ et al (2025) DNA methylation differences between cord and adult white blood cells reflect postnatal immune cell maturation. Commun Biol 8:237 Zeilinger S et al (2013) Tobacco smoking leads to extensive genome-wide changes in DNA methylation. PLoS ONE 8:e63812 May JE, Marques MB, Reddy VVB, Gangaraju R (2019) Three neglected numbers in the CBC: The RDW, MPV, and NRBC count. Cleve Clin J Med 86:167–172 Gervin K et al (2016) Cell type specific DNA methylation in cord blood: A 450K-reference data set and cell count-based validation of estimated cell type composition. Epigenetics 11:690–698 de Goede OM et al (2015) Nucleated red blood cells impact DNA methylation and expression analyses of cord blood hematopoietic cells. Clin Epigenetics 7:95 Yang L et al (2022) Maternal cigarette smoking before or during pregnancy increases the risk of birth congenital anomalies: a population-based retrospective cohort study of 12 million mother-infant pairs. BMC Med 20:4 Tang Z et al (2024) Former smoking associated with epigenetic modifications in human granulosa cells among women undergoing assisted reproduction. Sci Rep 14:5009 Laffan SB, Posobiec LM, Uhl JE, Vidal JD (2018) Species Comparison of Postnatal Development of the Female Reproductive System. Birth Defects Res 110:163–189 Kuriyama S et al (2016) The Tohoku Medical Megabank Project: Design and Mission. J Epidemiol 26:493–511 Tohoku M, Megabank O (2025) Birth and Three-Generation Cohort Study. Vol. (Tohoku University, 2025) Komaki S et al (2023) Epigenetic profile of Japanese supercentenarians: a cross-sectional study. Lancet Healthy Longev 4:e83–e90 Fang F, Andersen AM, Philibert R, Hancock DB (2023) Epigenetic biomarkers for smoking cessation. Addict Neurosci 6 Tables Tables 1 and 2 are available in the Supplementary Files section. Additional Declarations There is NO Competing Interest. Supplementary Files Table1sm2.xlsx Table 1 Table2sm2.xlsx Table 2 250512SIsm2.docx Supplemental Information SupTable1sm2.xlsx Supplemental Table 1 SupTable2sm1.xlsx Supplemental Table 2 SupTable3sm2.xlsx Supplemental Table 3 SupTable4sm1.xlsx Supplemental Table 4 SupTable5sm1.xlsx Supplemental Table 5 SupTable6sm1.xlsx Supplemental Table 6 SupTable7sm1.xlsx Supplemental Table 7 SupTable8sm1.xlsx Supplemental Table 8 SupFig250513.pdf Supplemental Figures Cite Share Download PDF Status: Posted 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-7314319","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":502659030,"identity":"bb92b212-dcdd-40dc-ac05-8ccdc3321b3c","order_by":0,"name":"Atsushi Shimizu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBklEQVRIiWNgGAWjYFACHoYDDAYMDPzMcJEEEMFGWItkM5B9gFgtYGBwAFULbsDff/bgoRsFdYmbj/MYfv5QwRDNz57A+OEHA18eLi0SN/ISDucYsCVuO8xjLHHgDEPuzJ4HzJI9DGzFOK25wWMA1MID1MK7QeJg2//cDTcSGKSBfklswKFD/vwZkBaJxM3NvJt/HPzHkLv/RgLzb3xaDA7kgLQYJG5g5t0mcbCBIXeDRAIbXlsMb4C1JBjPOMz/zeLMMYbcGWcetln2GOD2i9z5M8afc/7Uyfb3H0u+UVHDkNvfnnz4xo+KYzhDDBtgBDrJ4FgCKVrAoIZ0LaNgFIyCUTBcAQA9mFu+e4LIBgAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-8307-2461","institution":"Iwate Medical University","correspondingAuthor":true,"prefix":"","firstName":"Atsushi","middleName":"","lastName":"Shimizu","suffix":""},{"id":502659031,"identity":"1669b4a8-ba5a-4d9c-8dd7-ee25cb9a4dd4","order_by":1,"name":"Shiori Minabe","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shiori","middleName":"","lastName":"Minabe","suffix":""},{"id":502659032,"identity":"3ef0ceaa-2b39-49a1-96b5-830320b88dc3","order_by":2,"name":"Hideki Ohmomo","email":"","orcid":"https://orcid.org/0000-0002-3524-6524","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hideki","middleName":"","lastName":"Ohmomo","suffix":""},{"id":502659033,"identity":"e2872eaf-dcb0-43bd-b5bb-84b82b59da7e","order_by":3,"name":"Akira Takashima","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Akira","middleName":"","lastName":"Takashima","suffix":""},{"id":502659034,"identity":"eff8796e-bcf7-4a25-9efb-2e2b74fe0111","order_by":4,"name":"Kanako Ono","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kanako","middleName":"","lastName":"Ono","suffix":""},{"id":502659035,"identity":"36e06487-8341-411b-a9ef-a630aa419fea","order_by":5,"name":"Takako Aoyama","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Takako","middleName":"","lastName":"Aoyama","suffix":""},{"id":502659036,"identity":"96f7519c-74ec-435d-bb18-3e7ff9d22250","order_by":6,"name":"Kumi Furusawa","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kumi","middleName":"","lastName":"Furusawa","suffix":""},{"id":502659037,"identity":"96022792-d759-4446-af2b-b45ec96fcf64","order_by":7,"name":"Shohei Komaki","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shohei","middleName":"","lastName":"Komaki","suffix":""},{"id":502659038,"identity":"ea31f67d-2dc5-4aca-9e4e-1931f572ba94","order_by":8,"name":"Yoichi Sutoh","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yoichi","middleName":"","lastName":"Sutoh","suffix":""},{"id":502659039,"identity":"a977856b-f8ce-435a-afb4-5a5b915b7e93","order_by":9,"name":"Yayoi Otsuka-Yamasaki","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yayoi","middleName":"","lastName":"Otsuka-Yamasaki","suffix":""},{"id":502659040,"identity":"f7f04ecf-efcc-4439-b0ec-d06c30ded749","order_by":10,"name":"So Umekage","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"So","middleName":"","lastName":"Umekage","suffix":""},{"id":502659041,"identity":"52950d23-129c-4e66-a58c-e9f4ac160751","order_by":11,"name":"Erika Kurokawa","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Erika","middleName":"","lastName":"Kurokawa","suffix":""},{"id":502659042,"identity":"f8bc4b4a-2b97-4e61-a335-d63e8dfee4a4","order_by":12,"name":"Kotoka Kikuchi","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kotoka","middleName":"","lastName":"Kikuchi","suffix":""},{"id":502659043,"identity":"a5a6f8a5-dbbe-4e8e-a2db-6648b19ba329","order_by":13,"name":"Gen Haba","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Gen","middleName":"","lastName":"Haba","suffix":""},{"id":502659044,"identity":"c30e0375-e9ee-43e4-b84d-f642e38c8181","order_by":14,"name":"Miyuki Terata","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Miyuki","middleName":"","lastName":"Terata","suffix":""},{"id":502659045,"identity":"82c934a8-4521-4007-9125-fd6f66abcad1","order_by":15,"name":"Chizuko Isurugi","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chizuko","middleName":"","lastName":"Isurugi","suffix":""},{"id":502659046,"identity":"02cd47b0-f607-4e8a-9e75-9cd5802b8673","order_by":16,"name":"Hanae Kawamura","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hanae","middleName":"","lastName":"Kawamura","suffix":""},{"id":502659047,"identity":"a8c6f8fb-d2ab-4e75-96e1-e7c7dc0deb56","order_by":17,"name":"Takanori Sato","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Takanori","middleName":"","lastName":"Sato","suffix":""},{"id":502659048,"identity":"549a9595-59a6-4e54-990b-d68282f2fdd8","order_by":18,"name":"Shinya Hatayama","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shinya","middleName":"","lastName":"Hatayama","suffix":""},{"id":502659049,"identity":"291eb559-9b9e-4582-9365-eeefd25fb2c8","order_by":19,"name":"Rie Oyama","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Rie","middleName":"","lastName":"Oyama","suffix":""},{"id":502659050,"identity":"5a0bcd97-8d86-4cec-9287-a3813d67657b","order_by":20,"name":"Tsukasa Baba","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Tsukasa","middleName":"","lastName":"Baba","suffix":""},{"id":502659051,"identity":"143833f6-bd1d-4a17-8e5e-296e2f5cf26a","order_by":21,"name":"Shu Tadaka","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Shu","middleName":"","lastName":"Tadaka","suffix":""},{"id":502659052,"identity":"0b006ed4-347d-4a45-ab9c-36e01bae3320","order_by":22,"name":"Satoshi Mizuno","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Satoshi","middleName":"","lastName":"Mizuno","suffix":""},{"id":502659053,"identity":"51d2dba3-f859-4b81-bae8-b628375a5a2c","order_by":23,"name":"Mami Ishikuro","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Mami","middleName":"","lastName":"Ishikuro","suffix":""},{"id":502659054,"identity":"080574b9-aa89-4c10-8ee8-83a7c9c511ac","order_by":24,"name":"Hisaaki Kudo","email":"","orcid":"https://orcid.org/0000-0002-4477-7095","institution":"Tohoku Medical Megabank Organization, Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Hisaaki","middleName":"","lastName":"Kudo","suffix":""},{"id":502659055,"identity":"4743517b-9512-483f-870a-a47b17e68ed8","order_by":25,"name":"Taku Obara","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Taku","middleName":"","lastName":"Obara","suffix":""},{"id":502659056,"identity":"df7e9901-16aa-47d2-b6cf-90a2f060dd46","order_by":26,"name":"Kazuki Kumada","email":"","orcid":"https://orcid.org/0000-0002-2970-7742","institution":"Department of Biobank Life Science, Tohoku Medical Megabank Organization, Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Kazuki","middleName":"","lastName":"Kumada","suffix":""},{"id":502659057,"identity":"4c7b7d5c-94ef-4a0e-be8b-0da720fa50bd","order_by":27,"name":"Fumiki Katsuoka","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Fumiki","middleName":"","lastName":"Katsuoka","suffix":""},{"id":502659058,"identity":"6a06718c-24a5-422a-abea-99aaf65d5fb5","order_by":28,"name":"Soichi Ogishima","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Soichi","middleName":"","lastName":"Ogishima","suffix":""},{"id":502659059,"identity":"56169b95-2d64-447d-9a35-3d3caeb6c355","order_by":29,"name":"Kengo Kinoshita","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Kengo","middleName":"","lastName":"Kinoshita","suffix":""},{"id":502659060,"identity":"cfa670c2-32c2-4f75-9f8b-4469c5642dd6","order_by":30,"name":"Masayuki Yamamoto","email":"","orcid":"https://orcid.org/0000-0002-9073-9436","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Masayuki","middleName":"","lastName":"Yamamoto","suffix":""},{"id":502659061,"identity":"171faad0-c92e-4de7-a899-6a8e04389334","order_by":31,"name":"Shinichi