Epigenome-wide analysis of DNA-methylation signatures following climate-related disasters.

Epigenome-wide analysis of DNA-methylation signatures following climate-related disasters - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Med . 2026 Apr 20;24:336. doi: 10.1186/s12916-026-04868-y Epigenome-wide analysis of DNA-methylation signatures following climate-related disasters Wenzhong Huang 1, 2, # , Rongbin Xu 3, 4, 5, # , Yao Wu 2 , Zhengyu Yang 2 , Zhoufeng Ye 6 , Ee Ming Wong 7, 8 , Melissa C Southey 4, 7, 8 , John L Hopper 6 , Michael J Abramson 5 , Shanshan Li 2 , Shuai Li 6, 7, ✉, # , Yuming Guo 2, ✉, # Author information Article notes Copyright and License information 1 Key Laboratory of Environmental Medicine and Engineering of Ministry of Education, and School of Public Health, Southeast University, Nanjing, Jiangsu, China 2 Climate, Air Quality Research Unit, School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia 3 Chongqing Emergency Medical Center, Chongqing University Central Hospital, School of Medicine, Chongqing University, Chongqing, China 4 Cancer Epidemiology Division, Cancer Council Victoria, Melbourne, VIC 3004 Australia 5 School of Public Health and Preventive Medicine, Monash University, Melbourne, VIC 3004 Australia 6 Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, VIC 3010 Australia 7 Precision Medicine, School of Clinical Sciences at Monash Health, Monash University, Clayton, VIC 3800 Australia 8 Department of Clinical Pathology, Melbourne Medical School, The University of Melbourne, Melbourne, VIC 3010 Australia ✉ Corresponding author. # Contributed equally. Received 2025 Dec 19; Accepted 2026 Apr 9; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13224649  PMID: 42010415 Abstract Background Floods and tropical cyclones (TCs), two of the most frequent and costliest climate-related disasters worldwide, have been linked to sustained health risks extending beyond acute hazards. However, evidence on the underlying epigenetic mechanisms remains scarce. We aimed to characterize DNA methylation patterns associated with exposure to floods and TCs of varying intensities. Methods We collected peripheral blood samples from 479 women (132 twin pairs and 215 of their sisters) across Australia. Blood-derived DNA methylation profiles were assessed using the Illumina HumanMethylation450 BeadChip array. Daily flood and TC exposure data for the 6 years preceding each blood draw were obtained from the Dartmouth Flood Observatory and the International Best Track Archive for Climate Stewardship, respectively, and linked to participants based on residential addresses. Using a within-sibship analytical framework that accounted for shared familial factors and other relevant covariates, we examined associations between flood and TC exposures of varying intensities and site-specific methylation at each cytosine-guanine dinucleotide (CpG). Differentially methylated regions (DMRs) were identified using a combination of the comb-p and DMRcate algorithms. Results There were 164 CpGs and 219 DMRs associated with flood and TC exposures (Bonferroni-adjusted p value < 0.05), mapping to 242 genes enriched in pathways related to inflammation and immune regulation. These genes have been implicated in a wide range of human diseases or phenotypes. The number of differentially methylated CpGs increased with more recent and higher-intensity exposures. Intensity-dependent gene regulation was observed, with genes such as AMT and C22orf45 consistently implicated across various exposure levels, whereas RNF39 and ACY3 emerged only at higher intensities. Conclusions Exposures to floods and TCs were associated with differentially DNA methylated signals across the human genome, exhibiting intensity-dependent patterns. The identified signals and related gene pathways may shed light on the biological mechanism underlying the profound health effects of climate-related disasters. Supplementary Information The online version contains supplementary material available at 10.1186/s12916-026-04868-y. Keywords: Epigenome-wide analysis, Climate-related disasters, DNA methylation, Gene regulation Background Floods and tropical cyclones (TCs) constitute two of the most prevalent and severe climate-related disasters globally [ 1 , 2 ]. Each year, floods and TCs collectively affect more than 600 million individuals globally and the impacts of these events have escalated substantially [ 3 , 4 ]. TCs alone lead to an average of US$78 million in daily damages, and cumulative losses in the current decade (around US$460 billion) are approximately double those recorded between 1980 and 1999 [ 5 ]. Emerging epidemiological evidence, extending from single-event case studies [ 6 – 9 ] to multi-event, cross-regional systematic analyses in recent years [ 10 – 16 ], associates flood and TC exposures with elevated risks for a broad spectrum of adverse outcomes beyond acute injuries, including prolonged increased risks of chronic diseases (e.g., cardiopulmonary diseases [ 16 – 18 ], diabetes [ 1 , 16 ], mental illness [ 19 , 20 ]) and cause-specific mortality [ 10 , 12 – 15 , 21 ]. These disasters are likely to remain important public health concerns and key drivers of climate-related hazards, particularly given projections of increased intensity as the climate warms [ 22 – 24 ]. However, the underlying biological mechanism by which these events affect human health remains largely unclear. Epigenetic modification, particularly DNA methylation, serves as a fundamental modifiable biological process through which environmental factors influence human health by regulating gene expression without altering the underlying DNA sequence. The indirect, persistent, or delayed adverse impacts of floods and TCs (e.g., via contaminated water supplies, population displacement, and associated physical, financial, and psychological hardships) may be manifested in altered DNA methylation profiles [ 1 , 2 ]. An experimental study on a cohort of rhesus macaques observed accelerated epigenetic aging following exposure to Hurricane Maria [ 25 ]. The exposed macaques exhibited biological ages approximately 2 years older than their unexposed counterparts, which corresponded to around 7 to 8 years in the human lifespan. However, to our best knowledge, no study has examined DNA methylation alterations in human following exposures of climate disasters like floods or TCs. In the present study, by leveraging blood DNA methylation data from participants across Australia and incorporating data on multiple flood and TC events, we aimed to evaluate the associations between disaster exposure and genome-wide methylation patterns, and to examine whether methylation changes varied by event intensity and timing. Methods Study design and sample We implemented a twin and family design by including 132 female twins and their sisters (66 monozygotic twin pairs, 66 dizygotic twin pairs, and 215 sisters of these twins) from 130 families across Australia through the Australian Twin Registry between 2004 and 2009. This design enabled within-family comparisons between each participant and her siblings, thereby reducing potential confounding by genetic background and shared familial factors (e.g., familial dietary behavior, childhood living environment) [ 26 , 27 ]. Participants were originally recruited to investigate environmental, genetic, and lifestyle factors associated with breast cancer, with the following inclusion criteria: (1) aged between 40 and 70 years; (2) not pregnant or breastfeeding; (3) without breast cancer. Details of the data collection procedures have been described previously [ 28 – 31 ]. DNA methylation data Blood samples were collected from each participant for DNA methylation measurements. Specifically, each participant provided a 27 mL peripheral whole blood sample via a visit to a designated pathology laboratory. Trained phlebotomists conducted blood collection at participants’ homes in cases where they were unable to access the designated laboratory. All samples were dispatched to the biobank for processing within 48 h of collection. DNA methylation was measured using the Illumina Infinium HumanMethylation450 BeadChip arrays [ 32 ]. Raw data preprocessing and quality control were conducted using the Bioconductor minfi package, following three standardized steps [ 33 ]: (1) correction for probe type I and II bias via subset-quantile within array normalization [ 34 ]; (2) adjustment for batch effects using an empirical Bayes approach (ComBat) [ 35 ]; and (3) normalization by Illumina reference factor-based methods ( preprocessIllumina ). Probes were excluded if the detection p value exceeded 0.01 in any sample, if they bound documented SNPs or the X chromosome, if they were non-specific to a single location, or if they were among negative control probes [ 31 ]. X-chromosome probes were excluded to avoid the biological and technical noise because X-linked methylation in females is strongly influenced by X-chromosome inactivation (XCI), mosaicism and age/tissue-related XCI skewing, and because many 450 K X-probes show higher technical noise or cross-reactivity [ 31 ]. A final of 411,394 probes were retained for analysis. Each probe measured the average methylation status at a specific cytosine-guanine dinucleotide (CpG), reported as a beta value ranging from 0 (unmethylated) to 1 (fully methylated). Raw and processed methylation data are available via Gene Expression Omnibus (accession number: GSE100227 ). To estimate global DNA methylation, we first calculated the mean beta value across all retained CpGs (i.e., the genome-wide average DNA methylation). Additional metrics included the average methylation of CpGs in three major repetitive element types: Arthrobacter luteus (Alu), long interspersed elements-1 (L1), and long terminal repeats (LTR). Given the partial coverage of repetitive element-associated CpGs on the Illumina 450 K array, a previously validated random forest algorithm was applied to predict methylation levels of unprofiled repetitive element-associated CpGs using methylation of proximal profiled CpGs, inter-CpG distances, and genomic features of the target CpG and repetitive element [ 36 ]. Predicted CpGs were then integrated with profiled CpGs to derive genome-wide average methylation levels for Alu, L1, and LTR elements. Thus, seven metrics (i.e., the genome-wide CpG mean, the profiled and predicted methylation levels for