Kuriyama","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Shinichi","middleName":"","lastName":"Kuriyama","suffix":""}],"badges":[],"createdAt":"2025-08-07 03:40:40","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7314319/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7314319/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89463914,"identity":"0e07ab0a-a353-4dc0-a055-75e70bbc0f9e","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":108265,"visible":true,"origin":"","legend":"\u003cp\u003eStudy workflow. (A) We constructed a comprehensive DNA methylation dataset from 1,093 individuals across 158 Hepta families, comprising pregnant women (mother, M), their newborns (N), the newborns’ fathers (F), and maternal grandmothers (MGM) and maternal grandfathers (MGF) and paternal grandmothers (PGM) and grandfathers (PGF), recruited through the Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study). Genome-wide methylation profiles were obtained using capture-based sequencing from 938 adult peripheral blood (PB) samples and 155 neonatal cord blood (CB) samples. Principal component analysis (PCA) clustering and biological annotation were performed on the Hepta-family dataset. As a representative epigenome-wide association analysis (EWAS), we examined the impact of maternal and maternal grandmother pre-pregnancy smoking on neonatal methylation. During EWAS, cell-type bias was corrected using a conventional estimation model for adult PB and a model developed for CB, ensuring robust adjustment for cell-type composition across sample types. (B) To develop a cell composition estimation model for sequence-based CB DNA methylation data, we also performed whole-genome bisulfite sequencing (WGBS) on purified nucleated red blood cells (nRBCs, n = 15) derived from neonatal CB.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/5b2246233532430542c5a95b.jpg"},{"id":89463917,"identity":"d9e53659-11d7-48bc-b731-d51189034396","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":53323,"visible":true,"origin":"","legend":"\u003cp\u003ePrincipal component analysis (PCA) of DNA methylation profiles and correlation with metadata variables. (A) PCA based on genome-wide DNA methylation profiles of 1,093 participants across three generations. (B) Heatmap of Pearson correlation coefficients between the top eight principal components (PC1–PC8) and variables, including age, sex, DNA concentration, and flow cytometry metrics. Statistically significant correlations are indicated with asterisks.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/fa1569e0931352d34ca01f7a.jpg"},{"id":89465660,"identity":"de421fd5-a2b7-4fbc-b426-93e95f159a5f","added_by":"auto","created_at":"2025-08-20 08:31:23","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":28248,"visible":true,"origin":"","legend":"\u003cp\u003eGlobal and regional DNA methylation levels across three generations. Mean DNA methylation levels were compared across newborns, parents, and grandparents in global and region-specific genomic contexts. Regions were defined based on CpG annotation categories: CpG islands, shores (up to 2 kb from islands), shelves (2–4 kb from islands), and open sea (all other CpGs not located within these regions). Each bar represents the mean methylation percentage at CpG sites with a call rate ≥ 95% within each category. Statistical significance between groups was assessed using nonparametric methods: the Kruskal–Wallis rank-sum test followed by Dunn’s post-hoc test with Benjamini–Hochberg correction for multiple comparisons. Asterisks indicate adjusted p-values: \u003csup\u003e*\u003c/sup\u003e\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05, \u003csup\u003e**\u003c/sup\u003e\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01, \u003csup\u003e***\u003c/sup\u003e\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/a67ce40b1ee2d34828415d48.jpg"},{"id":89464551,"identity":"c8f6c222-082e-45ad-9ef1-d892dd289978","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":96949,"visible":true,"origin":"","legend":"\u003cp\u003eManhattan plots and quantile–quantile plot of epigenome-wide association analysis (EWAS) of maternal or maternal grandmother’s pre-pregnancy smoking using DNA from newborns. (a) EWAS of maternal pre-pregnancy smoking status using a three-category variable: never smokers (n = 114), former smokers who quit within 5 years prior to pregnancy (n = 12), and former smokers who quit ≥ 5 years prior to pregnancy (n = 19). (b) EWAS comparing newborns whose maternal grandmothers had smoked for \u0026gt; 1 year prior to pregnancy (n = 15) and those whose maternal grandmothers had never smoked before pregnancy (n = 122). The red line indicates a Bonferroni significance cutoff of 5.0 × 10\u003csup\u003e–8\u003c/sup\u003e, whereas the gray line indicates a suggestive significance cutoff of 1.0 × 10\u003csup\u003e–5\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/ad2aeaaeb82c614b45a7c391.jpg"},{"id":89465662,"identity":"6ccad421-4787-464c-b686-8d23832f6a6d","added_by":"auto","created_at":"2025-08-20 08:31:23","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":88234,"visible":true,"origin":"","legend":"\u003cp\u003eManhattan plots and quantile–quantile plot of smoking and smoking cessation epigenome-wide association analysis (EWAS) in the grandparent generation. (a) EWAS comparing continuous smokers (Smk, n = 86) and non-smokers (Ctr, n = 299). (b) EWAS of smoking cessation status using a three-category variable: never smokers (n = 302), those who quit within 5 years (n = 31), and those who quit ≥ 5 years ago (n = 167). The red line indicates a Bonferroni significance cutoff of 5.0 × 10\u003csup\u003e–8\u003c/sup\u003e, whereas the gray line indicates a suggestive significance cutoff of 1.0 × 10\u003csup\u003e–5\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/4240eb2b6a284f494a90b0c9.jpg"},{"id":89463929,"identity":"6e6c0034-e918-443b-a493-74f5b584a2f1","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":54370,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of smoking-associated CpG sites in adult peripheral blood (PB) and neonatal cord blood (CB). (a) Overlap of differentially methylated CpG sites associated with current smoking (Smk), smoking cessation (Smk_Cessation), maternal pre-pregnancy smoking (M_Smk), and grandmaternal pre-pregnancy smoking (MGM_Smk). Neonatal CB sites associated with maternal or grandmaternal smoking did not overlap with well-established smoking- or cessation-associated markers identified in adult PB. (b) Cell-type-adjusted DNA methylation levels at four representative CpG sites included in the adult smoking-associated marker set shown in (a) are plotted for four groups: never smoking grandparents (Ctr), current smoking grandparents (Smk), neonates born to nonsmoking mothers (M_Ctr), and neonates born to mothers with a history of pre-pregnancy smoking (M_Smk). Adjustment was performed using CB-specific and PB-specific reference models, respectively. Box plots display the median and interquartile range.\u003c/p\u003e","description":"","filename":"6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/e1ec0c04bcd108a0ea0d3254.jpg"},{"id":93067404,"identity":"8b05cfb9-7851-4cfa-bcd2-29f831783247","added_by":"auto","created_at":"2025-10-08 17:02:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1651476,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/aefe653c-b156-436a-91d1-ec18ae49d660.pdf"},{"id":89463913,"identity":"3a82f0f5-b317-4ab8-938a-c7c78b0631c1","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11452,"visible":true,"origin":"","legend":"Table 1","description":"","filename":"Table1sm2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/090a38c6a46185ef48731d3a.xlsx"},{"id":89464545,"identity":"8246101b-8080-4497-aa62-4880635355e3","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":15603,"visible":true,"origin":"","legend":"Table 2","description":"","filename":"Table2sm2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/28b24ff975f55da9934cddff.xlsx"},{"id":89463919,"identity":"f1b25f66-c2d2-49d9-b3aa-3f9ff64b486c","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":27030,"visible":true,"origin":"","legend":"Supplemental Information","description":"","filename":"250512SIsm2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/c1d5db38bd6971ce419409b1.docx"},{"id":89464548,"identity":"3c8b00e5-f799-447a-83e7-2708bfbef7c0","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":12510,"visible":true,"origin":"","legend":"Supplemental Table 1","description":"","filename":"SupTable1sm2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/8200f75b528721a90b103c3c.xlsx"},{"id":89464550,"identity":"86495b2d-b9a5-494e-8444-20736d73d114","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":10587,"visible":true,"origin":"","legend":"Supplemental Table 2","description":"","filename":"SupTable2sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/7f81fa0d60fcc3eb92cb2113.xlsx"},{"id":89463928,"identity":"24db82e3-4ac4-4632-acde-1e6b49fc71e0","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":12538,"visible":true,"origin":"","legend":"Supplemental Table 3","description":"","filename":"SupTable3sm2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/44eea60493b5b97e10246392.xlsx"},{"id":89464552,"identity":"3678f2a2-f7d5-4210-b923-46d4fd2f0f41","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":11456,"visible":true,"origin":"","legend":"Supplemental Table 4","description":"","filename":"SupTable4sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/fbbdac29858d58db97d5511e.xlsx"},{"id":89463925,"identity":"ff0dc1e6-4abf-4c19-b55f-131420b92350","added_by":"auto","created_at":"2025-08-20 08:15:23","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":11746,"visible":true,"origin":"","legend":"Supplemental Table 5","description":"","filename":"SupTable5sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/5e8fc96b6b070a1869c38dd4.xlsx"},{"id":89464554,"identity":"d7d919c7-f8ec-48ff-997b-813311a5b60a","added_by":"auto","created_at":"2025-08-20 08:23:23","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":27485,"visible":true,"origin":"","legend":"Supplemental Table 6","description":"","filename":"SupTable6sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/9009b08c93e49f588e64e6f2.xlsx"},{"id":89463936,"identity":"d97812bf-3ffd-4588-ac52-ffe68815eac6","added_by":"auto","created_at":"2025-08-20 08:15:24","extension":"xlsx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":12272,"visible":true,"origin":"","legend":"Supplemental Table 7","description":"","filename":"SupTable7sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/c4fa81e3ffcc7ddf1f622b12.xlsx"},{"id":89465663,"identity":"12af6353-ea1b-442c-83a0-b98e805086ec","added_by":"auto","created_at":"2025-08-20 08:31:23","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":20156,"visible":true,"origin":"","legend":"Supplemental Table 8","description":"","filename":"SupTable8sm1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/ae2862783adcbcdda79287c7.xlsx"},{"id":89464568,"identity":"99e6cc3a-1ddb-4860-ae7b-652222b48727","added_by":"auto","created_at":"2025-08-20 08:23:24","extension":"pdf","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":1404880,"visible":true,"origin":"","legend":"Supplemental Figures","description":"","filename":"SupFig250513.