Alu, L1, and LTR) were used to measure global DNA methylation. Flood and TC exposure assessment We obtained flood data between 1997 and 2009 from the Dartmouth Flood Observatory (DFO), which is a comprehensive and validated collection of global flood events from satellite observations, news, government, and instrumental sources [ 37 , 38 ]. DFO provides the spatial coverage (i.e., the flooded areas), occurrence date, and duration of each flood event, and has been commonly used in previous studies of flood climatology and health risk assessments [ 10 , 39 , 40 ]. The DFO classified flood events into three levels (large, very large, and extreme flood events) based on structural and agricultural damage, associated fatalities, and the return period. In this study, a participant was considered to be exposed to flood if their home was within the flooded areas [ 10 ]. For TC exposure assessment, we applied an improved wind field model [ 41 ] to estimate the historical temporal dynamics of the windspeed associated with TCs. This wind field model has been widely used in previous studies [ 42 – 44 ] and has been detailed in our previous work [ 15 ]. Briefly, based on the global TC information between 1997 and 2009 from the International Best Track Archive for Climate Stewardship (IBTrACS) [ 45 ], we used a parametric wind model to generate gridded hourly wind fields (i.e., daily maximum of 1-min sustained wind speeds associated with the TC) for each TC event in IBTrACS with a spatial resolution of 1/12° (~ 10 km). For this analysis, we aggregated the data to 0.5° (~ 55 km) spatial and daily temporal resolution, matching the local time zones of our collected data. Grid cells with a TC-related maximum sustained wind speed of at least 34 knots (i.e., 17.5 m/s) were considered to be TC-exposed [ 13 , 17 ]. We defined TC exposure for each study participant if the grid cell containing their residential address was TC-exposed [ 10 , 15 , 46 ]. TC exposures were further classified into gale to violent storms (TC-related maximum sustained wind speed ≥ 48 knots) and hurricanes (≥ 64 knots) [ 13 , 17 ]. Finally, our primary exposure, long-term flood and TC exposure, was defined for each participant as the total number of flood or TC exposure days during the 12 months before their blood draw. We constructed this combined exposure based on two main considerations: (1) floods and TCs frequently occur concurrently and are closely interrelated, leading to highly overlapping secondary impacts (e.g., infrastructure damage, disruption of medical services, and socio-economic displacement) that can drive long-term health effects; and (2) from a biological perspective, we hypothesized that DNA methylation and epigenetic aging responses are predominantly driven by these shared downstream stressors rather than by the specific meteorological trigger. Given that residential histories were not collected, we assumed that the participants lived at the same residence during the 12 months before blood collection. This assumption was supported by a national survey conducted in 2007–2008 that reported that nearly 80% of Australians 40–79 years of age had not changed residences in the last 5 years [ 47 ]. Each exposure day was categorized into three levels according to the flood and TC exposure levels described above: level 1 (days with any flood or TC occurrences), level 2 (days characterized by very large flood events or gale to violent storms), and level 3 (days with extreme flood events or hurricanes). Assessment of covariates We collected demographic information and self-reported lifestyles from each participant through a telephone-administered questionnaire. Detailed questions were reported in our previous work [ 27 ]. The telephone interview was conducted within 1 year before the blood draw for the vast majority of participants (95%, 457/479), and 1–2 years before the blood draw or up to 3 months after the blood draw for the remaining participants [ 27 ]. Consistent with previous studies [ 48 , 49 ], we included the following covariates: chronological age (the time elapsed in years from the participant’s birth to the date of blood sample collection), educational level (secondary or below, vocational training, or university), marital status (married or living as married, never married, and widowed/separated/divorced), smoking and drinking status (current, former, and never), area-level socioeconomic status (SES), proportions of leukocyte cells, and annual mean temperature and relative humidity (RH). Area-level SES was calculated using the Index of Relative Socioeconomic Advantage and Disadvantage (IRSAD) [ 49 ], which is a composite measure comprising 21 indicators of community-level SES (e.g., household income, internet connection, occupation, and education). A higher IRSAD indicated greater socioeconomic advantage. Leukocyte cell composition, including the proportions of CD8 + T cells, CD4 + T cells, natural killer cells, B cells, monocytes, and granulocytes, was included to account for potential confounding by cell type due to their distinct methylation signatures [ 50 ]. We further collected and defined each participant’s urban–rural status based on their residential addresses according to the Australian Statistical Geography Standard 2011 (the earliest year with such data available) [ 51 ]. Urban areas were categorized as “major urban” (urban centers with a population of 100,000 or more) or “other urban” (urban centers with a population between 1000 and 99,999). Each participant’s annual mean temperature and RH were calculated based on a comprehensive Australian climate data archive with a spatial resolution of 0.05° × 0.05° [ 52 ], representing the average ambient temperature and RH within the year preceding their blood sample collection. Statistical analysis Global and CpG-specific analyses We employed within-sibship analyses to investigate associations between global and CpG-specific DNA methylation and long-term flood and TC exposures. Specifically, we constructed generalized estimating equation (GEE) models with an identity link function and an independent correlation structure to account for familial clustering, specified as follows: E Y ij = β 0 + β within ∗ ( Expo ij - Expo ¯ i ) + β between ∗ Expo ¯ i + β ∗ X ij where Y ij represents the methylation level of a specific CpG site or a global DNA methylation metric for subject j in family i ; Expo ij denotes the long-term flood and TC exposures for subject j from family i (i.e., the total number of flood or TC exposure days during the year preceding blood draw); and Expo ¯ i is the average exposure days of all subjects in family i . In line with previous studies [ 13 , 18 ], we assumed a linear relationship between the number of exposure days and DNA methylation. The X ij represents a set of covariates as detailed above, including age, educational level, marital status, smoking and drinking status, area-level SES, temperature, RH, survey year, season, and cell-type proportions. Missing values (two for educational level and one for area-level SES) were imputed using five-fold multiple imputation with default settings in the mice R package [ 53 ]. The regression coefficients β within and β between represent the association estimates for within- and between-sibship comparisons, respectively, quantifying the average change in methylation level per additional flood or TC exposure day. We focused on the within-sibship estimate β within , as it was less likely to be confounded by shared genetic and familial environmental factors. We considered CpGs with Bonferroni-adjusted p values < 0.05 as statistically significant. CpGs with Benjamini–Hochberg false discovery rates (FDRs) < 0.05 were also reported [ 54 ]. The potential inflation or deflation of p values for CpG-specific associations was assessed using Quantile–Quantile (Q-Q) plots and Bayesian inflation factors (BIF) calculated with the bacon R package [ 55 ]. Analyses were conducted separately for each exposure intensity level (levels 1, 2, and 3) of flood and TC exposures. We conducted following sensitivity analyses to test the result robustness: (1) re-performing the analyses by applying alternative exposure windows ranging from 1 to 6 years prior to blood collection to represent long-term flood and TC exposures; (2) treating exposure as a binary variable (exposed vs unexposed) to evaluate the potential influence of the skewed, zero-inflated exposure distribution on the effect estimates; (3) additionally adjusting for participants’ urban–rural status (major urban, other urban, rural) to assess potential confounding by residential context or regional differences in urbanicity; (4) re-estimating the models excluding adjustments for smoking status and alcohol consumption, given the lack of conclusive evidence linking these behaviors to climate-disaster exposure; (5) re-estimating the models without adjustment for blood cell-type proportions to evaluate extrinsic differences in DNA methylation associated with flood and TC exposure; these extrinsic differences encompass variation in DNA methylation among leukocyte subpopulations, as well as cell-intrinsic methylation changes occurring independently of leukocyte composition [ 56 ]; (6) to account for possible misclassification in self-reported smoking status, re-estimating the models using a DNA methylation–derived smoking index, constructed from nine smoking-associated CpG sites [ 57 ], in place of self-reported smoking status. Differentially methylated region (DMR) analyses We employed two widely used methods, comb-p (version 0.32) [ 58 ] and DMRcate (version 1.18.0) [ 59 ], to identify DMRs associated with flood and TC exposures at varying intensity levels. Comb-p combined low p values of CpG sites within sliding windows to detect significant DMRs, applying the Šidák correction to adjust for multiple testing. In contrast, DMRcate utilized a tunable kernel smoothing approach on CpG-specific effect estimates to identify significant DMRs. Given their distinct algorithms and complementary detection principles, DMRs detected by both methods are less likely to represent false positives compared to those identified by only one method [ 49 , 60 , 61 ]. Input parameters for both methods were consistent with prior studies [ 62 , 63 ] and were detailed in Additional file 1: Table S1. DMRs were considered significant if they met all of the following criteria: (1) consisted of at least two CpGs; (2) had adjacent CpGs spaced less than 1000 base pairs apart; and (3) surpassed the significance thresholds for both comb-p (Šidák p value < 0.01) and DMRcate (FDR < 0.01), thereby minimizing false-positive findings. Gene-disease associations and pathway enrichments We used HumanMethylation 450 microarray annotation to annotate significant CpGs and DMRs with human genes. These annotated genes then queried in DisGeNET (version 25.1, https://disgenet.com/ ), one of the largest gene-disease association databases, to identify linked diseases and phenotypes [ 64 ]. Pathway and Gene Ontology (GO) enrichment analyses were performed with the “gometh” function from the missMethyl R package [ 65 , 66 ]. This approach mapped CpGs to genes, then tested for overrepresentation of gene sets in KEGG pathways and GO terms using Wallenius’ non-central hypergeometric distribution. The key advantage of “gometh” over alternative methods was its adjustment for inherent bias in Illumina methylation arrays, where genes with more measured CpG sites had a higher likelihood of detection as differentially methylated [ 65 ]. This correction yielded more unbiased gene set enrichment results. Only GO terms and pathways meeting nominal enrichment p < 0.05 and containing at least three mapped genes were considered significant. Input for enrichment included all Bonferroni-significant CpGs plus all CpGs from significant DMRs. To facilitate interpretation, redundant GO terms were reduced using the “simplify” function from clusterProfiler , which employs semantic similarity measures to retain only representative terms from highly similar groups [ 67 , 68 ]. All analyses were performed in R software (version 4.1.0) with packages “geepack” (for GEE model, version 1.3–2), “ENmix” (for comb-p analyses, version 1.28.8), “mice” (multiple imputation, version 3.13.0), “DMRcate” (DMRcate analyses, version 1.18.0), “REMP” (extract and predict repetitive elements), “missMethyl” (pathway enrichment, version 1.24.0), “bacon” (Q-Q plot, version 1.18.0), and “clusterProfiler” (remove redundant GO terms, version 3.18.1). Results Descriptive statistics We included 479 women aged 40 to 78 years (mean ± standard deviation [SD]: 56.4 ± 7.9 years) from diverse regions across Australia (Table 1 ). Educational attainment ranged from primary and secondary school (198/479, 41.3%) to university-level education (144/479, 30.1%). Most of the participants were married or in a de facto relationship (369/479, 77.0%). The majority of participants were never-smokers (291/479, 60.8%) or former smokers (147/479, 30.7%). Approximately half of participants were current alcohol drinkers (235/479, 49.1%). Table 1. Summary of the characteristics of the study participants by Australian states and territories Characteristics Overall VIC NSW QLD WA Other a Women, n (%) 479 (100.0%) 142 (100.0%) 137 (100.0%) 128 (100.0%) 41 (100.0%) 31 (100.0%) Age (y), mean (SD) 56.4 (7.9) 55.3 (7.5) 57.8 (7.8) 56.8 (8.5) 55.1 (6.7) 55.5 (8.5) Age group (y), n (%) < 65 years 402 (83.9%) 123 (86.6%) 111 (81.0%) 105 (82.0%) 38 (92.7%) 25 (80.6%) ≥ 65 years 77 (16.1%) 19 (13.4%) 26 (19.0%) 23 (18.0%) 3 (7.32%) 6 (19.4%) Education level, n (%) Secondary or below 198 (41.3%) 50 (35.2%) 59 (43.1%) 56 (43.8%) 17 (41.5%) 16 (51.6%) Vocational 135 (28.2%) 36 (25.4%) 41 (29.9%) 40 (31.2%) 11 (26.8%) 7 (22.6%) University 144 (30.1%) 55 (38.7%) 36 (26.3%) 32 (25.0%) 13 (31.7%) 8 (25.8%) Area-level SES percentile in Australia, n (%) (0, 25] 88 (18.4%) 15 (10.6%) 32 (23.4%) 30 (23.4%) 5 (12.2%) 6 (19.4%) (25, 50] 103 (21.5%) 33 (23.2%) 20 (14.6%) 32 (25.0%) 12 (29.3%) 6 (19.4%) (50, 75] 136 (28.4%) 48 (33.8%) 33 (24.1%) 37 (28.9%) 10 (24.4%) 8 (25.8%) (75, 100] 151 (31.5%) 46 (32.4%) 51 (37.2%) 29 (22.7%) 14 (34.1%) 11 (35.5%) Marital status, n (%) Married or de facto 369 (77.0%) 112 (78.9%) 98 (71.5%) 95 (74.2%) 36 (87.8%) 28 (90.3%) Never married 28 (5.9%) 9 (6.3%) 9 (6.6%) 8 (6.3%) 2 (4.9%) 0 (0.0%) Widowed/separated/divorced 82 (17.1%) 21 (14.8%) 30 (21.9%) 25 (19.5%) 3 (7.3%) 3 (9.7%) Smoking behavior, n (%) Current 41 (8.6%) 10 (7.0%) 17 (12.4%) 9 (7.0%) 0 (0.0%) 5 (16.1%) Former 147 (30.7%) 44 (31.0%) 38 (27.7%) 35 (27.3%) 17 (41.5%) 13 (41.9%) Never 291 (60.8%) 88 (62.0%) 82 (59.9%) 84 (65.6%) 24 (58.5%) 13 (41.9%) Alcohol use, n (%) Current 235 (49.1%) 77 (54.2%) 66 (48.2%) 58 (45.3%) 23 (56.1%) 11 (35.5%) Former 52 (10.9%) 11 (7.8%) 14 (10.2%) 18 (14.1%) 3 (7.3%) 6 (19.4%) Never 192 (40.1%) 54 (38.0%) 57 (41.6%) 52 (40.6%) 15 (36.6%) 14 (45.2%) Proportions of cell type, mean (SD) CD8 + T 9.0 (3.7) 8.5 (3.6) 8.8 (3.7) 9.5 (3.7) 9.3 (4.1) 9.4 (4.1) CD4 + T 15.2 (5.7) 15.1 (6.4) 15.0 (5.6) 16.1 (5.3) 13.4 (4.6) 15.1 (6.0) Natural killer (NK) 6.5 (4.3) 5.9 (4.6) 6.5 (4.3) 7.2 (4.0) 6.4 (4.6) 6.3 (3.7) B cells 5.4 (2.3) 5.1 (2.5) 5.4 (2.2) 5.6 (2.1) 5.9 (2.4) 5.6 (2.5) Monocytes 6.8 (2.2) 7.6 (2.3) 6.4 (2.1) 6.4 (2.1) 6.9 (1.9) 6.7 (2.1) Granulocytes 59.4 (7.9) 59.8 (7.9) 60.2 (7.9) 57.7 (8.3) 60.8 (7.2) 59.0 (7.5) Annual average ambient temperature (°C), mean (SD) 17.7 (2.8) 14.9 (0.9) 17.7 (1.4) 21.3 (1.6) 17.7 (1.2) 15.4 (1.8) Annual average relative humidity (%), mean (SD) 69.8 (3.8) 70.5 (3.1) 69.0 (3.9) 71.7 (2.8) 67.7 (2.9) 64.9 (3.9) Open in a new tab DFO Dartmouth Flood Observatory, NSW New South Wales, QLD Queensland, SD standard deviation, SES socioeconomic status, TC tropical cyclone, VIC Victoria, WA Western Australia a Other regions comprised the Australian Capital Territory, Tasmania, and South Australia Flood and TC exposure frequency varied substantially among participants, with mean total exposure days of 2.7 (SD: 6.5) for level 1, 1.6 (SD: 5.7) for level 2, and 1.5 (SD: 5.6) for level 3 within the year preceding blood sampling (Fig. 1 ). The highest exposure frequencies were observed among participants residing in Western Australia and Queensland, whereas relatively low exposure levels were predominantly clustered among those living in New South Wales. Fig. 1. Open in a new tab Geographical distribution of study participants and annual total number of flood/tropical cyclone (TC) exposure days by intensity level. Level 1 ( a ) represented days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 ( b ) denoted days with very large flood events or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); level 3 ( c ) comprised days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots) Methylation level associated with flood/TC over different exposure windows Among the exposure windows spanning 1 to 6 years, flood and TC exposures within the 1-year period preceding blood sampling exhibited the strongest associations with DNA methylation, as suggested by the greatest number of statistically significant CpG sites (Additional file 1: Table S2). A discernible trend was observed whereby more recent exposures were associated with a greater number of statistically significant CpG loci (Additional file 1: Table S2). At the level of global DNA methylation, no statistically significant associations were detected between flood/TC exposure and any of the seven global DNA methylation metrics (Additional file 1: Fig. S1 and Table S3). However, there was a trend that more recent flood/TC exposures were associated with greater reductions in global DNA methylation across these measures (Additional file 1: Fig. S1). We therefore concentrated subsequent analyses on the most recent exposure window, specifically flood and TC exposures occurring within the 1-year period prior to blood sampling. The effect estimates remained generally robust across the sensitivity analyses (Additional file 1: Figs. S2–S6). Genome-wide CpG-specific DNA methylation associated with flood/TC exposures Among the 411,394 CpGs included across the genome, flood and TC exposures at higher intensity levels generally exhibited stronger associations with DNA methylation, as suggested by the lower p values and larger effect sizes (Fig. 2 b). As demonstrated by the Q-Q plot and BIF, the p value distributions of CpG-specific associations with flood/TC exposures across intensity levels were largely consistent with the theoretical null distribution, with only mild inflation (BIF = 1.08, 1.10, and 1.11 for intensity levels 1 through 3, respectively) (Additional file 1: Fig. S7). Fig. 2. Open in a new tab Associations of global DNA methylation ( a ) and genome-wide CpG-specific DNA methylation ( b ) with flood/tropical cyclone (TC) exposures by intensity levels. Level 1 represented days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 denoted days with very large flood events or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); and level 3 indicated days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots). Notes: in a , the point and error bar represented point estimates and 95% confidence interval of the regression coefficients, respectively. The regression coefficients represented the average change in methylation level for each additional exposure day, adjusting for individual’s age, educational level, marital status, smoking behavior, alcohol use, area-level socioeconomic status, survey year, survey season, annual average ambient temperature, and annual average relative humidity, and cell-type proportions. In b , the red horizontal line represented the threshold of Bonferroni-adjusted p value equal to 0.05 Under a FDR threshold of < 0.05, we identified 2742 significant CpG sites with methylation levels associated with flood/TC exposures across varying intensity levels (509, 1876, and 2335 significant CpGs for intensity levels 1 through 3, respectively; Additional file 2). Under the stricter significance criteria (Bonferroni-adjusted p value < 0.05), 164 CpG sites remained significant, with higher exposure intensity corresponding to an increased number of significant sites (31, 97, and 135 significant CpGs for intensity levels 1 through 3, respectively; Fig. 3 and Additional file 3). The most significant associations (with the smallest Bonferroni-adjusted p value) were observed for cg15639684 (MCOLN2) at intensity level 1, and cg25457064 (ZFR2) at intensity levels 2 and 3, all of which exhibited positive directions (Table 2 ). Detailed information (e.g., genome location and coefficients of the associations) of all significant CpGs was provided in Tables S4 and S5. Fig. 3. Open in a new tab Manhattan plots for the association of genome-wide DNA methylation with flood/tropical cyclone exposures by intensity