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7314319/v1/3600044d1c00f7d127078529.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Mapping the human epigenetic landscape across three generations: A DNA methylation resource from TMM BirThree","fulltext":[{"header":"Introduction","content":"\u003cp\u003eUnderstanding the origins of complex human diseases requires an integrated view of genetic inheritance and environmental exposure across the life course. The Developmental Origins of Health and Disease (DOHaD) concept has been well-established through numerous epidemiological studies linking early-life environments to adult disease risk \u003csup\u003e\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Seminal human studies \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e have suggested that prenatal and even ancestral exposures can influence offspring and grand-offspring health outcomes.\u003c/p\u003e\u003cp\u003eDNA methylation is thought to mediate these associations, at least in part \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Epigenetic mechanisms offer a plausible molecular basis for how early-life environmental cues become embedded and potentially transmitted across generations. Animal studies have demonstrated that prenatal exposure to endocrine disruptors \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, nutritional restriction \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, or maternal stress \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e can induce heritable methylation changes, providing evidence for transgenerational epigenetic inheritance. However, in humans, molecular evidence for such inheritance remains scarce, largely because of the difficulty of conducting large-scale and long-term cohort studies that can integrate multigenerational exposure and epigenomic data by collecting biological and socioeconomic information across generations. However, these studies face substantial logistical, financial, and ethical barriers \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Large national studies were eventually terminated owing to challenges in participant recruitment and the complexity of the study design \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThe Tohoku Medical Megabank Project\u0026rsquo;s Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study) has been established as a prospective genomic cohort to support precision medicine and long-term health monitoring in post-disaster Japan \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Through longitudinal collection of biospecimens and detailed metadata on medical history, lifestyle, diet, and psychosocial stress, the TMM BirThree cohort provides a framework for investigating how early-life and ancestral exposure may influence molecular phenotypes, including DNA methylation. Its multigenerational structure and deep phenotyping make it particularly well-suited for examining research questions aligned with the DOHaD concept.\u003c/p\u003e\u003cp\u003eTo investigate the molecular basis of epigenetic inheritance and the DOHaD concept in humans, we established two types of DNA methylation resources: a three-generation dataset from 158 Japanese families (the Hepta-family dataset) \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e and a population-specific reference panel based on whole-genome bisulfite sequencing (WGBS) of nucleated red blood cells (nRBCs) from Japanese newborns used to develop a cell-type composition estimation model. The Hepta-family dataset includes methylation profiles obtained by targeted bisulfite sequencing from newborns (umbilical cord blood [CB]), their parents, and all four grandparents: maternal grandmother (MGM), maternal grandfather (MGF), paternal grandmother (PGM), and paternal grandfather (PGF), with peripheral blood (PB) used for all samples except newborns. To demonstrate the utility of this multigenerational resource, we performed a representative analysis focusing on the epigenetic impact of maternal and grandmaternal pre-pregnancy smoking, an environmental exposure with well-established and reproducible methylation signatures \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. This analysis suggests that the epigenetic response to smoking may vary according to developmental stage and tissue context, highlighting the importance of considering stage-specific effects when investigating transgenerational epigenetic signals.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStructure and characteristics of the three generation Hepta family\u003c/h2\u003e\u003cp\u003eThe dataset comprised 158 mothers, 158 fathers, 157 MGMs, 156 MGFs, 155 PGMs, 154 PGFs, and 155 neonates (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Among them, 148 families were complete with all 7 members available for analysis. One family included twins, resulting in an eight-member structure (Supplementary Fig.\u0026nbsp;1). In addition, the dataset included 154 maternal-neonate pairs, 468 spousal pairs across three generations, and 464 parent\u0026ndash;child trios. These trios comprised paternal (father, paternal grandparent, and neonate) and maternal (mother, maternal grandparent, and neonate) configurations (Supplementary Fig.\u0026nbsp;1).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;1 summarizes the characteristics of the newborns based on the maternal and neonatal medical records at admission and parent-administered questionnaires completed by the mothers. Among the newborns, 54.2% were male. The mean gestational age at birth was 39.2\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4 weeks, and the mean birth weight was 3,078.0\u0026thinsp;\u0026plusmn;\u0026thinsp;397.6 g. The proportion of low birthweight, defined by the World Health Organization as a birthweight of an infant of \u0026le;\u0026thinsp;2,499 g regardless of gestational age \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e, was 8.4% (13 of 155 newborns).\u003c/p\u003e\u003cp\u003eThe characteristics of the pregnant women (neonate mothers), fathers, and grandparents are shown in Table\u0026nbsp;2. Regarding household income, 23.4% of mothers reported an annual household income of \u0026lt;\u0026thinsp;4\u0026nbsp;million yen, and 35\u0026ndash;39.0% of grandparents fell within this category. Household income data were partially missing, particularly for 6.4\u0026ndash;7.1% of grandfathers and 15.5\u0026ndash;16.6% of grandmothers, and were unavailable for fathers.\u003c/p\u003e\u003cp\u003eData on smoking and alcohol consumption were available for all pregnant mothers (Table\u0026nbsp;2). Among them, 35 were former smokers and none were current smokers. In contrast, 24.1% of mothers reported current alcohol consumption. Among MGMs, 4.5% were current smokers and 10.2% were former smokers, a trend that was similar among PGMs. Regarding grandfathers, 17.3% of MGFs and 29.9% of PGFs were current smokers, whereas 59.0% and 51.3%, respectively, were former smokers.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eSequencing and mapping statistics of the Hepta-family DNA methylation dataset\u003c/h3\u003e\n\u003cp\u003eWe analyzed\u0026thinsp;\u0026gt;\u0026thinsp;1\u0026nbsp;million CpG sites spanning the promoter and gene body regions across \u0026gt;\u0026thinsp;16,500 annotated genes using the common DNA methylation variations (CDMV) \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e version 3 (CDMVv3) probe set probe set for captured methylation sequencing. The mean sequencing depth across target regions ranged from 26.2 to 26.9\u0026times;, and ~\u0026thinsp;98.0% of the regions were covered at least once, ensuring reliable methylation quantification and demonstrating efficient hybridization and sequencing performance (Supplementary Table\u0026nbsp;1). The consistency in sequencing depth across generations and between sexes further supports the reproducibility of our method. The detailed statistics are shown in Supplementary Table\u0026nbsp;1.\u003c/p\u003e\n\u003ch3\u003ePrincipal component analysis (PCA) clustering and biological annotation in the Hepta-family DNA methylation dataset\u003c/h3\u003e\n\u003cp\u003ePCA was performed using DNA methylation data from all 1,093 individuals included in the Hepta-family dataset. The first and second principal components (PCs) (PC1 and PC2, respectively) explained 10.9% and 5.9% of the variance, respectively, with neonates forming a distinct cluster clearly separated from adult family members (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The correlation heatmap in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB was used to assess the potential batch effects by evaluating the associations between PCs and various factors, including age, sample type (CB vs. PB), flow cell ID, sequencing batch, and sex. The results indicated no significant batch effects, confirming the robustness of the dataset for subsequent analyses.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo further explore the biological relevance of the major axes of variation, PC1 and PC2, (Supplementary Fig.\u0026nbsp;2A), we performed gene enrichment analysis on CpG sites with high loading scores (\u0026ge;\u0026thinsp;0.4). Among the 823,553 CpG sites analyzed (call rate [CR]\u0026thinsp;=\u0026thinsp;100%), 162,802 CpG sites for PC1 and 81,574 for PC2 met the cutoff (Supplementary Fig.\u0026nbsp;2B). The CpG sites contributing to PC1 were enriched in genes associated with neurodevelopmental processes, including axonogenesis, regulation of nervous system development, and sensory system formation (Supplementary Fig.\u0026nbsp;2C, \u003cem\u003eupper\u003c/em\u003e). Tissue enrichment analysis further supported these findings, revealing a strong overrepresentation of genes specifically expressed in the brain (Supplementary Fig.\u0026nbsp;2D, \u003cem\u003eupper\u003c/em\u003e). In contrast, CpG sites contributing to PC2 were enriched in genes involved in immune-related processes, such as lymphocyte differentiation, leukocyte activation, and cell\u0026ndash;cell adhesion (Supplementary Fig.\u0026nbsp;2C, \u003cem\u003elower\u003c/em\u003e). These findings were further supported by tissue enrichment analysis, which revealed a significant overlap with genes specifically expressed in the lymph nodes, spleen, and bone marrow (Supplementary Fig.\u0026nbsp;2D, \u003cem\u003elower\u003c/em\u003e), indicating that PC2 captures immunologically driven interindividual epigenetic variability.\u003c/p\u003e\n\u003ch3\u003eGlobal and regional methylation levels across generations\u003c/h3\u003e\n\u003cp\u003eTo investigate generational differences in DNA methylation levels, we compared mean methylation rates across newborns, parents, and grandparents using CpG sites with a CR\u0026thinsp;\u0026ge;\u0026thinsp;95% (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Significant generational differences in global methylation levels were observed (adjusted \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Newborns had significantly higher global methylation levels than both parents and grandparents (adjusted \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and a smaller yet significant difference was observed between parents and grandparents (adjusted \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Significant effects of generation were also observed in the CpG islands, shores, shelves, and open-sea regions (all adjusted \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eDevelopment and validation of a cell-type composition estimation method for CB in sequencing-based DNA methylation data\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo develop a reference panel for cell-type composition estimation in sequencing-based DNA methylation data from CB, nRBCs were isolated from CB samples collected from 24 newborns. All samples underwent cell sorting to isolate CD3⁻CD14⁻CD19⁻CD71⁺CD235⁺ cell populations. Of them, 15 samples passed quality control based on cell purity (\u0026ge;\u0026thinsp;94%) and DNA integrity and were included in downstream analyses. The average maternal age for the 15 newborns (10 males and 5 females) was 35.3\u0026thinsp;\u0026plusmn;\u0026thinsp;4.8 years, and the mean gestational age at delivery was 37.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.0 weeks. Extracted genomic DNA showed high integrity with a mean DNA Integrity Number of 8.2\u0026thinsp;\u0026plusmn;\u0026thinsp;0.6, and nuclear staining confirmed the identity of the isolated nRBC.\u003c/p\u003e\u003cp\u003eWGBS data from 15 newborn-derived nRBC samples revealed approximately 27\u0026nbsp;million CpG sites (\u0026ge;\u0026thinsp;1\u0026times; depth) detected on both DNA strands. The raw sequencing depth exceeded 30\u0026times; coverage, ensuring a sufficient read depth for downstream analysis (Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). The DNA methylation profiles of nRBCs were compared with the WGBS data from eight other blood cell types (monocytes, neutrophils, CD4\u003csup\u003e+\u003c/sup\u003e T cells, CD8\u003csup\u003e+\u003c/sup\u003e T cells, NK cells, B cells, PB mononuclear cells, and leukocytes, n\u0026thinsp;=\u0026thinsp;20) previously published in the iMETHYL database \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Moreover, WGBS data from erythroblasts (n\u0026thinsp;=\u0026thinsp;2) obtained from the IHEC Data Portal and DNA methylation profiles of CBs obtained from 155 neonates were generated using targeted bisulfite sequencing in this study. PCA using 9,880 CpG sites detected across all datasets demonstrated that nRBCs formed a distinct cluster, indicating a unique DNA methylation profile (Supplementary Fig.\u0026nbsp;3).\u003c/p\u003e\u003cp\u003eNext, we incorporated the nRBC WGBS data into a reference panel to estimate the cell-type composition in the CB samples. This model was applied to 938 adult PB and 155 CB samples from the Hepta-family cohort. Neutrophils were the predominant cell type in both sample types (69.7\u0026thinsp;\u0026plusmn;\u0026thinsp;10.3% in PB, 54.9\u0026thinsp;\u0026plusmn;\u0026thinsp;9.8% in CB). As expected, nRBCs were nearly absent in PB (0.6\u0026thinsp;\u0026plusmn;\u0026thinsp;0.8%) but present at 12.8\u0026thinsp;\u0026plusmn;\u0026thinsp;3.6% in CB, and this difference was statistically significant (t-test, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Supplementary Fig.\u0026nbsp;4A and 4B). The coefficient sets derived from both CB- and adult-specific models are available in the iMETHYL database.\u003c/p\u003e\u003cp\u003eTo evaluate the performance of the CB-specific model, we conducted an epigenome-wide association studies (EWAS) using neonatal CB methylation data from the Hepta-family dataset with maternal smoking history as the exposure variable. Compared with the conventional adult-based estimation model (Model 1, Supplementary Fig.\u0026nbsp;5, \u003cem\u003eright\u003c/em\u003e), the CB-specific model (Model 2, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, \u003cem\u003eright\u003c/em\u003e) demonstrated improved bias correction, as reflected by reduced genomic inflation factor (λ), which decreased from 1.027 (95% confidence interval [CI], 1.023\u0026ndash;1.032) to 1.010 (95% CI, 1.006\u0026ndash;1.015). Furthermore, Model 2 identified a greater number of CpG sites surpassing the genome-wide suggestive threshold (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;1.0 \u0026times; 10⁻\u003csup\u003e5\u003c/sup\u003e), with 14 CpG sites detected (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, \u003cem\u003eleft\u003c/em\u003e) compared with 9 CpG sites by Model 1 (Supplementary Fig.\u0026nbsp;5, \u003cem\u003eleft\u003c/em\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eDemonstration 1: Impact of maternal ancestral smoking history on DNA methylation in neonatal CB\u003c/h3\u003e\n\u003cp\u003eCharacteristic features related to maternal smoking among the 153 MGM-mother\u0026ndash;newborn families with available epigenomic data are summarized in Supplementary Table\u0026nbsp;3. Among the 153 mothers with available smoking history, 29 quit smoking before pregnancy. Of them, 27 had smoked for \u0026gt;\u0026thinsp;1 year prior to conception, and two had smoked for \u0026lt;\u0026thinsp;1 year. The average duration of smoking cessation before pregnancy was 7.1\u0026thinsp;\u0026plusmn;\u0026thinsp;3.9 years. Information on smoking was available for 146 MGMs, among whom, 23 were identified as ever-smokers (either current or former smokers). The pre-pregnancy smoking status of each MGM was estimated by comparing the reported age at smoking initiation with the calculated age at delivery based on the birth dates of the grandmother and mother. Among the 23 ever-smokers, 16 had initiated smoking\u0026thinsp;\u0026gt;\u0026thinsp;1 year before pregnancy. Of these, 12 were current smokers, and 4 were former smokers who had quit either before or during pregnancy. The average duration of smoking prior to pregnancy was 5.4\u0026thinsp;\u0026plusmn;\u0026thinsp;3.3 years, and the mean duration since smoking cessation (for former smokers) was 2.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.9 years. Information regarding whether smoking continued throughout pregnancy was unavailable.