levels. Level 1 ( A ) represented days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 ( B ) denoted days with very large flood events or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); level 3 ( C ) indicated days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots). Notes: the red horizontal line represented the threshold of Bonferroni-adjusted p value equal to 0.05. The models were adjusted for individual’s age, educational level, marital status, smoking behavior, alcohol use, area-level socioeconomic status, survey year, survey season, annual average ambient temperature, and annual average relative humidity, and cell-type proportions Table 2. Associations between DNA methylation status at genome-wide CpG sites (top ten significant sites ranked by smallest Bonferroni-adjusted p values) and flood/TC exposures by intensity levels CpG CHR Position Gene Location in gene Relation to CpG islands Coefficients (× 10 −2 ) SE (× 10 −2 ) p values Bonferroni-adjusted p values Level 1 a cg15639684 1 85,464,110 MCOLN2 TSS1500 Island 0.15 0.02 2.98 × 10 −11 1.23 × 10 −5 cg25457064 19 3,825,390 ZFR2 Body Island 0.11 0.02 3.43 × 10 −11 1.41 × 10 −5 cg05641656 19 58,728,610 IGR Island 0.16 0.02 5.63 × 10 −11 2.32 × 10 −5 cg04541474 1 44,873,592 RNF220 5′UTR Island − 0.08 0.01 5.73 × 10 −10 < 0.001 cg22635075 22 19,551,881 IGR Shelf 0.08 0.01 1.55 × 10 −9 0.001 cg10424846 17 33,894,970 IGR Island 0.03 0.01 2.25 × 10 −9 0.001 cg14097568 1 94,792,942 IGR Opensea 0.13 0.02 2.93 × 10 −9 0.001 cg27319192 12 96,883,564 IGR Island 0.02 0.00 7.23 × 10 −9 0.003 cg05385164 12 54,144,855 IGR Island 0.11 0.02 1.15 × 10 −8 0.005 cg13613388 17 44,803,928 NSF Body Opensea 0.03 0.01 1.21 × 10 −8 0.005 Level 2 a cg25457064 19 3,825,390 ZFR2 Body Island 0.10 0.01 4.35 × 10 −15 1.79 × 10 −9 cg24926121 1 27,763,886 WASF2 5′UTR Opensea − 0.10 0.01 5.61 × 10 −14 2.31 × 10 −8 cg26354412 4 156,382,201 IGR Opensea − 0.05 0.01 2.45 × 10 −13 1.01 × 10 −7 cg22635075 22 19,551,881 IGR Shelf 0.08 0.01 2.17 × 10 −12 8.92 × 10 −7 cg25671977 8 40,755,559 ZMAT4 TSS1500 Shore 0.11 0.02 3.15 × 10 −12 1.29 × 10 −6 cg20775162 11 8,248,503 LMO1 Body Opensea − 0.17 0.02 6.62 × 10 −12 2.72 × 10 −6 cg22558389 4 90,225,384 GPRIN3 5′UTR Shelf − 0.09 0.01 8.03 × 10 −12 3.31 × 10 −6 cg19480937 3 16,467,544 RFTN1 Body Opensea − 0.05 0.01 1.77 × 10 −11 7.28 × 10 −6 cg09891100 19 12,893,992 IGR Island 0.16 0.02 6.05 × 10 −11 2.49 × 10 −5 cg24122124 19 5,339,092 PTPRS 5′UTR Island 0.06 0.01 1.02 × 10 −10 4.21 × 10 −5 Level 3 a cg25457064 19 3,825,390 ZFR2 Body Island 0.10 0.01 4.26 × 10 −15 1.75 × 10 −9 cg26354412 4 156,382,201 IGR Opensea − 0.06 0.01 4.44 × 10 −15 1.83 × 10 −9 cg22558389 4 90,225,384 GPRIN3 5′UTR Shelf − 0.10 0.01 3.93 × 10 −14 1.62 × 10 −8 cg24926121 1 27,763,886 WASF2 5′UTR Opensea − 0.11 0.01 1.02 × 10 −13 4.20 × 10 −8 cg19480937 3 16,467,544 RFTN1 Body Opensea − 0.05 0.01 2.97 × 10 −13 1.22 × 10 −7 cg22635075 22 19,551,881 IGR Shelf 0.08 0.01 1.23 × 10 −12 5.08 × 10 −7 cg24686610 3 147,125,782 ZIC4 TSS1500 Shore 0.09 0.01 1.82 × 10 −12 7.49 × 10 −7 cg25671977 8 40,755,559 ZMAT4 TSS1500 Shore 0.11 0.01 2.47 × 10 −12 1.02 × 10 −6 cg04541474 1 44,873,592 RNF220 5′UTR Island − 0.09 0.01 1.00 × 10 −11 4.13 × 10 −6 cg19940194 6 30,072,384 TRIM31 Body Shore 0.10 0.01 1.94 × 10 −11 8.00 × 10 −6 Open in a new tab CpG cytosine-guanine dinucleotide, CHR chromosome, IGR intergenic region, SE standard error, TC tropical cyclone a Flood/TC exposure was classified into three levels based on the event intensity: level 1 denoted days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 represented days with very large flood events as defined by the Dartmouth Flood Observatory or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); and level 3 corresponded to days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots) Notes: Complete list of significant CpG sites (Bonferroni-adjusted p values < 0.05) is provided in Additional file 2. The regression coefficients represented the average change in methylation level for each additional exposure day during the 1-year period preceding blood sampling, adjusting for individual’s age, educational level, marital status, smoking behavior, alcohol use, area-level socioeconomic status, survey year, survey season, annual average ambient temperature, and annual average relative humidity, and cell-type proportions. Position refers to Genome Research Consortium human genome build 37 (GRCh37)/UCSC human genome 19 (hg19) DMR analyses Based on the DMRcate method (FDR < 0.01), a total of 353, 696, and 949 significant DMRs were identified for flood/TC exposure at intensity levels 1 through 3, respectively (Additional file 1: Table S4). By comparison, the comb-p method found 74, 108, and 140 significant (Šidák p value < 0.01) DMRs. In total, there were 74, 108, and 138 DMRs for flood/TC exposure at intensity levels 1 through 3 identified by both DMRcate (FDR < 0.01) and comb-p (Šidák p value < 0.01) (219 unique DMRs; Additional file 1: Table S4). We provided detailed information (e.g., genome location, CpGs and genes) of all these DMRs in Additional file 4. Among the top ten significant DMRs, several genes, including AMT and C22orf45 , consistently appeared across different exposure intensities, whereas genes such as RNF39 and ACY3 emerged only at higher intensities (levels 2 and 3; Table 3 ). Higher exposure intensities generally involved DMRs with a greater number of CpG sites and stronger statistical significance. Table 3. Top ten significant DMRs a (ranked by lowest Šidák-adjusted p values) associated with flood/TC exposure by intensity levels CHR Start End No. of CpGs Šidák p value Gene Level 1 b 3 49,459,855 49,460,178 8 1.03 × 10 −14 AMT 8 144,120,335 144,120,707 7 2.43 × 10 −10 C8orf31 1 949,392 949,894 5 3.58 × 10 −10 ISG15 1 85,464,000 85,464,262 4 1.33 × 10 −9 MCOLN2 22 24,890,690 24,891,130 8 7.15 × 10 −9 C22orf45 5 140,857,192 140,858,108 9 8.82 × 10 −8 PCDHGA4 8 40,755,559 40,755,865 2 1.23 × 10 −6 ZMAT4 4 154,711,183 154,711,676 13 1.54 × 10 −6 SFRP2 17 80,708,279 80,708,633 4 7.72 × 10 −6 FN3K/TBCD 19 6,271,786 6,272,159 4 2.07 × 10 −5 MLLT1 Level 2 b 3 49,459,855 49,460,178 8 2.69 × 10 −12 AMT 6 30,038,754 30,039,601 37 1.06 × 10 −10 RNF39 11 67,417,958 67,418,406 13 2.74 × 10 −10 ACY3 11 60,775,545 60,776,400 6 9.07 × 10 −9 CD6 11 69,285,980 69,286,353 5 1.56 × 10 −8 1 179,555,999 179,556,254 2 2.18 × 10 −8 17 14,206,572 14,207,531 10 4.64 × 10 −8 HS3ST3B1 20 43,935,222 43,935,552 10 5.59 × 10 −8 MATN4/RBPJL 22 24,890,690 24,891,130 8 6.72 × 10 −8 C22orf45 1 85,464,000 85,464,262 4 2.12 × 10 −7 MCOLN2 Level 3 b 6 30,038,754 30,039,802 38 1.31 × 10 −13 RNF39 20 43,935,222 43,935,552 10 2.49 × 10 −12 MATN4/RBPJL 17 14,206,572 14,207,531 10 9.95 × 10 −12 HS3ST3B1 11 67,417,958 67,418,406 13 1.87 × 10 −11 ACY3 3 49,459,855 49,460,178 8 1.87 × 10 −11 AMT 22 24,890,690 24,891,142 9 6.21 × 10 −10 C22orf45 2 113,992,762 113,993,314 8 3.90 × 10 −9 PAX8 13 47,126,023 47,126,552 5 1.60 × 10 −8 LRCH1 1 179,555,999 179,556,254 2 1.74 × 10 −8 11 69,285,980 69,286,353 5 1.87 × 10 −8 Open in a new tab CpG cytosine-guanine dinucleotide, CHR chromosome, DFO Dartmouth Flood Observatory, DMR differentially methylated region, FDR false discovery rate, TC tropical cyclone a DMRs were identified by both DMRcate (FDR < 0.01) and comb-p (Šidák p value < 0.01) b Level 1 denotes days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 represents days with very large flood events or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); and level 3 corresponds to days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots) Notes: Complete list of significant DMRs is provided in Additional file 4. Position refers to Genome Research Consortium human genome build 37 (GRCh37)/UCSC human genome 19 (hg19) Annotated genes and the related human diseases For flood/TC exposure at intensity levels 1 through 3, the significant CpGs (Bonferroni-adjusted p value < 0.05) and significant DMRs (FDR < 0.01 and Šidák p value < 0.01) were mapped to 81, 165, and 122 genes, respectively (242 unique genes; Additional file 1: Table S5). Complete list of the genes (e.g., genome location and coefficients of the associations) by exposure intensity level was provided in Additional file 5. The most significant genes were ZFR2 at flood/TC exposure intensity levels 1 and 2 and ARHGEF10 at level 3, both showing positive associations between their methylation levels and flood/TC exposures (Fig. 3 ). According to the DisGeNET, the genes mapped to the significant CpGs and DMRs have been related to many human diseases or phenotypes (Additional file 1: Table S6). These mappings reflect overlaps between the genes implicated by our epigenetic signals and gene sets reported in disease databases, rather than providing evidence of a direct causal link to disease risk. Specifically, we observed that at higher flood/TC exposure intensities, the associated gene sets appeared to transition from those primarily involved in metabolic and endocrine disturbances to those associated with congenital genetic syndromes (e.g., SHORT syndrome) and hematologic malignancies (e.g., acute myelocytic leukemia). Enriched GO terms and KEGG pathways We identified 55, 60, and 53 significantly enriched GO terms for CpGs and DMRs related to flood/TC exposure at intensity levels 1 through 3, respectively (Additional file 1: Table S5). Detailed information of these GO terms (e.g., ontology and genes) by exposure intensity level was provided in Additional file 6. The significantly enriched GO terms were dominated by biological process across different exposure intensity levels. The most significantly enriched GO terms were consistently homophilic cell adhesion via plasma membrane adhesion molecules (biological process) and calcium ion binding (molecular function) across various flood/TC exposure intensity levels (Fig. 4 ). We identified 4 significantly enriched KEGG pathways for significant CpGs and DMRs related to flood/TC exposures, including Acute myeloid leukemia , Neuroactive ligand-receptor interaction , Adherens junction , and Epstein-Barr virus infection . Fig. 4. Open in a new tab Top ten significant genes and enriched GO terms mapped to significant CpGs by flood/tropical cyclone exposure intensity levels. The genes and GO terms were ranked by a smallest p value. The complete lists of all significant genes and GO terms were provided in Additional file 5 and Additional file 6, respectively. Notes: the mean coefficient referred to the average change in CpG-specific methylation levels associated with additional flood/tropical cyclone exposure day for all identified CpGs located in each specific gene. Flood/tropical cyclone exposure was classified into three levels based on the event intensity: level 1 denoted days with any flood events or TCs (maximum sustained wind speeds ≥ 34 knots); level 2 represented days with very large flood events or gale to violent storms (maximum sustained wind speeds ≥ 48 knots); and level 3 corresponded to days with extreme flood events or hurricanes (maximum sustained wind speeds ≥ 64 knots) Discussion Based on this twin and family study, we found that flood and TC exposures were associated with altered blood DNA methylation at many genomic loci (i.e., 164 CpGs and 219 DMRs). These loci mapped to 242 genes and were enriched for multiple biological processes, including homophilic cell adhesion and calcium-ion binding. These genes have been implicated in a range of human diseases or phenotypes, particularly endocrine and metabolic disorders (e.g., thyroid dysfunction and abnormalities in calcium homeostasis). More recent and higher-intensity exposures were linked to more extensive methylation changes, reflected by an increased number of significant CpGs and implicated genes. To date, studies on the potential impacts of climate disasters on human epigenetic modifications remain scarce. The majority of evidence has focused on other environmental exposures such as ambient air pollution and temperature [ 69 – 71 ]. Generally consistent findings demonstrated measurable alterations in DNA methylation associated with these exposures [ 48 , 69 , 72 , 73 ]. In comparison, we observed that flood and TC exposures were associated with markedly greater epigenetic perturbation, yielding substantially larger numbers of differentially methylated CpG sites and DMRs in exposed populations. For example, recent epigenome-wide association studies have identified 26 CpGs and 33 DMRs associated with long-term wildfire-related fine particulate matter exposure, and similarly modest methylation effects for temperature variation and urban air pollution exposures [ 49 , 73 , 74 ]. By contrast, we detected 164 differentially methylated CpG sites and 219 DMRs (mapping to 242 genes) associated with flood and TC exposure. These contrasts suggest that climatic events may elicit broader and more readily detectable perturbations of the peripheral-blood methylome than many chronic or lower-amplitude exposures. The high magnitude of the DNA methylation changes associated with TC and flood events may reflect the multiple concurrent stressors, including acute physical trauma, interruption of medical care, large-scale displacement, as well as acute and chronic psychosocial stress and rapid changes in housing and economic circumstances caused by such events [ 1 , 2 ]. This pattern underscores the importance of considering these understudied extreme climatic events as critical modifiers of genome-wide epigenetic regulation in population-level environmental health research. A consistent and biologically plausible pattern in our results is that more recent and higher-intensity flood and TC exposures were associated with a greater number of statistically significant CpG loci and more robust DMRs. Among exposure windows spanning 1 to 6 years, we found that exposures occurring within the 1-year period before blood sampling produced the strongest associations, and increasing exposure intensity was associated with increased number of CpG sites and DMRs characterized by more CpGs and larger effect sizes. Together, these findings suggest a proximate, dose-responsive epigenetic response to acute climatic disasters rather than a slow, monotonic accumulation of global methylation change. Similar patterns have been reported in studies of other environmental exposures, such as air pollution, where shorter-term or recent exposures were linked to more pronounced DNA methylation changes [ 71 , 75 – 77 ]. Although we did not find statistically significant associations at the level of global DNA methylation, there was a notable and consistent trend toward greater global methylation reductions with more recent and intense flood/TC exposures. The associations and differences might reach statistical significance if there was a larger sample size. Future larger and longitudinal studies are therefore warranted to confirm these trends. The understanding of biological mechanisms underlying the increased morbidity and mortality observed both acutely and long-term after floods and TCs is crucial for guiding intervention strategies, yet remains an area of limited evidence [ 78 ]. Our findings of the enriched pathways could provide some important biological insights into the health impacts of climate-related disasters. For example, we found that the differentially methylated CpGs and DMRs mapped to loci such as AMT , ZFR2 , ARHGEF10 , RNF39 , and ACY3 and showed consistent enrichment for homophilic cell adhesion via plasma membrane adhesion molecules and calcium ion binding across different exposure intensities. These suggest that the processes that are central to immune activation, barrier integrity, neuronal signaling, and endocrine regulation could be key biological processes affected by flood/TC exposures [ 79 – 81 ]. Notably, genes such as AMT and C22orf45 were consistently identified across all levels of exposure, while RNF39 and ACY3 emerged only under higher intensities, highlighting potential intensity-dependent gene regulation. This gradient may reflect that more severe disasters could provoke broader and more coordinated systemic responses (e.g., heightened inflammation, stronger HPA-axis activation, greater toxicant exposure and infection risk) that recruit additional biological pathways into the epigenetic response [ 82 – 84 ]. Alternatively, higher-intensity events could increase signal-to-noise, permitting detection of subtle or context-dependent loci relevant to the epigenetic response [ 85 , 86 ]. Intensity-specific loci may therefore represent (i) genes that are highly stress-sensitive and activated only under severe perturbation [ 87 ], (ii) genes influenced by secondary exposures induced by the severe events (e.g., contaminants, pathogens) [ 88 , 89 ], or (iii) signals arising from more pronounced activation or altered cellular states after intense events [ 90 , 91 ]. DisGeNET-based disease-mapping revealed that the differentially methylated genes not only overlapped with pathways relevant to inflammation and cellular communication but were also mapped to genomic regions previously associated with a spectrum of human diseases and phenotypes. At lower exposure intensities, biological signatures overlapped predominantly with endocrine-metabolic disorders, while higher flood/TC intensities were characterized by an enrichment of genes implicated in congenital syndromes and hematologic malignancies. While genetic mutations in these loci during embryonic development lead to congenital phenotypes, in adulthood, these genes continue to participate in core tissue maintenance, growth, and stress-signaling networks. Intense systemic stress or adverse exposure can disrupt these networks, leading to detectable epigenetic dysregulation that overlaps with genes associated with developmental processes and hematopoietic functions [ 92 – 94 ]. For example, enrichment for hematologic malignancy pathways plausibly reflected disaster-related perturbation of immune-hematopoietic processes (e.g., inflammation-driven hematopoietic stem/progenitor cell activation, viral reactivation, or genotoxic exposures), which could produce epigenetic signatures overlapping hematologic malignancy pathways [ 95 – 97 ]. Crucially, these findings represent pathway- or gene-level overlaps between our epigenetic signals and established disease-associated gene sets, rather than evidence that the observed methylation changes directly predict clinical disease risk. To our knowledge, this is the first study evaluating the human epigenetic impacts of floods and TCs. Leveraging a twin and family study design, we benefitted from within-sibship analyses, which controlled for confounding due to shared genetic background and family environment. This approach is particularly valuable given that genetic variation accounts for up to 20% of the variance in blood DNA methylation measured by the Illumina 450 K array [ 98 ]. Nonetheless, several limitations should be acknowledged. The absence of replication and functional validation datasets, such as gene expression profiles, limits further validation of our findings. Additionally, the lack of longitudinal methylation data precludes assessment of the persistence or reversibility of flood/TC-induced methylation changes over time. Future work should prioritize longitudinal, within-individual designs to distinguish transient versus persistent methylation changes, integrating methylation data with transcriptomic, proteomic, and cell-type-specific analyses to clarify functional consequences. Furthermore, our relatively small sample size has constrained statistical power to detect significant associations and conduct more granular analyses (e.g., subgroup or event-specific investigations [ 14 ]). Future studies with larger samples and a greater number of distinct events may be needed to disentangle hazard-specific effects. In addition, our analysis focused exclusively on autosomal CpGs. While this approach follows standard quality control protocols to minimize noise from XCI and potential probe cross-reactivity, it may have nonetheless omitted potentially relevant epigenetic signals. Future studies utilizing X-specific analytical frameworks may provide further insights into the epigenetic impact of climate-related disasters. Finally, we only included middle-aged women from Australia (mainly of European ancestry), and methylation was assessed in peripheral blood, which may not fully reflect epigenetic patterns in other relevant tissues such as lung, heart, brain, or kidney [ 99 ]. Further studies in men, children, other ethnicities, or DNA methylation profiling in other tissues are warranted to expand the knowledge on the epigenetic impacts of climate disasters. Conclusions In summary, exposure to floods and TCs was associated with differential blood DNA methylation at many genomic loci in