\u003c/p\u003e\u003cp\u003eTo investigate the effect of maternal pre-pregnancy smoking on DNA methylation in the offspring, we compared CB DNA methylation profiles (998,877 CpG sites) across three groups: never smokers (n\u0026thinsp;=\u0026thinsp;114), former smokers who quit within 5 years prior to pregnancy (n\u0026thinsp;=\u0026thinsp;12), and former smokers who quit\u0026thinsp;\u0026ge;\u0026thinsp;5 years prior to pregnancy (n\u0026thinsp;=\u0026thinsp;19). The EWAS identified one CpG site that surpassed the Bonferroni-corrected significance threshold (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.1 \u0026times; 10⁻⁸) and 13 CpG sites that reached the suggestive threshold (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;1.0 \u0026times; 10⁻\u003csup\u003e5\u003c/sup\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Among them, six CpG sites mapped to six distinct genes (Supplementary Table\u0026nbsp;4). Notably, several of these genes were previously implicated in EWAS associated with tobacco exposure. These genes included \u003cem\u003eUBR4\u003c/em\u003e, \u003cem\u003eKIFC3\u003c/em\u003e, \u003cem\u003eMYOD1\u003c/em\u003e, \u003cem\u003eADORA2A-AS1\u003c/em\u003e, and \u003cem\u003eCACNA1C\u003c/em\u003e, indicating their biological relevance \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eNext, to examine the effect of MGMs\u0026rsquo; pre-pregnancy smoking on DNA methylation in grandchildren, we compared the CB DNA methylation profiles (991,676 CpG sites) between newborns whose MGMs had smoked for \u0026gt;\u0026thinsp;1 year prior to pregnancy (n\u0026thinsp;=\u0026thinsp;15) and those whose MGMs had never smoked before pregnancy (n\u0026thinsp;=\u0026thinsp;122) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). EWAS did not identify any CpG sites that reached genome-wide significance; however, 19 CpG sites showed suggestive associations (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;1.0 \u0026times; 10⁻\u003csup\u003e5\u003c/sup\u003e). Among them, 10 CpG sites mapped to 10 distinct genes (Supplementary Table\u0026nbsp;5). Several of these genes were previously implicated in EWAS related to tobacco exposure. These included \u003cem\u003eANXA6\u003c/em\u003e, \u003cem\u003eTERT\u003c/em\u003e, \u003cem\u003eHSPG2\u003c/em\u003e, \u003cem\u003eOGDHL\u003c/em\u003e, and \u003cem\u003eEHMT1\u003c/em\u003e, indicating potential biological relevance \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eDemonstration 2: EWAS of smoking and smoking cessation in the grandparent generation\u003c/h2\u003e\u003cp\u003eTo identify smoking-related DNA methylation markers detectable in the Hepta-family dataset, we conducted EWAS using epigenomic data from the grandparent generation. This cohort was selected because it included a substantial number of unrelated individuals and sufficient number of continuous smokers. Of the 622 grandparents, 39 with missing information on smoking status or household income were excluded.\u003c/p\u003e\u003cp\u003eThe analysis included 86 continuous smokers (Smk group) and 299 nonsmokers (Ctr group), with DNA methylation profiles assessed at 988,567 CpG sites using smoking status as the trait of interest. The EWAS identified 59 significant CpG sites surpassing the Bonferroni-corrected threshold (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.1 \u0026times; 10⁻⁸) and 206 CpG sites with suggestive associations (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;1.0 \u0026times; 10⁻\u003csup\u003e5\u003c/sup\u003e) (Supplementary Table\u0026nbsp;6). The Manhattan plot of smoking EWAS (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA) revealed peaks in genomic regions previously associated with smoking, including 2q37.1, 11q13.4, and genes such as \u003cem\u003eGFI1\u003c/em\u003e, \u003cem\u003eKIF5C\u003c/em\u003e, \u003cem\u003eEXOC2\u003c/em\u003e, \u003cem\u003eAHRR\u003c/em\u003e, \u003cem\u003eALPP\u003c/em\u003e, \u003cem\u003eCNTNAP2\u003c/em\u003e, \u003cem\u003eMGAT3\u003c/em\u003e, \u003cem\u003ePLAT\u003c/em\u003e, \u003cem\u003eMYOM1\u003c/em\u003e, and \u003cem\u003eF2RL3\u003c/em\u003e \u003csup\u003e18,19\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo explore the long-term epigenetic legacy of smoking after cessation, we conducted EWAS using a three-category variable for smoking status. The categories were never smokers (n\u0026thinsp;=\u0026thinsp;302), former smokers who quit within 5 years (n\u0026thinsp;=\u0026thinsp;31), and those who quit\u0026thinsp;\u0026ge;\u0026thinsp;5 years ago (n\u0026thinsp;=\u0026thinsp;167). DNA methylation was assessed at 989,219 CpG sites. The EWAS identified three significant CpG sites surpassing the Bonferroni-corrected threshold (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;5.1 \u0026times; 10⁻⁸) and 64 suggestive CpG sites (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;1.0 \u0026times; 10⁻\u003csup\u003e5\u003c/sup\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB; Supplementary Table\u0026nbsp;7). Manhattan plots revealed signal CpG sites at 2q37.1 and 11q13.4 loci and within \u003cem\u003eKIF5C\u003c/em\u003e genes, consistent with those observed in smoking EWAS (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Supplementary Table\u0026nbsp;6).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eDemonstration 3: Comparison of smoking-associated CpG sites in PB and CB\u003c/h3\u003e\n\u003cp\u003eCpG sites showing differential DNA methylation in neonatal CB from mothers with a history of smoking did not overlap with known smoking- associated or smoking-cessation-associated CpG markers previously identified in adult PB (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Similarly, altered CpG sites in neonates whose MGMs had a history of smoking showed no overlap with adult-associated markers.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eWhen comparing cell-type-adjusted methylation levels at known smoking-associated CpG sites, we found notable differences between adult PB and neonatal CB. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB and Supplementary Table\u0026nbsp;8, \u003cem\u003eF2RL3\u003c/em\u003e, \u003cem\u003eNFE2L2\u003c/em\u003e, \u003cem\u003eRPTOR\u003c/em\u003e, and \u003cem\u003ePPP1R15A\u003c/em\u003e were hypermethylated in neonatal CB compared with adult PB, whereas 2q37.1, \u003cem\u003eMGAT3\u003c/em\u003e, and \u003cem\u003eEXOC2\u003c/em\u003e were hypomethylated. Similar methylation patterns were observed in publicly available datasets. Using the iMETHYL database, we compared cell-type-adjusted methylation levels between adult PB (control group of the Kidney Cancer study) and gestational age-specific CB samples \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eAlthough most smoking-associated CpG sites exhibited age-dependent differences in DNA methylation between newborns and adults, a small subset showed stable methylation patterns regardless of developmental stage. These included the CpG sites located at chr1:92947586 and chr1:92947588 in \u003cem\u003eGFI1\u003c/em\u003e, chr2:149824190 in \u003cem\u003eKIF5C\u003c/em\u003e, chr5:373651 in \u003cem\u003eAHRR\u003c/em\u003e, chr7:145814670 in \u003cem\u003eCNTNAP2\u003c/em\u003e, chr12:53613078 in \u003cem\u003eRARG\u003c/em\u003e, and chr5:17000346 in \u003cem\u003eF2RL3\u003c/em\u003e (Supplementary Table\u0026nbsp;8). These consistently methylated sites may serve as robust biomarkers of tobacco exposure and may be less influenced by age-related epigenetic variations.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we constructed a three-generation epigenomic dataset comprising 1,093 samples with detailed family structures and cohort metadata. All experimental steps, including DNA library preparation, targeted bisulfite sequencing, and bioinformatics analysis, were performed at a single facility to minimize technical variability and ensure data consistency across samples. To further enhance the accuracy of EWAS, we generated WGBS data for nRBCs and developed a cell-type composition estimation model specifically for neonatal CB. We also applied an existing cell-type estimation model for adult PB, enabling cell-type adjustment in both neonatal and adult samples derived from the same cohort. Summary data from the Hepta-family dataset have been made publicly available via the iMETHYL database in the form of summary-level epigenomic data, serving as a reference panel for epigenetic studies in Japanese populations. Notably, the demographic and phenotypic characteristics of our participants closely mirror those of the original three-generation cohort \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e, indicating that the dataset is representative of the general Japanese population and suitable for use as a reference in future epigenetic studies. The family-based structure provides a framework for investigating intergenerational effects and environmental exposure within a controlled genetic framework.\u003c/p\u003e\u003cp\u003eSubstantial differences in DNA methylation profiles were observed between neonatal CB and adult PB. PCA revealed a clear separation between the two. PC1 reflected variations associated with neurodevelopment and organogenesis and was enriched in brain-specific gene sets, whereas PC2 captured immunological diversity associated with lymphoid tissue-expressed genes. These differences were not fully explained by cell-type composition alone, suggesting coordinated, developmentally regulated epigenetic programs. Moreover, neonates showed significantly lower global and region-specific methylation levels, particularly in CpG islands, than parents and grandparents. This widespread hypomethylation, together with functional enrichment patterns, supports the epigenomic plasticity of early development \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e and emphasizes the need to consider the developmental stage and cellular context when interpreting methylation data across generations.\u003c/p\u003e\u003cp\u003eRecent studies have further demonstrated that DNA methylation profiles differ markedly between neonatal and adult immune cells, even within the same cell type, reflecting lineage-specific developmental maturation \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. These findings underscore that even with appropriate cell-type adjustment, comparisons between neonatal and adult samples must be interpreted with caution. Even after adjusting for cell-type composition, methylation levels at almost all smoking marker CpGs, such as those in \u003cem\u003eF2RL3\u003c/em\u003e and 2q37.1 loci \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, differed markedly between newborns and adults. Moreover, although our dataset replicated several well-established smoking-associated CpG sites \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e found in the PB, the sites associated with maternal or grandmaternal smoking history in the descendants did not overlap with these known markers. These observations suggest that the epigenetic response to smoking exposure differs depending on the developmental stage and tissue type. CpG sites that respond to environmental exposure may vary depending on the tissue and developmental context, and extrapolating adult-derived findings to neonatal samples may lead to misinterpretations. This is particularly important for studies investigating transgenerational epigenetic inheritance.\u003c/p\u003e\u003cp\u003eIn EWAS using blood-derived DNA methylation data, it is standard practice to adjust for potential confounders to reduce bias in statistical associations. This adjustment is particularly important in studies using CB, which contains nRBCs \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Consequently, conventional estimation models based on adult blood cell profiles are insufficient to accurately correct cell-type heterogeneity in CB-derived methylation data \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. In this study, we applied a newly developed cell-type composition estimation model specifically optimized for CB and compatible with sequencing-based methylation data. This model revealed a significantly higher proportion of nRBCs in newborn samples than in adult PB, which is biologically consistent with the well-documented enrichment of nRBCs in the CB \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. When applied to the neonatal EWAS, the CB-specific model outperformed the conventional adult PB model. Specifically, it yielded a lower inflation factor, indicating improved bias correction, and identified a greater number of CpG sites, surpassing the suggested threshold. Thus, the Hepta-family epigenomic dataset included resources for cell composition-related bias correction models tailored to both peripheral and CB tissues, enabling reliable and biologically appropriate epigenome-wide analyses across different life stages and sample types.\u003c/p\u003e\u003cp\u003eAs a representative application, we explored the potential effect of maternal and grandmaternal pre-pregnancy smoking on neonatal DNA methylation. Changes in DNA methylation were observed in the CB of neonates born to mothers who had smoked for over a year and quit within 5 years before conception, including at CpG sites near genes previously associated with smoking exposure. Epidemiological studies have reported a higher incidence of congenital anomalies in the offspring of former smokers, even when cessation occurred before pregnancy \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, and epigenetic changes have also been detected in the granulosa cells of former smokers undergoing assisted reproductive technology \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, suggesting persistent molecular signatures of smoking in the oocyte microenvironment. Thus, taken together with these previous studies, our dataset provides valuable insights for future studies on the epigenetic inheritance and germline persistence of environmental signals.\u003c/p\u003e\u003cp\u003eIn contrast, the epigenetic impact of maternal smoking diminished substantially compared with that observed in current smokers, with only a single significant CpG site detected. Moreover, no significant methylation changes were observed in relation to pre-pregnant grandmaternal smoking, suggesting that the intergenerational signal might not be retained in the third generation at the neonatal stage. These findings support the notion that a longer interval between exposure and conception may attenuate epigenetic alterations in offspring. Given that humans require a substantially longer period to reach reproductive maturity than rodent models \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, this extended developmental window may allow for partial epigenetic recovery from adverse effects induced by early-life or pre-conceptional environmental exposures. As a family-based multigenerational resource, this dataset enables human research on transgenerational epigenetic effects, including the timing of exposure, potential for epigenetic recovery, and implications for health outcomes.\u003c/p\u003e\u003cp\u003eIn conclusion, we established a multigenerational DNA methylation dataset with detailed cohort information, offering a valuable resource for exploring how environmental exposure and genetic factors shape the human epigenome. By including both adult and neonatal samples, the dataset captures key developmental differences in methylation patterns, highlighting the stage-specific nature of epigenetic regulation. This study provides a foundation for future studies on transgenerational epigenetic inheritance and life-course epigenomics in humans.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch2\u003eEthics\u003c/h2\u003e\u003cp\u003e This study was approved by two independent research protocols reviewed by the institutional ethics committee. One of them involved the construction of a DNA methylation dataset of the Hepta family and their detailed analysis with family history of smoking as part of the Tohoku Medical Megabank Organization’s birth cohort study, which was approved by the Ethics Committees of Tohoku University (Approval ID: 2020-4-058; September 14, 2020) and Iwate Medical University (IMU) (Approval ID: HG2020-008; June 4, 2020). All adult participants provided written informed consent before participation. For neonatal participants with an insufficient ability to understand the study protocol at any age, informed consent was obtained from their guardians, with the approval of the ethics committee. Another study involving the WGBS analysis of purified nRBCs and the development of a model of cell-type composition estimation for EWAS with umbilical CB was approved by the Ethics Committee of the School of Medicine, IMU (Approval ID: HG2019-025; June 4, 2020). Written informed consent was obtained from pregnant women who delivered at IMU Hospital.\u003c/p\u003e\u003ch2\u003eStudy design and setting\u003c/h2\u003e\u003cp\u003eThe sample for Hepta-family study was drawn from the TMM BirThree Cohort Study in Japan; details of the TMM BirThree Cohort Study have been described elsewhere \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. The study population consisted of 158 Hepta families, comprising 1,107 individuals. DNA methylation data were analyzed for 1,093 individuals who provided informed consent for EWAS analyses incorporating diverse environmental factors and were included in the Hepta-family dataset for this study (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). The Hepta families comprised a family lineage of seven members, including parents and grandparents, as viewed from the center of the neonate (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e\u003cp\u003e For the nRBC WGBS analysis, Japanese pregnant women who planned to deliver at the IMU Hospital were recruited, and written informed consent was obtained from 24 participants between March 2020 and November 2020. The inclusion criteria were mothers who had a singleton birth, were scheduled for caesarean section, and had no perinatal complications. The exclusion criteria were as follows: (1) multiple births, (2) mothers with underlying medical conditions, (3) fetuses with congenital anomalies, (4) withdrawal of consent to participate in this study, and (5) CB samples that required extended collection time.\u003c/p\u003e\u003cp\u003eAll statistical analyses were performed using Tohoku Medical Megabank Organization Supercomputer System.\u003c/p\u003e\u003ch2\u003eCaptured methylation sequencing and DNA methylation profiling of Hepta family\u003c/h2\u003e\u003cp\u003eGenomic DNA samples derived from the PB of adult participants and umbilical CB of neonates were retrieved from the TMM biobank. Library preparation was performed using 1.1 µg of DNA per sample with a targeted bisulfite sequencing method with Agilent SureSelect Human Methyl-Seq Custom Capture Kits with customized probes (i.e., CDMV\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e version 3 probe set) on an Agilent Bravo system (Agilent Technologies, Santa Clara, CA, USA), following previously described protocols. Sequencing and subsequent data processing, including quality control of raw reads and methylation calling, were conducted, as described previously \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. CpG sites with a call rate ≥ 95% were annotated using R package annotatr, with promoter regions (\u0026lt; 1 kb upstream of the TSS) and gene bodies, defined as 5'UTRs, 3'UTRs, exons, and introns.\u003c/p\u003e\u003ch2\u003ePCA and functional enrichment analysis of Hepta-family methylation data set\u003c/h2\u003e\u003cp\u003ePrincipal component analysis was performed on unadjusted beta values from the sequencing data to explore variability and detect batch effects or population structures. CpG sites with missing values (NA) were excluded to ensure the availability of complete data for analysis. The analysis was performed using the prcomp function in R, and the results were visualized using the ggplot2 package to highlight clustering patterns and associations with variables, such as age, sex, and sequencing batch.\u003c/p\u003e\u003cp\u003eEigenvalues derived from the PCA were used to generate a scree plot, which identified the PC1 and PC2 as the most informative contributors to variance. For downstream analysis, CpG sites were selected based on their PC loading scores, with thresholds of ≥ 0.4. Only CpG sites with a CR of 100% were included in the analysis. CpG sites in the associated genes were annotated using the R package annotatr. Gene ontology analysis of the biological process terms was performed using the clusterProfiler package, and tissue enrichment analysis was performed using the TissueEnrich package.