Australian women. These signals mapped to a range of human diseases and to biological pathways involved in inflammation, immune function, and development, with intensity-dependent shifts in methylation patterns. Such signatures may serve as candidate biomarkers of disaster exposure and offer insights into the biological mechanisms underlying both acute and long-term adverse health outcomes following such events. Further research is highly warranted to expand the currently limited epigenetic evidence on the health effects of extreme weather events. Supplementary Information 12916_2026_4868_MOESM1_ESM.docx (9.9MB, docx) Additional file 1. Tables S1–S6; Figures S1–S7. Table S1 Input parameters for DMRcate and comb-p for identifying significant DMRs. Table S2 Number of significant CpGs with methylation level associated with flood/TC exposures in different exposure windows. Table S3 Associations of global DNA methylation with flood/TC exposures by intensity levels. Table S4 Number of identified DMRs associated with flood/TC exposure by intensity levels. Table S5 Number of genes and significantly enriched GO terms by flood/TC exposure intensity levels. Table S6 Top 3 diseases related to genes with methylation level associated with flood/TC exposures. Figure S1 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC by exposure windows. Figure S2 Global and CpG-specific DNA methylation associated with flood/TC exposure versus non-exposure, stratified by exposure intensity. Figure S3 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels, with and without adjustment for urban–rural residency. Figure S4 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels, with and without adjustment for smoking and drinking status. Figure S5 Intrinsic and extrinsicassociations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels. Figure S6 Comparison of DNA methylation associations with flood/TC exposure intensity using self-reported smoking status versus a DNA methylation-based smoking index. Figure S7 Q-Q plots for the p values of CpG-specific associated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM2_ESM.xlsx (510.3KB, xlsx) Additional file 2. Full list of genome-wide CpG sites with methylation levels significantlyassociated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM3_ESM.xlsx (41.2KB, xlsx) Additional file 3. Full list of genome-wide CpG sites with methylation levels significantlyassociated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM4_ESM.xlsx (36.9KB, xlsx) Additional file 4. Full list of DMRs with methylation levels significantly associated with flood/TC exposure by intensity levels, identified by both comb-pand DMRcate. 12916_2026_4868_MOESM5_ESM.xlsx (31KB, xlsx) Additional file 5. Full list of genes mapped to significant CpGs and DMRs by flood/TC exposure intensity levels. 12916_2026_4868_MOESM6_ESM.xlsx (26.1KB, xlsx) Additional file 6. Full list of significantly enriched GO terms by flood/TC exposure intensity levels. Acknowledgements We thank all the twins and their family members who participated in the study. Authors’ contributions WH: Conceptualization, Data curation, Methodology, Formal analysis, Visualization, Software, Writing – original draft; RX: Conceptualization, Data curation, Methodology, Writing – review & editing; YW: Data curation, Writing – review & editing; ZY: Data curation, Writing – review & editing; EMW: Data curation, Writing – review & editing; MCS: Data curation, Writing – review & editing; JLH: Data curation, Resources, Writing – review & editing; MJA: Supervision, Writing – review & editing; SL1: Supervision, Writing – review & editing; SL2: Supervision, Data curation, Methodology, Software, Writing – review & editing; YG: Conceptualization, Supervision, Resources, Methodology, Funding acquisition, Writing – review & editing. All authors read and approved the final manuscript. Funding This work was supported by China Scholarship Council (grant numbers 202006380055 and 202006010044), VicHealth Postdoctoral Research Fellowships, Monash Graduate Scholarship, Australian National Health and Medical Research Council (grant numbers APP1163693, APP1109193, APP1155163, 1050561, and 1079102), Victorian Cancer Agency (grant number ECRF 19020), and Cancer Australia and National Breast Cancer Foundation (grant number 509307). Data availability The raw and processed DNA methylation data set are open accessed or free on Gene Expression Omnibus (accession number GSE100227 ). As required by the ethics approval, the other data (e.g., covariates) are not allowed to be open for access. However, these data could be accessed on reasonable request to Dr. Shuai Li (shuai.li@unimelb.edu.au). Declarations Ethics approval and consent to participate Written informed consent was obtained from all participants prior to enrollment. Ethical approval was received from the Monash University Human Research Ethics Committee (reference number: 17813). Consent for publication This manuscript does not contain any individual person’s data in any form (including individual details, images, or videos). Therefore, consent for publication was not required. Competing interests The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: MJA held investigator-initiated grants from Pfizer, Boehringer-Ingelheim, GlaxoSmithKline and Sanofi for unrelated research. He has undertaken an unrelated consultancy for Sanofi. He also received a speaker’s fee from GSK. The other authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Wenzhong Huang and Rongbin Xu contributed equally to this work. Shuai Li and Yuming Guo are joint senior authors. Contributor Information Shuai Li, Email: shuai.li@unimelb.edu.au. Yuming Guo, Email: yuming.guo@monash.edu. References 1. Huang W, Gao Y, Xu R, Yang Z, Yu P, Ye T, et al. Health effects of cyclones: a systematic review and meta-analysis of epidemiological studies. Environ Health Perspect. 2023;131(8):86001. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Yang Z, Huang W, McKenzie JE, Yu P, Ju K, Wu Y, et al. Mortality and morbidity risks associated with floods: a systematic review and meta-analysis. Environ Res. 2024;263(3):120263. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Liu Q, Du M, Wang Y, Deng J, Yan W, Qin C, et al. Global, regional and national trends and impacts of natural floods, 1990-2022. Bull World Health Organ. 2024;102(6):410–20. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Jing R, Heft-Neal S, Chavas DR, Griswold M, Wang Z, Clark-Ginsberg A, et al. Global population profile of tropical cyclone exposure from 2002 to 2019. Nature. 2024;626(7999):549–54. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Allianz SE. The global economic ripple effect of cyclones 2024. Available from: https://www.allianz.com/en/economic_research/insights/publications/specials_fmo/241009-economy-cyclones.html . 6. Cruz-Cano R, Mead EL. Causes of excess deaths in Puerto Rico after Hurricane Maria: a time-series estimation. Am J Public Health. 2019;109(7):1050–2. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Kim S, Kulkarni PA, Rajan M, Thomas P, Tsai S, Tan C, et al. Hurricane Sandy (New Jersey): mortality rates in the following month and quarter. Am J Public Health. 2017;107(8):1304–7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Dosa DM, Skarha J, Peterson LJ, Jester DJ, Sakib N, Ogarek J, et al. Association between exposure to Hurricane Irma and mortality and hospitalization in Florida nursing home residents. JAMA Netw Open. 2020;3(10):e2019460. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Dosa D, Feng Z, Hyer K, Brown LM, Thomas K, Mor V. Effects of Hurricane Katrina on nursing facility resident mortality, hospitalization, and functional decline. Disaster Med Public Health Prep. 2010;4 Suppl 1(0 1):S28–32. [ DOI ] [ PMC free article ] [ PubMed ] 10. Yang Z, Huang W, McKenzie JE, Xu R, Yu P, Ye T, et al. Mortality risks associated with floods in 761 communities worldwide: time series study. BMJ. 2023;383:e075081. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Lynch VD, Sullivan JA, Flores AB, Xie X, Aggarwal S, Nethery RC, et al. Large floods drive changes in cause-specific mortality in the United States. Nat Med. 2025;31(2):663–71. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Young R, Hsiang S. Mortality caused by tropical cyclones in the United States. Nature. 2024;635(8037):121–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Parks RM, Benavides J, Anderson GB, Nethery RC, Navas-Acien A, Dominici F, et al. Association of tropical cyclones with county-level mortality in the US. JAMA. 2022;327(10):946–55. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Huang W, Yang Z, Zhang Y, Vogt T, Armstrong B, Yu W, et al. Tropical cyclone-specific mortality risks and the periods of concern: a multicountry time-series study. PLoS Med. 2024;21(1):e1004341. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Huang W, Li S, Vogt T, Xu R, Tong S, Molina T, et al. Global short-term mortality risk and burden associated with tropical cyclones from 1980 to 2019: a multi-country time-series study. Lancet Planet Health. 2023;7(8):e694–705. [ DOI ] [ PubMed ] [ Google Scholar ] 16. Yang Z, Huang W, McKenzie JE, Xu R, Yu P, Wu Y, et al. Hospitalization risks associated with floods in a multi-country study. Nat Water. 2025;3(5):561–70. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Parks RM, Anderson GB, Nethery RC, Navas-Acien A, Dominici F, Kioumourtzoglou MA. Tropical cyclone exposure is associated with increased hospitalization rates in older adults. Nat Commun. 2021;12(1):1545. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Huang W, Yang Z, Zhang Y, Vogt T, Park J, Yu W, et al. Hospitalization risk and burden for cause-specific cardiovascular diseases following tropical cyclones: a multicountry study. Sci Adv. 2025;11(31):eadr0800. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Bozick R. The effects of Hurricane Harvey on the physical and mental health of adults in Houston. Health Place. 2021;72:102697. [ DOI ] [ PubMed ] [ Google Scholar ] 20. An R, Qiu Y, Xiang X, Ji M, Guan C. Impact of Hurricane Katrina on mental health among US adults. Am J Health Behav. 2019;43(6):1186–99. [ DOI ] [ PubMed ] [ Google Scholar ] 21. Parks RM, Kontis V, Anderson GB, Baldwin JW, Danaei G, Toumi R, et al. Short-term excess mortality following tropical cyclones in the United States. Sci Adv. 2023;9(33):eadg6633. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Knutson T, Camargo SJ, Chan JCL, Emanuel K, Ho C-H, Kossin J, et al. Tropical cyclones and climate change assessment: part II: projected response to anthropogenic warming. Bull Am Meteorol Soc. 2020;101(3):E303–22. [ Google Scholar ] 23. Wang G, Wu L, Mei W, Xie SP. Ocean currents show global intensification of weak tropical cyclones. Nature. 2022;611(7936):496–500. [ DOI ] [ PubMed ] [ Google Scholar ] 24. Huang X, Yin J, Slater LJ, Kang S, He S, Liu P. Global projection of flood risk with a bivariate framework under 1.5–3.0°C warming levels. Earths Future. 2024;12(4):e2023EF004312. [ Google Scholar ] 25. Watowich MM, Chiou KL, Montague MJ, Simons ND, Horvath JE, Ruiz-Lambides AV, et al. Natural disaster and immunological aging in a nonhuman primate. Proc Natl Acad Sci U S A. 2022;119(8):e2121663119. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Li S, Wong EM, Southey MC, Hopper JL. Association between DNA methylation at SOCS3 gene and body mass index might be due to familial confounding. Int J Obes (Lond). 2017;41(6):995–6. [ DOI ] [ PubMed ] [ Google Scholar ] 27. Xu R, Li S, Li SS, Wong EM, Southey MC, Hopper JL, et al. Surrounding greenness and biological aging based on DNA methylation: a twin and family study in Australia. Environ Health Persp. 2021;129(8):87007. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Odefrey F, Stone J, Gurrin LC, Byrnes GB, Apicella C, Dite GS, et al. Common genetic variants associated with breast cancer and mammographic density measures that predict disease. Cancer Res. 2010;70(4):1449–58. [ DOI ] [ PubMed ] [ Google Scholar ] 29. Li S, Wong EM, Joo JE, Jung CH, Chung J, Apicella C, et al. Genetic and environmental causes of variation in the difference between biological age based on DNA methylation and chronological age for middle-aged women. Twin Res Hum Genet. 2015;18(6):720–6. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Li S, Wong EM, Bui M, Nguyen TL, Joo JHE, Stone J, et al. Causal effect of smoking on DNA methylation in peripheral blood: a twin and family study. Clin Epigenetics. 2018;10:18. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Li S, Wong EM, Bui M, Nguyen TL, Joo JHE, Stone J, et al. Inference about causation between body mass index and DNA methylation in blood from a twin family study. Int J Obesity. 2019;43(2):243–52. [ DOI ] [ PubMed ] [ Google Scholar ] 32. Sandoval J, Heyn HA, Moran S, Serra-Musach J, Pujana MA, Bibikova M, et al. Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome. Epigenetics-Us. 2011;6(6):692–702. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30(10):1363–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. Maksimovic J, Gordon L, Oshlack A. SWAN: subset-quantile within array normalization for illumina infinium HumanMethylation450 BeadChips. Genome Biol. 2012;13(6):R44. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. 2007;8(1):118–27. [ DOI ] [ PubMed ] [ Google Scholar ] 36. Zheng Y, Joyce BT, Liu L, Zhang Z, Kibbe WA, Zhang W, et al. Prediction of genome-wide DNA methylation in repetitive elements. Nucleic Acids Res. 2017;45(15):8697–711. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Carozza DA, Boudreault M. A global flood risk modeling framework built with climate models and machine learning. J Adv Model Earth Syst. 2021;13(4):e2020MS002221. [ Google Scholar ] 38. Global active archive of large flood events. Dartmouth Flood Observatory, University of Colorado. Available from: https://floodobservatory.colorado.edu/ (accessed 12 December 2023). 39. Najibi N, Devineni N. Recent trends in the frequency and duration of global floods. Earth Syst Dynam. 2018;9(2):757–83. [ Google Scholar ] 40. Tellman B, Sullivan JA, Kuhn C, Kettner AJ, Doyle CS, Brakenridge GR, et al. Satellite imaging reveals increased proportion of population exposed to floods. Nature. 2021;596(7870):80–6. [ DOI ] [ PubMed ] [ Google Scholar ] 41. Holland G. A revised hurricane pressure–wind model. Mon Weather Rev. 2008;136(9):3432–45. [ Google Scholar ] 42. Peduzzi P, Chatenoux B, Dao H, De Bono A, Herold C, Kossin J, et al. Global trends in tropical cyclone risk. Nat Clim Chang. 2012;2(4):289–94. [ Google Scholar ] 43. Lange S, Volkholz J, Geiger T, Zhao F, Vega I, Veldkamp T, et al. Projecting exposure to extreme climate impact events across six event categories and three spatial scales. Earths Future. 2020;8(12):e2020EF001616. [ Google Scholar ] 44. Geiger T, Gütschow J, Bresch DN, Emanuel K, Frieler K. Double benefit of limiting global warming for tropical cyclone exposure. Nat Clim Chang. 2021;11(10):861–6. [ Google Scholar ] 45. Knapp KR, Kruk MC, Levinson DH, Diamond HJ, Neumann CJ. The international best track archive for climate stewardship (IBTrACS) unifying tropical cyclone data. Bull Am Meteorol Soc. 2010;91(3):363–76. [ Google Scholar ] 46. He C, Chen R, Kim H, Hashizume M, Lee W, Honda Y, et al. Tropical cyclone and daily respiratory mortality across East Asia: a time series study. Eur Respir J. 2023;62(1):2300546. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 47. Australian Bureau of Statistics. Australian social trends 2010. Available from: https://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/4102.0Main+Features1Dec%202010?OpenDocument . 48. Choi EY, Ailshire JA. Ambient outdoor heat and accelerated epigenetic aging among older adults in the US. Sci Adv. 2025;11(9):eadr0616. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Xu R, Li S, Wu Y, Yue X, Wong EM, Southey MC, et al. Wildfire-related PM(2.5) and DNA methylation: an Australian twin and family study. Environ Int. 2023;171:107704. [ DOI ] [ PubMed ] 50. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012;13:86. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Australian Bureau of Statistics. Australian Statistical Geography Standard (ASGS): Volume 4 – Significant Urban Areas, Urban Centres and Localities, Section of State, July 2011. Canberra: Australian Bureau of Statistics; Report No.: 1270.0.55.004. Available from: https://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/1270.0.55.004July%202011 . 52. Jeffrey SJ, Carter JO, Moodie KB, Beswick AR. Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environ Model Softw. 2001;16(4):309–30. [ Google Scholar ] 53. van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45(3):1–67. [ Google Scholar ] 54. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc: Ser B (Methodol). 1995;57(1):289–300. [ Google Scholar ] 55. van Iterson M, van Zwet EW, Heijmans BT. Controlling bias and inflation in epigenome- and transcriptome-wide association studies using the empirical null distribution. Genome Biol. 2017;18(1):19. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Horvath S, Raj K. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet. 2018;19(6):371–84. [ DOI ] [ PubMed ] [ Google Scholar ] 57. Gao X, Zhang Y, Saum KU, Schöttker B, Breitling LP, Brenner H. Tobacco smoking and smoking-related DNA methylation are associated with the development of frailty among older adults. Epigenetics US. 2017;12(2):149–56. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Pedersen BS, Schwartz DA, Yang IV, Kechris KJ. Comb-p: software for combining, analyzing, grouping and correcting spatially correlated p-values. Bioinformatics. 2012;28(22):2986–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 59. Peters TJ, Buckley MJ, Statham AL, Pidsley R, Samaras K, R VL, et al. De novo identification of differentially methylated regions in the human genome. Epigenetics Chromatin. 2015;8:6. [ DOI ] [ PMC free article ] [ PubMed ] 60. Lent S, Xu H, Wang L, Wang Z, Sarnowski C, Hivert MF, et al. Comparison of novel and existing methods for detecting differentially methylated regions. BMC Genet. 2018;19(Suppl 1):84. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 61. Bondhus L, Wei A, Arboleda VA. Dmrscaler: a scale-aware method to identify regions of differential DNA methylation spanning basepair to multi-megabase features. BMC Bioinformatics. 2022;23(1):364. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 62. Gruzieva O, Xu CJ, Yousefi P, Relton C, Merid SK, Breton CV, et al. Prenatal particulate air pollution and DNA methylation in newborns: an epigenome-wide meta-analysis. Environ Health Perspect. 2019;127(5):57012. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 63. Lee MK, Xu CJ, Carnes MU, Nichols CE, Ward JM, Kwon SO, et al. Genome-wide DNA methylation and long-term ambient air pollution exposure in Korean adults. Clin Epigenetics. 2019;11(1):37. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 64. Piñero J, Ramírez-Anguita JM, Saüch-Pitarch J, Ronzano F, Centeno E, Sanz F, et al. The DisGeNET knowledge platform for disease genomics: 2019 update. Nucleic Acids Res. 2020;48(D1):D845–55. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Maksimovic J, Oshlack A, Phipson B. Gene set enrichment analysis for genome-wide DNA methylation data. Genome Biol. 2021;22(1):173. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 66. Phipson B, Maksimovic J, Oshlack A. missMethyl: an R package for analyzing data from Illumina’s HumanMethylation450 platform. Bioinformatics. 2016;32(2):286–8. [ DOI ] [ PubMed ] [ Google Scholar ] 67. Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation (Camb). 2021;2(3):100141. [ DOI ] [ PMC free article ] [ PubMed ] 68. Yu G, Li F, Qin Y, Bo X, Wu Y, Wang S. GOSemSim: an R package for measuring semantic similarity among GO terms and gene products. Bioinformatics. 2010;26(7):976–8. [ DOI ] [ PubMed ] [ Google Scholar ] 69. Xu R, Li S, Guo S, Zhao Q, Abramson MJ, Li S, et al. Environmental temperature and human epigenetic modifications: a systematic review. Environ Pollut. 2020;259:113840. [ DOI ] [ PubMed ] [ Google Scholar ] 70. Zetzsche J, Fallet M. To live or let die? Epigenetic adaptations to climate change-a review. Environ Epigenet. 2024;10(1):dvae009. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 71. Eze IC, Jeong A, Schaffner E, Rezwan FI, Ghantous A, Foraster M, et al. Genome-wide DNA methylation in peripheral blood and long-term exposure to source-specific transportation noise and air pollution: the SAPALDIA study. Environ Health Perspect. 2020;128(6):67003. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 72. Bind MA, Lepeule J, Zanobetti A, Gasparrini A, Baccarelli A, Coull BA, et al. Air pollution and gene-specific methylation in the Normative Aging Study: association, effect modification, and mediation analysis. Epigenetics-Us. 2014;9(3):448–58. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 73. Chen Y, Dong GH, Li S, Liu Y, Li S, Guo Y, et al. The associations between exposure to ambient air pollution and coagulation markers and the potential effects of DNA methylation. J Hazard Mater. 2024;480:136433. [ DOI ] [ PubMed ] [ Google Scholar ] 74. Wu Y, Xu R, Li S, Ming Wong E, Southey MC, Hopper JL, et al. Epigenome-wide association study of short-term temperature fluctuations based on within-sibship analyses in Australian females. Environ Int. 2023;171:107655. [ DOI ] [ PubMed ] [ Google Scholar ] 75. Zhang M, Hu T, Ma T, Huang W, Wang Y. Epigenetics and environmental health. Front Med. 2024;18(4):571–96. [ DOI ] [ PubMed ] [ Google Scholar ] 76. Baccarelli A, Wright RO, Bollati V, Tarantini L, Litonjua AA, Suh HH, et al. Rapid DNA methylation changes after exposure to traffic particles. Am J Respir Crit Care Med. 2009;179(7):572–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 77. Dhingra R, Nwanaji-Enwerem JC, Samet M, Ward-Caviness CK. DNA methylation age-environmental influences, health impacts, and its role in environmental epidemiology. Curr Environ Health Rep. 2018;5(3):317–27. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 78. Prada D, Baccarelli AA, Kupsco A, Parks RM. Climate change and health: understanding mechanisms will inform mitigation and prevention strategies. Nat Med. 2024;30(6):1522–4. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 79. Harjunpää H, Llort Asens M, Guenther C, Fagerholm SC. Cell adhesion molecules and their roles and regulation in the immune and tumor microenvironment. Front Immunol. 2019;10:1078. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 80. Yamada S, Nelson WJ. Synapses: sites of cell recognition, adhesion, and functional specification. Annu Rev Biochem. 2007;76:267–94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 81. Bellassen L, David K, Lampert B, Sarusi-Portuguez A, Tsoory M, Lubliner J, et al. The immune receptor SLAMF5 regulates myeloid-cell mediated neuroinflammation in multiple sclerosis. PLoS Biol. 2025;23(9):e3003373. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 82. Dobson GP, Morris JL, Letson HL. Immune dysfunction following severe trauma: a systems failure from the central nervous system to mitochondria. Front Med. 2022;9:968453. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 83. Sharma A, Steven S, Bosmann M. The pituitary gland prevents shock-associated death by controlling multiple inflammatory mediators. Biochem Biophys Res Commun. 2019;509(1):188–93. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 84. Van den Berghe G, Téblick A, Langouche L, Gunst J. The hypothalamus-pituitary-adrenal axis in sepsis- and hyperinflammation-induced critical illness: gaps in current knowledge and future translational research directions. EBioMedicine. 2022;84:104284. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 85. Sumner JA, Gambazza S, Gao X, Baccarelli AA, Uddin M, McLaughlin KA. Epigenetics of early-life adversity in youth: cross-sectional and longitudinal associations. Clin Epigenetics. 2022;14(1):48. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 86. Lussier AA, Zhu Y, Smith BJ, Simpkin AJ, Smith A, Suderman MJ, et al. Sensitive periods for the effect of childhood adversity on DNA methylation: updated results from a prospective, longitudinal study. Biol Psychiatry Glob Open Sci. 2023;3(3):567–71. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 87. Vojta A, Zoldoš V. Adaptation or malignant transformation: the two faces of epigenetically mediated response to stress. Biomed Res Int. 2013;2013:954060. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 88. Matosin N, Cruceanu C, Binder EB. Preclinical and clinical evidence of DNA methylation changes in response to trauma and chronic stress. Chron Stress. 2017;1:2470547017710764. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 89. Teker T, Sefer Ö, Gazdağlı A, Yörük E, Varol Gİ, Albayrak G. α-Thujone exhibits an antifungal activity against F. graminearum by inducing oxidative stress, apoptosis, epigenetics alterations and reduced toxin synthesis. Eur J Plant Pathol. 2021;160(3):611–22. [ Google Scholar ] 90. Gatta E, Saudagar V, Auta J, Grayson DR, Guidotti A. Epigenetic landscape of stress surfeit disorders: key role for DNA methylation dynamics. Int Rev Neurobiol. 2021;156:127–83. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 91. Leung CS, Kosyk O, Welter EM, Dietrich N, Archer TK, Zannas AS. Chronic stress-driven glucocorticoid receptor activation programs key cell phenotypes and functional epigenomic patterns in human fibroblasts. iScience. 2022;25(9):104960. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 92. Aref-Eshghi E, Rodenhiser DI, Schenkel LC, Lin H, Skinner C, Ainsworth P, et al. Genomic DNA methylation signatures enable concurrent diagnosis and clinical genetic variant classification in neurodevelopmental syndromes. Am J Hum Genet. 2018;102(1):156–74. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 93. Petroff RL, Dolinoy DC, Wang K, Montrose L, Padmanabhan V, Peterson KE, et al. Translational toxicoepigenetic meta-analyses identify homologous gene DNA methylation reprogramming following developmental phthalate and lead exposure in mouse and human offspring. Environ Int. 2024;186:108575. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 94. Brennan K, Zheng H, Fahrner JA, Shin JH, Gentles AJ, Schaefer B, et al. NSD1 mutations deregulate transcription and DNA methylation of bivalent developmental genes in Sotos syndrome. Hum Mol Genet. 2022;31(13):2164–84. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 95. Nwosu GO, Powell JA, Pitson SM. Targeting the integrated stress response in hematologic malignancies. Exp Hematol Oncol. 2022;11(1):94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 96. Kim MJ, Kwak J, Kim HS, Han J, Kang JA, Jo J, et al. A novel squaramide derivative, HR-19011, induces the integrated stress response via the HRI-eIF2α-ATF4 pathway, effectively inhibiting hematologic malignancies. Mol Cancer Ther. 2025;24(10):1546–60. [ DOI ] [ PubMed ] [ Google Scholar ] 97. Hasan M, Tannu S, Saini S, Ugidos M, Garcia-Manero G, Platzbecker U, et al. Luspatercept reduces clonal hematopoiesis-associated cardiac stress via modulation of mTORC1 and inflammatory signaling pathways. Blood. 2024;144:4046. [ Google Scholar ] 98. van Dongen J, Nivard MG, Willemsen G, Hottenga JJ, Helmer Q, Dolan CV, et al. Genetic and environmental influences interact with age and sex in shaping the human methylome. Nat Commun. 2016;7:11115. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 99. Mill J, Heijmans BT. From promises to practical strategies in epigenetic epidemiology. Nat Rev Genet. 2013;14(8):585–94. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials 12916_2026_4868_MOESM1_ESM.docx (9.9MB, docx) Additional file 1. Tables S1–S6; Figures S1–S7. Table S1 Input parameters for DMRcate and comb-p for identifying significant DMRs. Table S2 Number of significant CpGs with methylation level associated with flood/TC exposures in different exposure windows. Table S3 Associations of global DNA methylation with flood/TC exposures by intensity levels. Table S4 Number of identified DMRs associated with flood/TC exposure by intensity levels. Table S5 Number of genes and significantly enriched GO terms by flood/TC exposure intensity levels. Table S6 Top 3 diseases related to genes with methylation level associated with flood/TC exposures. Figure S1 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC by exposure windows. Figure S2 Global and CpG-specific DNA methylation associated with flood/TC exposure versus non-exposure, stratified by exposure intensity. Figure S3 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels, with and without adjustment for urban–rural residency. Figure S4 Associations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels, with and without adjustment for smoking and drinking status. Figure S5 Intrinsic and extrinsicassociations of global DNA methylation and genome-wide CpG-specific DNA methylation with flood/TC exposures by intensity levels. Figure S6 Comparison of DNA methylation associations with flood/TC exposure intensity using self-reported smoking status versus a DNA methylation-based smoking index. Figure S7 Q-Q plots for the p values of CpG-specific associated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM2_ESM.xlsx (510.3KB, xlsx) Additional file 2. Full list of genome-wide CpG sites with methylation levels significantlyassociated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM3_ESM.xlsx (41.2KB, xlsx) Additional file 3. Full list of genome-wide CpG sites with methylation levels significantlyassociated with flood/TC exposures by intensity levels. 12916_2026_4868_MOESM4_ESM.xlsx (36.9KB, xlsx) Additional file 4. Full list of DMRs with methylation levels significantly associated with flood/TC exposure by intensity levels, identified by both comb-pand DMRcate. 12916_2026_4868_MOESM5_ESM.xlsx (31KB, xlsx) Additional file 5. Full list of genes mapped to significant CpGs and DMRs by flood/TC exposure intensity levels. 12916_2026_4868_MOESM6_ESM.xlsx (26.1KB, xlsx) Additional file 6. Full list of significantly enriched GO terms by flood/TC exposure intensity levels. Data Availability Statement The raw and processed DNA methylation data set are open accessed or free on Gene Expression Omnibus (accession number GSE100227 ). As required by the ethics approval, the other data (e.g., covariates) are not allowed to be open for access. However, these data could be accessed on reasonable request to Dr. Shuai Li (shuai.li@unimelb.edu.au). Articles from BMC Medicine are provided here courtesy of BMC ACTIONS View on publisher site PDF (3.2 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Evaluating SNP confounding in epigenetic clocks: Application to Native Hawaiian and Pacific Islander DNA methylation data#1166538
• DNA Methylation Markers in Veterans Exposed to Open Burn Pits#885973
• Communicating Climate Hazards and Risks: A Systematic Review of Climate Change Perception in Small Island Communities#14535
• Revitalizing SDG 17 for Climate Disaster Resilience : A Grassroots Study from Bogura, Bangladesh#20761
• Revitalizing SDG 17 for Climate Disaster Resilience : A Grassroots Study from Bogura, Bangladesh#20760