\u003c/p\u003e\u003ch2\u003eGenerational differences in DNA methylation levels and variability across genomic regions\u003c/h2\u003e\u003cp\u003eTo investigate DNA methylation levels and interindividual variability across generations and genomic contexts, we calculated the mean β-value at each CpG site for newborns, parents, and grandparents. CpGs with a CR ≥ 95% were retained. Genomic annotations, including CpG island-related categories (CpG islands, shores, shelves, and open sea) and gene-based features (promoters, exons, introns, untranslated regions, and intergenic areas), were assigned using the R packages AnnotationHub and GenomicRanges. Group-wise comparisons across generations within each genomic category were performed using nonparametric methods: the Kruskal–Wallis rank-sum test, followed by Dunn’s post-hoc test with Benjamini–Hochberg correction for multiple testing.\u003c/p\u003e\u003ch2\u003eSmoking and other variables in Hepta-family methylation dataset\u003c/h2\u003e\u003cp\u003eCohort information, including variables for EWAS and baseline characteristics, was obtained from a curated version of the TMM BirThree Cohort Study dataset \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. For this study, data corresponding to 1,095 individuals selected for DNA methylation analysis were extracted, excluding those who withdrew consent. Collected variables included demographic characteristics (age, sex, body mass index), lifestyle factors (e.g., smoking status and alcohol consumption obtained from individual questionnaires), and clinical data. These variables were used to describe the baseline characteristics of the study population and covariates or confounding factors in the EWAS analysis.\u003c/p\u003e\u003ch2\u003eCell composition estimation for sequencing-based methylation data of neonatal CB\u003c/h2\u003e\u003cp\u003eTo address cell-type heterogeneity in CB-derived DNA methylation data, we developed a cell-type composition estimation model optimized for sequencing-based data. This model incorporates reference methylation profiles for seven blood cell types, including WGBS data obtained from purified nRBCs from the CB of Japanese newborns. Reference profiles for six other cell types (natural killer cells, B cells, CD4\u003csup\u003e+\u003c/sup\u003e T cells, CD8\u003csup\u003e+\u003c/sup\u003e T cells, monocytes, and neutrophils) were obtained from previously established datasets. The estimation was implemented by modifying the “estimateCellCounts” function from the R package minfi. The DNA methylation profiles of nRBCs were compared with those of other blood cell types to confirm their distinct epigenetic signatures (see Supplementary Information for full methods and results). A cell composition estimation model for adult PB was developed and utilized in our previous study \u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. The WGBS summary data for nRBCs are publicly available in the iMETHYL database \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003ch2\u003eEpigenome‑wide association study\u003c/h2\u003e\u003cp\u003eTo investigate the potential transmission of smoking-related DNA methylation signatures across generations, we conducted an EWAS focusing on maternal and grandmaternal pre-pregnancy smoking. Specifically, we examined whether smoking exposure in the preceding generations was associated with neonatal DNA methylation patterns and compared CpG sites identified across generations to assess potential transgenerational epigenetic inheritance.\u003c/p\u003e\u003cp\u003eCpG sites with a CR of \u0026lt; 95% were excluded. All EWAS were performed using linear regression models, with DNA methylation beta values as the dependent variable and smoking-related traits as the independent variables. The Bonferroni method was used to correct multiple tests. The results were visualized using the qqman and ggplot2 R packages, and CpG sites were annotated based on genomic features using the AnnotationHub, GenomicRanges, rtracklayer, and org.Hs.eg.db R packages.\u003c/p\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eNeonatal EWAS1 – Maternal pre-pregnancy smoking\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eMaternal smoking status before pregnancy was categorized into three groups: never smokers, former smokers who quit within 5 years, and those who quit ≥ 5 years ago. The model was adjusted for the estimated cell-type composition (natural killer cells, B cells, CD4 + T cells, CD8 + T cells, monocytes, neutrophils, and nRBCs), maternal parity (primiparous or multiparous), sex, household income (\u0026lt; 4 or ≥ 4\u0026nbsp;million yen), and partner smoking status. Additionally, to assess the performance of the cell-type composition model developed for CB, an EWAS for pre-pregnant maternal smoking was conducted using cell-type estimates derived from an adult PB model. Lambda (λ) values from both models were compared to evaluate the effectiveness of cell-type adjustment in controlling statistical inflation.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eNeonatal EWAS2 – Maternal grandmother’s (MGM’s) pre-pregnancy smoking\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eMGM’s smoking status before pregnancy was categorized as a binary variable (never smokers vs ever smokers), as only one participant had a grandmother who had quit smoking ≥ 5 years before pregnancy. This model was further adjusted for maternal pre-pregnancy smoking, in addition to the covariates described in Neonatal EWAS1.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eGrandparental EWAS1 – Own smoking status\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eSmoking status was treated as a binary trait, comparing ever smokers (including former and current smokers) with never smokers in an EWAS conducted among grandparents (MGM, MGF, PGM, and PGF). The model was adjusted for the estimated cell-type composition (natural killer cells, B cells, CD4 + T cells, CD8 + T cells, monocytes, and neutrophils), age, sex, and household income. If the household income information for an individual was unavailable, the corresponding partner’s household income data were used as a proxy for adjustment.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eGrandparental EWAS2 – Own smoking cessation status\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eSmoking cessation was treated as a three-category variable representing smoking cessation status (never smokers, former smokers who quit within 5 years, and those who quit ≥ 5 years ago). These categorizations were based on Fang et al. (2023) \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. The EWAS was performed using the same covariates as in Grandparental EWAS1.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003ch2\u003eStatistical analyses of smoking-associated CpG sites\u003c/h2\u003e\u003cp\u003eTo evaluate the DNA methylation status of smoking-related CpG sites across generations, we compared the CpG sites identified in each EWAS (smoking, smoking cessation, pre-pregnant maternal smoking, and pre-pregnant MGM’s smoking). We visualized their overlap using Venn diagrams implemented in R (GenomicRanges, rtracklayer, VennDiagram, and UpSetR packages).\u003c/p\u003e\u003cp\u003eTo compare the methylation levels between neonates and adults (grandparents), we focused on the CpG sites identified in this study as smoking-associated DNA methylation markers. Four groups were included: neonates born to mothers with a history of pre-pregnancy smoking (M_Smk, n = 31), neonates born to nonsmoking mothers (M_Ctr, n = 114), currently smoking grandparents (Smk, n = 86), and never-smoking grandparents (Ctr, n = 299). Cell-type composition bias was corrected using a linear regression model fitted separately for each CpG site, with the DNA methylation β-value as the dependent variable and estimated proportions of six or seven cell types as the independent variables. The adjusted methylation value was calculated as the model residual plus the original mean β-value (residual + mean method). The cell-type composition was estimated using reference models appropriate for CB and adult PB. Two-way analysis of variance was performed on the adjusted β-values for each CpG site, testing for the main effects of smoking status (smoking: yes/no), age group (AgeGroup: adult/neonatal), and their interaction. \u003cem\u003eP\u003c/em\u003e-values were corrected for multiple testing using the Bonferroni method (n = 156), with adjusted \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 considered statistically significant.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSummary data of DNA methylation profiles in the Hepta-family cohort and WGBS summary data of nRBCs from the CB of Japanese newborns are available in the iMETHYL database (https://imethyl.ihec-epigenomes.org/). Owing to ethical considerations, including the protection of participant privacy and the prevention of unintended identification, individual-level data from the TMM BirThree Cohort Study are not publicly available. Access to these datasets may be granted upon request and is subject to approval from the Ethics Committee of IMU and the Materials and Information Distribution Review Committee of the TMM Project. Researchers interested in accessing these datasets must contact the corresponding author to initiate the request process.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003cstrong\u003ecknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank all the participants who provided specimens and data. We are also grateful to Miyuki Horie, Yukino Nakamura, Hiroko Nakamura, and Anna Kudo for their assistance with the experiments. This study was supported by the Tohoku Medical Megabank Project (Special Account for the Reconstruction of the Great East Japan Earthquake) of the Ministry of Education, Culture, Sports, Science and Technology and Japan Agency for Medical Research and Development (AMED) (grant numbers JP19km0105004 and JP20km0105004). Supercomputer resources were provided by an AMED Research Grant (JP20km0405001). This work was also supported by the Japan Endocrine Society Grant for Promising Investigator and JSPS KAKENHI Grant-in-Aid for Young Scientists (grant number 23K14437).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eS. Minabe designed the study, performed analyses, and wrote the manuscript. H.O. managed the Hepta-family experiments and drafted the nRBC methods. S.U. coordinated WGBS experiments. E.K., K. Kikuchi, G. H., M. T., C. I., H. Kawamura, T. S., S. H. obtained informed consent and collected umbilical cord blood. R. O. coordinated scheduling with obstetricians and participants. A.T. processed sequencing data. K.O., T.A., and K.F. conducted laboratory work. S. Komaki managed the iMETHYL database. K. Kumada, S. Mizuno, and H. Kudo handled biospecimen provision. S.T., M.I., and T.O. managed cohort data. F.K., S.O., and K. Kinoshita. supported supercomputer infrastructure. T. B. and A. S. conceived the plan for using nRBC in this study and managed its implementation. M.Y. and S. Kuriyama planned the Hepta-family project and oversaw the cohort. H.O., A.T., K.O., T.A., K.F., S. Komaki, Y.S., Y.O., A.S. reviewed and revised the manuscript. A.S. supervised the project. All authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have nothing to disclose.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBarker DJ (2007) The origins of the developmental origins theory. J Intern Med 261:412\u0026ndash;417\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGluckman PD, Hanson MA, Cooper C, Thornburg KL (2008) Effect of in utero and early-life conditions on adult health and disease. N Engl J Med 359:61\u0026ndash;73\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Dijk SJ et al (2015) Epigenetics and human obesity. Int J Obes (Lond) 39:85\u0026ndash;97\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePembrey M, Saffery R, Bygren LO, Network (2014) in Epigenetic, E. \u0026amp; Network in Epigenetic, E. Human transgenerational responses to early-life experience: potential impact on development, health and biomedical research. \u003cem\u003eJ Med Genet\u003c/em\u003e 51, 563\u0026thinsp;\u0026ndash;\u0026thinsp;72\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePembrey ME (2010) Male-line transgenerational responses in humans. Hum Fertil (Camb) 13:268\u0026ndash;271\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHeijmans BT et al (2008) Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A 105:17046\u0026ndash;17049\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoubert BR et al (2016) DNA Methylation in Newborns and Maternal Smoking in Pregnancy: Genome-wide Consortium Meta-analysis. Am J Hum Genet 98:680\u0026ndash;696\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAnway MD, Cupp AS, Uzumcu M, Skinner MK (2005) Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 308:1466\u0026ndash;1469\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHoile SP, Lillycrop KA, Thomas NA, Hanson MA, Burdge GC (2011) Dietary protein restriction during F0 pregnancy in rats induces transgenerational changes in the hepatic transcriptome in female offspring. PLoS ONE 6:e21668\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFranklin TB et al (2010) Epigenetic transmission of the impact of early stress across generations. \u003cem\u003eBiol Psychiatry\u003c/em\u003e 68, 408\u0026thinsp;\u0026ndash;\u0026thinsp;15\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHebbring S (2019) Genomic and Phenomic Research in the 21st Century. Trends Genet 35:29\u0026ndash;41\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePearson H (2015) Massive UK baby study cancelled. Nature 526:620\u0026ndash;621\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuriyama S et al (2020) Cohort Profile: Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study): rationale, progress and perspective. Int J Epidemiol 49:18\u0026ndash;19m\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMiyake K et al (2018) Association between DNA methylation in cord blood and maternal smoking: The Hokkaido Study on Environment and Children's Health. Sci Rep 8:5654\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWorld Health O (2004) United Nations Children's. F. Low birthweight: country, regional and global estimates. World Health Organization, Geneva\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHachiya T et al (2017) Genome-wide identification of inter-individually variable DNA methylation sites improves the efficacy of epigenetic association studies. NPJ Genom Med 2:11\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKomaki S et al (2018) iMETHYL: an integrative database of human DNA methylation, gene expression, and genomic variation. Hum Genome Var 5:18008\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHoang TT et al (2024) Comprehensive evaluation of smoking exposures and their interactions on DNA methylation. EBioMedicine 100:104956\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOhmomo H et al (2022) DNA Methylation Abnormalities and Altered Whole Transcriptome Profiles after Switching from Combustible Tobacco Smoking to Heated Tobacco Products. Cancer Epidemiol Biomarkers Prev 31:269\u0026ndash;279\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOhmomo H et al (2022) Potential DNA methylation biomarkers for the detection of clear cell renal cell carcinoma identified by a whole blood-based epigenome-wide association study. Epigenetics Commun 2:1\u0026ndash;11\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJones MJ et al (2025) DNA methylation differences between cord and adult white blood cells reflect postnatal immune cell maturation. Commun Biol 8:237\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZeilinger S et al (2013) Tobacco smoking leads to extensive genome-wide changes in DNA methylation. PLoS ONE 8:e63812\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMay JE, Marques MB, Reddy VVB, Gangaraju R (2019) Three neglected numbers in the CBC: The RDW, MPV, and NRBC count. Cleve Clin J Med 86:167\u0026ndash;172\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGervin K et al (2016) Cell type specific DNA methylation in cord blood: A 450K-reference data set and cell count-based validation of estimated cell type composition. Epigenetics 11:690\u0026ndash;698\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ede Goede OM et al (2015) Nucleated red blood cells impact DNA methylation and expression analyses of cord blood hematopoietic cells. Clin Epigenetics 7:95\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang L et al (2022) Maternal cigarette smoking before or during pregnancy increases the risk of birth congenital anomalies: a population-based retrospective cohort study of 12 million mother-infant pairs. BMC Med 20:4\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTang Z et al (2024) Former smoking associated with epigenetic modifications in human granulosa cells among women undergoing assisted reproduction. Sci Rep 14:5009\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLaffan SB, Posobiec LM, Uhl JE, Vidal JD (2018) Species Comparison of Postnatal Development of the Female Reproductive System. Birth Defects Res 110:163\u0026ndash;189\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuriyama S et al (2016) The Tohoku Medical Megabank Project: Design and Mission. J Epidemiol 26:493\u0026ndash;511\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTohoku M, Megabank O (2025) Birth and Three-Generation Cohort Study. Vol. (Tohoku University, 2025)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKomaki S et al (2023) Epigenetic profile of Japanese supercentenarians: a cross-sectional study. Lancet Healthy Longev 4:e83\u0026ndash;e90\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFang F, Andersen AM, Philibert R, Hancock DB (2023) Epigenetic biomarkers for smoking cessation. Addict Neurosci 6\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 and 2 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7314319/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7314319/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eTransgenerational epigenetic inheritance has been demonstrated in rodent models but not in humans. To address this gap, we established a comprehensive DNA methylation resource derived from 158 three-generation Japanese families. The dataset integrates genome-wide methylation profiles with extensive clinical and lifestyle data. Using targeted bisulfite sequencing, we profiled \u0026gt; 1 million CpG sites across the genome, covering the promoter and gene body regions of \u0026gt; 16,500 annotated genes, in 938 adult peripheral blood and 155 neonatal cord blood samples. To demonstrate the utility of this resource, we performed a representative analysis focusing on the intergenerational impact of maternal and grandmaternal pre-pregnancy smoking. We identified persistent methylation marks in neonates associated with ancestral smoking history, suggesting the potential transgenerational transmission of environmental effects in humans. This multigenerational epigenomic resource provides a valuable foundation for future studies on intergenerational epigenetic mechanisms and their role in shaping human health trajectories.\u003c/p\u003e","manuscriptTitle":"Mapping the human epigenetic landscape across three generations: A DNA methylation resource from TMM BirThree","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-20 08:15:18","doi":"10.21203/rs.3.rs-7314319/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e0a9c68a-7de7-4781-a0da-c31f15dfad77","owner":[],"postedDate":"August 20th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":53395888,"name":"Health sciences/Risk factors"},{"id":53395889,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":53395890,"name":"Biological sciences/Genetics/Epigenetics/DNA methylation"},{"id":53395891,"name":"Biological sciences/Genetics/Epigenomics"}],"tags":[],"updatedAt":"2025-10-08T16:45:55+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-20 08:15:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7314319","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7314319","identity":"rs-7314319","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}